{
    "segments": [
        {
            "id": 1,
            "seek": 2764,
            "start": 4.1,
            "end": 11.22,
            "text": " Here's our second talk of the day. It's Eric Gallican.",
            "tokens": [
                50412,
                1692,
                311,
                527,
                1150,
                751,
                295,
                264,
                786,
                13,
                467,
                311,
                9336,
                14588,
                8914,
                13,
                50888
            ],
            "temperature": 0.0,
            "avg_logprob": -0.35620470841725665,
            "compression_ratio": 1.3563218390804597,
            "no_speech_prob": 0.22802734375,
            "words": [
                {
                    "start": 4.1,
                    "end": 4.78,
                    "word": " Here's",
                    "probability": 0.625732421875
                },
                {
                    "start": 4.78,
                    "end": 4.9,
                    "word": " our",
                    "probability": 0.79296875
                },
                {
                    "start": 4.9,
                    "end": 5.16,
                    "word": " second",
                    "probability": 0.79150390625
                },
                {
                    "start": 5.16,
                    "end": 5.44,
                    "word": " talk",
                    "probability": 0.8134765625
                },
                {
                    "start": 5.44,
                    "end": 5.58,
                    "word": " of",
                    "probability": 0.90087890625
                },
                {
                    "start": 5.58,
                    "end": 5.68,
                    "word": " the",
                    "probability": 0.81591796875
                },
                {
                    "start": 5.68,
                    "end": 5.96,
                    "word": " day",
                    "probability": 0.90234375
                },
                {
                    "start": 5.96,
                    "end": 6.64,
                    "word": ".",
                    "probability": 0.85693359375
                },
                {
                    "start": 7.62,
                    "end": 8.3,
                    "word": " It's",
                    "probability": 0.94384765625
                },
                {
                    "start": 8.3,
                    "end": 8.74,
                    "word": " Eric",
                    "probability": 0.6591796875
                },
                {
                    "start": 8.74,
                    "end": 10.88,
                    "word": " Gallican",
                    "probability": 0.547607421875
                },
                {
                    "start": 10.88,
                    "end": 11.22,
                    "word": ".",
                    "probability": 0.8564453125
                }
            ]
        },
        {
            "id": 2,
            "seek": 2764,
            "start": 12.32,
            "end": 19.72,
            "text": " He's a... well, you see his Who Am I page, but he's also a PhD student at Brandeis and",
            "tokens": [
                50952,
                634,
                311,
                257,
                485,
                731,
                11,
                291,
                536,
                702,
                2102,
                2012,
                286,
                3028,
                11,
                457,
                415,
                311,
                611,
                257,
                14476,
                3107,
                412,
                11119,
                68,
                271,
                293,
                51328
            ],
            "temperature": 0.0,
            "avg_logprob": -0.35620470841725665,
            "compression_ratio": 1.3563218390804597,
            "no_speech_prob": 0.22802734375,
            "words": [
                {
                    "start": 12.32,
                    "end": 12.68,
                    "word": " He's",
                    "probability": 0.783203125
                },
                {
                    "start": 12.68,
                    "end": 13.1,
                    "word": " a...",
                    "probability": 0.44256591796875
                },
                {
                    "start": 13.1,
                    "end": 13.46,
                    "word": " well",
                    "probability": 0.13525390625
                },
                {
                    "start": 13.46,
                    "end": 13.58,
                    "word": ",",
                    "probability": 0.791015625
                },
                {
                    "start": 13.58,
                    "end": 13.7,
                    "word": " you",
                    "probability": 0.853515625
                },
                {
                    "start": 13.7,
                    "end": 13.98,
                    "word": " see",
                    "probability": 0.64208984375
                },
                {
                    "start": 13.98,
                    "end": 14.24,
                    "word": " his",
                    "probability": 0.8642578125
                },
                {
                    "start": 14.24,
                    "end": 14.4,
                    "word": " Who",
                    "probability": 0.5009765625
                },
                {
                    "start": 14.4,
                    "end": 14.48,
                    "word": " Am",
                    "probability": 0.4462890625
                },
                {
                    "start": 14.48,
                    "end": 14.6,
                    "word": " I",
                    "probability": 0.99853515625
                },
                {
                    "start": 14.6,
                    "end": 15.0,
                    "word": " page",
                    "probability": 0.57470703125
                },
                {
                    "start": 15.0,
                    "end": 15.16,
                    "word": ",",
                    "probability": 0.587890625
                },
                {
                    "start": 15.16,
                    "end": 15.4,
                    "word": " but",
                    "probability": 0.86474609375
                },
                {
                    "start": 15.4,
                    "end": 15.7,
                    "word": " he's",
                    "probability": 0.948486328125
                },
                {
                    "start": 15.7,
                    "end": 15.9,
                    "word": " also",
                    "probability": 0.82421875
                },
                {
                    "start": 16.32,
                    "end": 16.8,
                    "word": " a PhD",
                    "probability": 0.84130859375
                },
                {
                    "start": 16.8,
                    "end": 17.4,
                    "word": " student",
                    "probability": 0.8349609375
                },
                {
                    "start": 17.4,
                    "end": 18.76,
                    "word": " at",
                    "probability": 0.89990234375
                },
                {
                    "start": 18.76,
                    "end": 19.28,
                    "word": " Brandeis",
                    "probability": 0.9148763020833334
                },
                {
                    "start": 19.28,
                    "end": 19.72,
                    "word": " and",
                    "probability": 0.54345703125
                }
            ]
        },
        {
            "id": 3,
            "seek": 2764,
            "start": 21.48,
            "end": 28.08,
            "text": " an all-around awesome guy. He does way too much stuff and is very enthusiastic and involved in",
            "tokens": [
                51408,
                364,
                439,
                12,
                25762,
                3476,
                2146,
                13,
                634,
                775,
                636,
                886,
                709,
                1507,
                293,
                307,
                588,
                28574,
                293,
                3288,
                294,
                51740
            ],
            "temperature": 0.0,
            "avg_logprob": -0.35620470841725665,
            "compression_ratio": 1.3563218390804597,
            "no_speech_prob": 0.22802734375,
            "words": [
                {
                    "start": 21.48,
                    "end": 22.16,
                    "word": " an",
                    "probability": 0.2493896484375
                },
                {
                    "start": 22.16,
                    "end": 22.28,
                    "word": " all",
                    "probability": 0.88037109375
                },
                {
                    "start": 22.28,
                    "end": 22.52,
                    "word": "-around",
                    "probability": 0.682373046875
                },
                {
                    "start": 22.52,
                    "end": 23.0,
                    "word": " awesome",
                    "probability": 0.8427734375
                },
                {
                    "start": 23.0,
                    "end": 23.34,
                    "word": " guy",
                    "probability": 0.80859375
                },
                {
                    "start": 23.34,
                    "end": 24.02,
                    "word": ".",
                    "probability": 0.60009765625
                },
                {
                    "start": 24.18,
                    "end": 24.54,
                    "word": " He",
                    "probability": 0.39453125
                },
                {
                    "start": 24.54,
                    "end": 24.68,
                    "word": " does",
                    "probability": 0.92236328125
                },
                {
                    "start": 24.68,
                    "end": 25.02,
                    "word": " way",
                    "probability": 0.8798828125
                },
                {
                    "start": 25.02,
                    "end": 25.18,
                    "word": " too",
                    "probability": 0.81982421875
                },
                {
                    "start": 25.18,
                    "end": 25.44,
                    "word": " much",
                    "probability": 0.81689453125
                },
                {
                    "start": 25.44,
                    "end": 25.74,
                    "word": " stuff",
                    "probability": 0.77734375
                },
                {
                    "start": 25.74,
                    "end": 26.08,
                    "word": " and",
                    "probability": 0.83056640625
                },
                {
                    "start": 26.08,
                    "end": 26.28,
                    "word": " is",
                    "probability": 0.8837890625
                },
                {
                    "start": 26.28,
                    "end": 26.66,
                    "word": " very",
                    "probability": 0.7333984375
                },
                {
                    "start": 26.66,
                    "end": 27.18,
                    "word": " enthusiastic",
                    "probability": 0.71435546875
                },
                {
                    "start": 27.18,
                    "end": 27.58,
                    "word": " and",
                    "probability": 0.89013671875
                },
                {
                    "start": 27.58,
                    "end": 27.94,
                    "word": " involved",
                    "probability": 0.7822265625
                },
                {
                    "start": 27.94,
                    "end": 28.08,
                    "word": " in",
                    "probability": 0.8994140625
                }
            ]
        },
        {
            "id": 4,
            "seek": 5598,
            "start": 28.08,
            "end": 34.16,
            "text": " the community. So I'm just going to let him talk about... on his intro talk about...",
            "tokens": [
                50364,
                264,
                1768,
                13,
                407,
                286,
                478,
                445,
                516,
                281,
                718,
                796,
                751,
                466,
                485,
                322,
                702,
                12897,
                751,
                466,
                485,
                50668
            ],
            "temperature": 0.0,
            "avg_logprob": -0.29166665607028536,
            "compression_ratio": 1.5862068965517242,
            "no_speech_prob": 0.03564453125,
            "words": [
                {
                    "start": 28.08,
                    "end": 28.22,
                    "word": " the",
                    "probability": 0.1629638671875
                },
                {
                    "start": 28.22,
                    "end": 28.62,
                    "word": " community",
                    "probability": 0.68701171875
                },
                {
                    "start": 28.62,
                    "end": 29.3,
                    "word": ".",
                    "probability": 0.92529296875
                },
                {
                    "start": 30.12,
                    "end": 30.52,
                    "word": " So",
                    "probability": 0.53759765625
                },
                {
                    "start": 30.52,
                    "end": 30.76,
                    "word": " I'm",
                    "probability": 0.8623046875
                },
                {
                    "start": 30.76,
                    "end": 30.92,
                    "word": " just",
                    "probability": 0.85498046875
                },
                {
                    "start": 30.92,
                    "end": 31.1,
                    "word": " going",
                    "probability": 0.54833984375
                },
                {
                    "start": 31.1,
                    "end": 31.28,
                    "word": " to",
                    "probability": 0.88623046875
                },
                {
                    "start": 31.28,
                    "end": 31.48,
                    "word": " let",
                    "probability": 0.89501953125
                },
                {
                    "start": 31.48,
                    "end": 31.66,
                    "word": " him",
                    "probability": 0.8935546875
                },
                {
                    "start": 31.66,
                    "end": 32.0,
                    "word": " talk",
                    "probability": 0.7705078125
                },
                {
                    "start": 32.0,
                    "end": 32.7,
                    "word": " about...",
                    "probability": 0.3070068359375
                },
                {
                    "start": 32.7,
                    "end": 33.02,
                    "word": " on",
                    "probability": 0.2081298828125
                },
                {
                    "start": 33.02,
                    "end": 33.22,
                    "word": " his",
                    "probability": 0.84375
                },
                {
                    "start": 33.22,
                    "end": 33.5,
                    "word": " intro",
                    "probability": 0.8603515625
                },
                {
                    "start": 33.5,
                    "end": 33.82,
                    "word": " talk",
                    "probability": 0.8203125
                },
                {
                    "start": 33.82,
                    "end": 34.16,
                    "word": " about...",
                    "probability": 0.4434814453125
                }
            ]
        },
        {
            "id": 5,
            "seek": 5598,
            "start": 34.86,
            "end": 36.72,
            "text": " for machine learning people.",
            "tokens": [
                50728,
                337,
                3479,
                2539,
                561,
                13,
                50780
            ],
            "temperature": 0.0,
            "avg_logprob": -0.29166665607028536,
            "compression_ratio": 1.5862068965517242,
            "no_speech_prob": 0.03564453125,
            "words": [
                {
                    "start": 34.86,
                    "end": 35.54,
                    "word": " for",
                    "probability": 0.8056640625
                },
                {
                    "start": 35.54,
                    "end": 35.8,
                    "word": " machine",
                    "probability": 0.72509765625
                },
                {
                    "start": 35.8,
                    "end": 36.04,
                    "word": " learning",
                    "probability": 0.81396484375
                },
                {
                    "start": 36.04,
                    "end": 36.38,
                    "word": " people",
                    "probability": 0.91845703125
                },
                {
                    "start": 36.38,
                    "end": 36.72,
                    "word": ".",
                    "probability": 0.990234375
                }
            ]
        },
        {
            "id": 6,
            "seek": 5598,
            "start": 40.93,
            "end": 46.83,
            "text": " Hi, everybody. Welcome to my talk. This is Baby's First 100 MLSat Words.",
            "tokens": [
                50864,
                2421,
                11,
                2201,
                13,
                4027,
                281,
                452,
                751,
                13,
                639,
                307,
                9425,
                311,
                2386,
                2319,
                376,
                19198,
                267,
                32857,
                13,
                51124
            ],
            "temperature": 0.0,
            "avg_logprob": -0.29166665607028536,
            "compression_ratio": 1.5862068965517242,
            "no_speech_prob": 0.03564453125,
            "words": [
                {
                    "start": 40.93,
                    "end": 41.61,
                    "word": " Hi",
                    "probability": 0.169921875
                },
                {
                    "start": 41.61,
                    "end": 41.71,
                    "word": ",",
                    "probability": 0.52978515625
                },
                {
                    "start": 41.71,
                    "end": 41.99,
                    "word": " everybody",
                    "probability": 0.7216796875
                },
                {
                    "start": 41.99,
                    "end": 42.21,
                    "word": ".",
                    "probability": 0.7470703125
                },
                {
                    "start": 42.21,
                    "end": 42.39,
                    "word": " Welcome",
                    "probability": 0.7822265625
                },
                {
                    "start": 42.39,
                    "end": 42.59,
                    "word": " to",
                    "probability": 0.9033203125
                },
                {
                    "start": 42.59,
                    "end": 42.75,
                    "word": " my",
                    "probability": 0.908203125
                },
                {
                    "start": 42.75,
                    "end": 43.05,
                    "word": " talk",
                    "probability": 0.775390625
                },
                {
                    "start": 43.05,
                    "end": 43.27,
                    "word": ".",
                    "probability": 0.97216796875
                },
                {
                    "start": 43.27,
                    "end": 43.41,
                    "word": " This",
                    "probability": 0.79443359375
                },
                {
                    "start": 43.41,
                    "end": 43.65,
                    "word": " is",
                    "probability": 0.90869140625
                },
                {
                    "start": 43.65,
                    "end": 44.19,
                    "word": " Baby's",
                    "probability": 0.68310546875
                },
                {
                    "start": 44.19,
                    "end": 44.53,
                    "word": " First",
                    "probability": 0.41357421875
                },
                {
                    "start": 44.53,
                    "end": 45.11,
                    "word": " 100",
                    "probability": 0.89990234375
                },
                {
                    "start": 45.11,
                    "end": 46.05,
                    "word": " MLSat",
                    "probability": 0.4436848958333333
                },
                {
                    "start": 46.05,
                    "end": 46.49,
                    "word": " Words",
                    "probability": 0.74658203125
                },
                {
                    "start": 46.49,
                    "end": 46.83,
                    "word": ".",
                    "probability": 0.98046875
                }
            ]
        },
        {
            "id": 7,
            "seek": 5598,
            "start": 47.21,
            "end": 52.29,
            "text": " My name is Eric Gallican. I'm an artificial intelligence researcher at Rapid7. I specialize",
            "tokens": [
                51160,
                1222,
                1315,
                307,
                9336,
                14588,
                8914,
                13,
                286,
                478,
                364,
                11677,
                7599,
                21751,
                412,
                44580,
                22,
                13,
                286,
                37938,
                51424
            ],
            "temperature": 0.0,
            "avg_logprob": -0.29166665607028536,
            "compression_ratio": 1.5862068965517242,
            "no_speech_prob": 0.03564453125,
            "words": [
                {
                    "start": 47.21,
                    "end": 47.51,
                    "word": " My",
                    "probability": 0.91259765625
                },
                {
                    "start": 47.51,
                    "end": 47.67,
                    "word": " name",
                    "probability": 0.8271484375
                },
                {
                    "start": 47.67,
                    "end": 47.81,
                    "word": " is",
                    "probability": 0.90966796875
                },
                {
                    "start": 47.81,
                    "end": 48.09,
                    "word": " Eric",
                    "probability": 0.76171875
                },
                {
                    "start": 48.09,
                    "end": 48.43,
                    "word": " Gallican",
                    "probability": 0.02685403823852539
                },
                {
                    "start": 48.43,
                    "end": 48.73,
                    "word": ".",
                    "probability": 0.89697265625
                },
                {
                    "start": 48.73,
                    "end": 48.97,
                    "word": " I'm",
                    "probability": 0.989990234375
                },
                {
                    "start": 48.97,
                    "end": 49.07,
                    "word": " an",
                    "probability": 0.91455078125
                },
                {
                    "start": 49.07,
                    "end": 49.33,
                    "word": " artificial",
                    "probability": 0.7158203125
                },
                {
                    "start": 49.33,
                    "end": 50.03,
                    "word": " intelligence",
                    "probability": 0.65673828125
                },
                {
                    "start": 50.03,
                    "end": 50.55,
                    "word": " researcher",
                    "probability": 0.75
                },
                {
                    "start": 50.55,
                    "end": 50.85,
                    "word": " at",
                    "probability": 0.8740234375
                },
                {
                    "start": 50.85,
                    "end": 51.41,
                    "word": " Rapid7",
                    "probability": 0.744140625
                },
                {
                    "start": 51.41,
                    "end": 51.67,
                    "word": ".",
                    "probability": 0.97412109375
                },
                {
                    "start": 51.67,
                    "end": 51.85,
                    "word": " I",
                    "probability": 0.99658203125
                },
                {
                    "start": 51.85,
                    "end": 52.29,
                    "word": " specialize",
                    "probability": 0.92138671875
                }
            ]
        },
        {
            "id": 8,
            "seek": 5598,
            "start": 52.29,
            "end": 59.71,
            "text": " in breaking machine learning stuff and applying machine learning to detecting bad things.",
            "tokens": [
                51424,
                294,
                7697,
                3479,
                2539,
                1507,
                293,
                9275,
                3479,
                2539,
                281,
                40237,
                1578,
                721,
                13,
                51744
            ],
            "temperature": 0.0,
            "avg_logprob": -0.29166665607028536,
            "compression_ratio": 1.5862068965517242,
            "no_speech_prob": 0.03564453125,
            "words": [
                {
                    "start": 52.29,
                    "end": 52.75,
                    "word": " in",
                    "probability": 0.90234375
                },
                {
                    "start": 52.75,
                    "end": 53.13,
                    "word": " breaking",
                    "probability": 0.7958984375
                },
                {
                    "start": 53.13,
                    "end": 53.59,
                    "word": " machine",
                    "probability": 0.7392578125
                },
                {
                    "start": 53.59,
                    "end": 53.95,
                    "word": " learning",
                    "probability": 0.83154296875
                },
                {
                    "start": 53.95,
                    "end": 54.49,
                    "word": " stuff",
                    "probability": 0.8046875
                },
                {
                    "start": 54.49,
                    "end": 55.35,
                    "word": " and",
                    "probability": 0.80859375
                },
                {
                    "start": 55.35,
                    "end": 56.01,
                    "word": " applying",
                    "probability": 0.72802734375
                },
                {
                    "start": 56.01,
                    "end": 56.55,
                    "word": " machine",
                    "probability": 0.744140625
                },
                {
                    "start": 56.55,
                    "end": 56.91,
                    "word": " learning",
                    "probability": 0.85498046875
                },
                {
                    "start": 56.91,
                    "end": 57.43,
                    "word": " to",
                    "probability": 0.9033203125
                },
                {
                    "start": 57.43,
                    "end": 57.99,
                    "word": " detecting",
                    "probability": 0.75390625
                },
                {
                    "start": 57.99,
                    "end": 58.51,
                    "word": " bad",
                    "probability": 0.87548828125
                },
                {
                    "start": 58.51,
                    "end": 59.07,
                    "word": " things",
                    "probability": 0.7255859375
                },
                {
                    "start": 59.07,
                    "end": 59.71,
                    "word": ".",
                    "probability": 0.998046875
                }
            ]
        },
        {
            "id": 9,
            "seek": 8446,
            "start": 59.71,
            "end": 65.13,
            "text": " I also do research at the Montreal AI Ethics Institute on applying DevOps principle to",
            "tokens": [
                50364,
                286,
                611,
                360,
                2132,
                412,
                264,
                34180,
                7318,
                10540,
                1167,
                9446,
                322,
                9275,
                43051,
                8665,
                281,
                50636
            ],
            "temperature": 0.0,
            "avg_logprob": -0.2058722455422957,
            "compression_ratio": 1.6043478260869566,
            "no_speech_prob": 0.00015115737915039062,
            "words": [
                {
                    "start": 59.71,
                    "end": 60.03,
                    "word": " I",
                    "probability": 0.9755859375
                },
                {
                    "start": 60.03,
                    "end": 60.29,
                    "word": " also",
                    "probability": 0.80322265625
                },
                {
                    "start": 60.29,
                    "end": 60.61,
                    "word": " do",
                    "probability": 0.9013671875
                },
                {
                    "start": 60.61,
                    "end": 61.09,
                    "word": " research",
                    "probability": 0.78759765625
                },
                {
                    "start": 61.09,
                    "end": 61.23,
                    "word": " at",
                    "probability": 0.8974609375
                },
                {
                    "start": 61.23,
                    "end": 61.35,
                    "word": " the",
                    "probability": 0.78173828125
                },
                {
                    "start": 61.35,
                    "end": 61.69,
                    "word": " Montreal",
                    "probability": 0.84765625
                },
                {
                    "start": 61.69,
                    "end": 62.05,
                    "word": " AI",
                    "probability": 0.8193359375
                },
                {
                    "start": 62.05,
                    "end": 62.45,
                    "word": " Ethics",
                    "probability": 0.923828125
                },
                {
                    "start": 62.45,
                    "end": 62.89,
                    "word": " Institute",
                    "probability": 0.8876953125
                },
                {
                    "start": 62.89,
                    "end": 63.31,
                    "word": " on",
                    "probability": 0.87255859375
                },
                {
                    "start": 63.31,
                    "end": 63.75,
                    "word": " applying",
                    "probability": 0.73193359375
                },
                {
                    "start": 63.75,
                    "end": 64.25,
                    "word": " DevOps",
                    "probability": 0.82666015625
                },
                {
                    "start": 64.25,
                    "end": 64.71,
                    "word": " principle",
                    "probability": 0.53369140625
                },
                {
                    "start": 64.71,
                    "end": 65.13,
                    "word": " to",
                    "probability": 0.890625
                }
            ]
        },
        {
            "id": 10,
            "seek": 8446,
            "start": 65.13,
            "end": 72.37,
            "text": " machine learning ethics. And last year, I was here at DEF CON in person, not here in my house",
            "tokens": [
                50636,
                3479,
                2539,
                19769,
                13,
                400,
                1036,
                1064,
                11,
                286,
                390,
                510,
                412,
                10113,
                37,
                16596,
                294,
                954,
                11,
                406,
                510,
                294,
                452,
                1782,
                51000
            ],
            "temperature": 0.0,
            "avg_logprob": -0.2058722455422957,
            "compression_ratio": 1.6043478260869566,
            "no_speech_prob": 0.00015115737915039062,
            "words": [
                {
                    "start": 65.13,
                    "end": 65.53,
                    "word": " machine",
                    "probability": 0.66259765625
                },
                {
                    "start": 65.53,
                    "end": 65.85,
                    "word": " learning",
                    "probability": 0.83544921875
                },
                {
                    "start": 65.85,
                    "end": 66.35,
                    "word": " ethics",
                    "probability": 0.8076171875
                },
                {
                    "start": 66.35,
                    "end": 66.99,
                    "word": ".",
                    "probability": 0.63818359375
                },
                {
                    "start": 67.29,
                    "end": 67.47,
                    "word": " And",
                    "probability": 0.66455078125
                },
                {
                    "start": 67.47,
                    "end": 67.85,
                    "word": " last",
                    "probability": 0.7685546875
                },
                {
                    "start": 67.85,
                    "end": 68.01,
                    "word": " year",
                    "probability": 0.88427734375
                },
                {
                    "start": 68.01,
                    "end": 68.03,
                    "word": ",",
                    "probability": 0.32568359375
                },
                {
                    "start": 68.03,
                    "end": 68.09,
                    "word": " I",
                    "probability": 0.9990234375
                },
                {
                    "start": 68.09,
                    "end": 68.21,
                    "word": " was",
                    "probability": 0.892578125
                },
                {
                    "start": 68.21,
                    "end": 68.33,
                    "word": " here",
                    "probability": 0.8095703125
                },
                {
                    "start": 68.33,
                    "end": 68.47,
                    "word": " at",
                    "probability": 0.90087890625
                },
                {
                    "start": 68.47,
                    "end": 68.65,
                    "word": " DEF",
                    "probability": 0.69140625
                },
                {
                    "start": 68.65,
                    "end": 68.95,
                    "word": " CON",
                    "probability": 0.88623046875
                },
                {
                    "start": 68.95,
                    "end": 69.69,
                    "word": " in",
                    "probability": 0.7822265625
                },
                {
                    "start": 69.69,
                    "end": 70.11,
                    "word": " person",
                    "probability": 0.71875
                },
                {
                    "start": 70.11,
                    "end": 70.41,
                    "word": ",",
                    "probability": 0.8212890625
                },
                {
                    "start": 70.41,
                    "end": 70.69,
                    "word": " not",
                    "probability": 0.892578125
                },
                {
                    "start": 70.69,
                    "end": 71.45,
                    "word": " here",
                    "probability": 0.8056640625
                },
                {
                    "start": 71.45,
                    "end": 71.73,
                    "word": " in",
                    "probability": 0.88671875
                },
                {
                    "start": 71.73,
                    "end": 71.97,
                    "word": " my",
                    "probability": 0.90966796875
                },
                {
                    "start": 71.97,
                    "end": 72.37,
                    "word": " house",
                    "probability": 0.80224609375
                }
            ]
        },
        {
            "id": 11,
            "seek": 8446,
            "start": 72.37,
            "end": 79.69,
            "text": " at DEF CON, presenting at the Cloud Village on a piece of malware that I wrote. Malware,",
            "tokens": [
                51000,
                412,
                10113,
                37,
                16596,
                11,
                15578,
                412,
                264,
                8061,
                22651,
                322,
                257,
                2522,
                295,
                40747,
                300,
                286,
                4114,
                13,
                5746,
                3039,
                11,
                51364
            ],
            "temperature": 0.0,
            "avg_logprob": -0.2058722455422957,
            "compression_ratio": 1.6043478260869566,
            "no_speech_prob": 0.00015115737915039062,
            "words": [
                {
                    "start": 72.37,
                    "end": 72.73,
                    "word": " at",
                    "probability": 0.8857421875
                },
                {
                    "start": 72.73,
                    "end": 73.07,
                    "word": " DEF",
                    "probability": 0.939453125
                },
                {
                    "start": 73.07,
                    "end": 73.35,
                    "word": " CON",
                    "probability": 0.92431640625
                },
                {
                    "start": 73.35,
                    "end": 73.97,
                    "word": ",",
                    "probability": 0.7470703125
                },
                {
                    "start": 73.97,
                    "end": 75.31,
                    "word": " presenting",
                    "probability": 0.63916015625
                },
                {
                    "start": 75.31,
                    "end": 75.51,
                    "word": " at",
                    "probability": 0.87548828125
                },
                {
                    "start": 75.51,
                    "end": 75.59,
                    "word": " the",
                    "probability": 0.75
                },
                {
                    "start": 75.59,
                    "end": 75.81,
                    "word": " Cloud",
                    "probability": 0.767578125
                },
                {
                    "start": 75.81,
                    "end": 76.29,
                    "word": " Village",
                    "probability": 0.609375
                },
                {
                    "start": 76.29,
                    "end": 76.77,
                    "word": " on",
                    "probability": 0.8896484375
                },
                {
                    "start": 76.99,
                    "end": 77.17,
                    "word": " a piece",
                    "probability": 0.814453125
                },
                {
                    "start": 77.17,
                    "end": 77.35,
                    "word": " of",
                    "probability": 0.90380859375
                },
                {
                    "start": 77.35,
                    "end": 77.67,
                    "word": " malware",
                    "probability": 0.880859375
                },
                {
                    "start": 77.67,
                    "end": 77.97,
                    "word": " that",
                    "probability": 0.87841796875
                },
                {
                    "start": 77.97,
                    "end": 78.13,
                    "word": " I",
                    "probability": 0.998046875
                },
                {
                    "start": 78.13,
                    "end": 78.39,
                    "word": " wrote",
                    "probability": 0.87890625
                },
                {
                    "start": 78.39,
                    "end": 78.73,
                    "word": ".",
                    "probability": 0.461181640625
                },
                {
                    "start": 78.73,
                    "end": 79.35,
                    "word": " Malware",
                    "probability": 0.772216796875
                },
                {
                    "start": 79.35,
                    "end": 79.69,
                    "word": ",",
                    "probability": 0.70166015625
                }
            ]
        },
        {
            "id": 12,
            "seek": 8446,
            "start": 79.69,
            "end": 88.19,
            "text": " not actually malware, just in case the legal team is listening. And I presented that piece of proof",
            "tokens": [
                51364,
                406,
                767,
                40747,
                11,
                445,
                294,
                1389,
                264,
                5089,
                1469,
                307,
                4764,
                13,
                400,
                286,
                8212,
                300,
                2522,
                295,
                8177,
                51784
            ],
            "temperature": 0.0,
            "avg_logprob": -0.2058722455422957,
            "compression_ratio": 1.6043478260869566,
            "no_speech_prob": 0.00015115737915039062,
            "words": [
                {
                    "start": 79.69,
                    "end": 80.11,
                    "word": " not",
                    "probability": 0.89111328125
                },
                {
                    "start": 80.11,
                    "end": 80.77,
                    "word": " actually",
                    "probability": 0.75439453125
                },
                {
                    "start": 80.77,
                    "end": 81.31,
                    "word": " malware",
                    "probability": 0.86376953125
                },
                {
                    "start": 81.31,
                    "end": 82.09,
                    "word": ",",
                    "probability": 0.5986328125
                },
                {
                    "start": 82.09,
                    "end": 82.93,
                    "word": " just",
                    "probability": 0.8759765625
                },
                {
                    "start": 82.93,
                    "end": 83.11,
                    "word": " in",
                    "probability": 0.90625
                },
                {
                    "start": 83.11,
                    "end": 83.35,
                    "word": " case",
                    "probability": 0.81103515625
                },
                {
                    "start": 83.35,
                    "end": 83.59,
                    "word": " the",
                    "probability": 0.81005859375
                },
                {
                    "start": 83.59,
                    "end": 83.87,
                    "word": " legal",
                    "probability": 0.8291015625
                },
                {
                    "start": 83.87,
                    "end": 84.13,
                    "word": " team",
                    "probability": 0.92236328125
                },
                {
                    "start": 84.13,
                    "end": 84.29,
                    "word": " is",
                    "probability": 0.89892578125
                },
                {
                    "start": 84.29,
                    "end": 84.57,
                    "word": " listening",
                    "probability": 0.7158203125
                },
                {
                    "start": 84.57,
                    "end": 85.21,
                    "word": ".",
                    "probability": 0.93896484375
                },
                {
                    "start": 85.83,
                    "end": 86.47,
                    "word": " And",
                    "probability": 0.8662109375
                },
                {
                    "start": 86.47,
                    "end": 86.61,
                    "word": " I",
                    "probability": 0.99365234375
                },
                {
                    "start": 86.61,
                    "end": 86.89,
                    "word": " presented",
                    "probability": 0.71337890625
                },
                {
                    "start": 86.89,
                    "end": 87.23,
                    "word": " that",
                    "probability": 0.87939453125
                },
                {
                    "start": 87.23,
                    "end": 87.53,
                    "word": " piece",
                    "probability": 0.81396484375
                },
                {
                    "start": 87.53,
                    "end": 87.81,
                    "word": " of",
                    "probability": 0.9033203125
                },
                {
                    "start": 87.81,
                    "end": 88.19,
                    "word": " proof",
                    "probability": 0.8193359375
                }
            ]
        },
        {
            "id": 13,
            "seek": 11322,
            "start": 88.19,
            "end": 95.83,
            "text": " of concept malware called Sassy Boy. But mostly I do boring math stuff. And if you can't see my",
            "tokens": [
                50364,
                295,
                3410,
                40747,
                1219,
                318,
                20712,
                9486,
                13,
                583,
                5240,
                286,
                360,
                9989,
                5221,
                1507,
                13,
                400,
                498,
                291,
                393,
                380,
                536,
                452,
                50744
            ],
            "temperature": 0.0,
            "avg_logprob": -0.19430693718466427,
            "compression_ratio": 1.5368852459016393,
            "no_speech_prob": 0.00035691261291503906,
            "words": [
                {
                    "start": 88.19,
                    "end": 88.35,
                    "word": " of",
                    "probability": 0.59033203125
                },
                {
                    "start": 88.35,
                    "end": 88.89,
                    "word": " concept",
                    "probability": 0.495849609375
                },
                {
                    "start": 88.89,
                    "end": 89.39,
                    "word": " malware",
                    "probability": 0.779296875
                },
                {
                    "start": 89.39,
                    "end": 90.17,
                    "word": " called",
                    "probability": 0.7197265625
                },
                {
                    "start": 90.17,
                    "end": 90.57,
                    "word": " Sassy",
                    "probability": 0.76318359375
                },
                {
                    "start": 90.57,
                    "end": 90.81,
                    "word": " Boy",
                    "probability": 0.59716796875
                },
                {
                    "start": 90.81,
                    "end": 91.29,
                    "word": ".",
                    "probability": 0.461669921875
                },
                {
                    "start": 91.87,
                    "end": 92.07,
                    "word": " But",
                    "probability": 0.83642578125
                },
                {
                    "start": 92.07,
                    "end": 92.31,
                    "word": " mostly",
                    "probability": 0.80078125
                },
                {
                    "start": 92.31,
                    "end": 92.49,
                    "word": " I",
                    "probability": 0.861328125
                },
                {
                    "start": 92.49,
                    "end": 92.61,
                    "word": " do",
                    "probability": 0.916015625
                },
                {
                    "start": 92.61,
                    "end": 92.95,
                    "word": " boring",
                    "probability": 0.82421875
                },
                {
                    "start": 92.95,
                    "end": 93.29,
                    "word": " math",
                    "probability": 0.8955078125
                },
                {
                    "start": 93.29,
                    "end": 93.67,
                    "word": " stuff",
                    "probability": 0.78515625
                },
                {
                    "start": 93.67,
                    "end": 94.15,
                    "word": ".",
                    "probability": 0.724609375
                },
                {
                    "start": 94.45,
                    "end": 94.61,
                    "word": " And",
                    "probability": 0.78955078125
                },
                {
                    "start": 94.61,
                    "end": 94.95,
                    "word": " if",
                    "probability": 0.8828125
                },
                {
                    "start": 94.95,
                    "end": 95.15,
                    "word": " you",
                    "probability": 0.8994140625
                },
                {
                    "start": 95.15,
                    "end": 95.49,
                    "word": " can't",
                    "probability": 0.949462890625
                },
                {
                    "start": 95.49,
                    "end": 95.63,
                    "word": " see",
                    "probability": 0.81787109375
                },
                {
                    "start": 95.63,
                    "end": 95.83,
                    "word": " my",
                    "probability": 0.90576171875
                }
            ]
        },
        {
            "id": 14,
            "seek": 11322,
            "start": 95.83,
            "end": 104.03,
            "text": " speaker video for some reason, I put a picture of my face on the right so that you can see it. Or",
            "tokens": [
                50744,
                8145,
                960,
                337,
                512,
                1778,
                11,
                286,
                829,
                257,
                3036,
                295,
                452,
                1851,
                322,
                264,
                558,
                370,
                300,
                291,
                393,
                536,
                309,
                13,
                1610,
                51152
            ],
            "temperature": 0.0,
            "avg_logprob": -0.19430693718466427,
            "compression_ratio": 1.5368852459016393,
            "no_speech_prob": 0.00035691261291503906,
            "words": [
                {
                    "start": 95.83,
                    "end": 96.17,
                    "word": " speaker",
                    "probability": 0.72998046875
                },
                {
                    "start": 96.17,
                    "end": 96.61,
                    "word": " video",
                    "probability": 0.7958984375
                },
                {
                    "start": 96.61,
                    "end": 96.99,
                    "word": " for",
                    "probability": 0.7197265625
                },
                {
                    "start": 96.99,
                    "end": 97.53,
                    "word": " some",
                    "probability": 0.81494140625
                },
                {
                    "start": 97.53,
                    "end": 98.13,
                    "word": " reason",
                    "probability": 0.90966796875
                },
                {
                    "start": 98.13,
                    "end": 98.69,
                    "word": ",",
                    "probability": 0.88671875
                },
                {
                    "start": 98.69,
                    "end": 99.13,
                    "word": " I",
                    "probability": 0.982421875
                },
                {
                    "start": 99.13,
                    "end": 99.35,
                    "word": " put",
                    "probability": 0.7216796875
                },
                {
                    "start": 99.47,
                    "end": 99.69,
                    "word": " a picture",
                    "probability": 0.779296875
                },
                {
                    "start": 99.69,
                    "end": 99.87,
                    "word": " of",
                    "probability": 0.9052734375
                },
                {
                    "start": 99.87,
                    "end": 100.01,
                    "word": " my",
                    "probability": 0.90087890625
                },
                {
                    "start": 100.01,
                    "end": 100.41,
                    "word": " face",
                    "probability": 0.81982421875
                },
                {
                    "start": 100.41,
                    "end": 100.87,
                    "word": " on",
                    "probability": 0.900390625
                },
                {
                    "start": 100.87,
                    "end": 101.01,
                    "word": " the",
                    "probability": 0.8134765625
                },
                {
                    "start": 101.01,
                    "end": 101.25,
                    "word": " right",
                    "probability": 0.8740234375
                },
                {
                    "start": 101.25,
                    "end": 101.81,
                    "word": " so",
                    "probability": 0.685546875
                },
                {
                    "start": 101.81,
                    "end": 102.03,
                    "word": " that",
                    "probability": 0.83251953125
                },
                {
                    "start": 102.03,
                    "end": 102.17,
                    "word": " you",
                    "probability": 0.8974609375
                },
                {
                    "start": 102.17,
                    "end": 102.41,
                    "word": " can",
                    "probability": 0.8984375
                },
                {
                    "start": 102.41,
                    "end": 103.31,
                    "word": " see",
                    "probability": 0.81591796875
                },
                {
                    "start": 103.31,
                    "end": 103.53,
                    "word": " it",
                    "probability": 0.892578125
                },
                {
                    "start": 103.53,
                    "end": 103.81,
                    "word": ".",
                    "probability": 0.302490234375
                },
                {
                    "start": 103.81,
                    "end": 104.03,
                    "word": " Or",
                    "probability": 0.77001953125
                }
            ]
        },
        {
            "id": 15,
            "seek": 11322,
            "start": 104.03,
            "end": 111.15,
            "text": " you can see it twice. You can see that my hair hasn't changed basically at all in years. So",
            "tokens": [
                51152,
                291,
                393,
                536,
                309,
                6091,
                13,
                509,
                393,
                536,
                300,
                452,
                2578,
                6132,
                380,
                3105,
                1936,
                412,
                439,
                294,
                924,
                13,
                407,
                51524
            ],
            "temperature": 0.0,
            "avg_logprob": -0.19430693718466427,
            "compression_ratio": 1.5368852459016393,
            "no_speech_prob": 0.00035691261291503906,
            "words": [
                {
                    "start": 104.03,
                    "end": 104.13,
                    "word": " you",
                    "probability": 0.859375
                },
                {
                    "start": 104.13,
                    "end": 104.23,
                    "word": " can",
                    "probability": 0.89501953125
                },
                {
                    "start": 104.23,
                    "end": 104.35,
                    "word": " see",
                    "probability": 0.80517578125
                },
                {
                    "start": 104.35,
                    "end": 104.45,
                    "word": " it",
                    "probability": 0.9013671875
                },
                {
                    "start": 104.45,
                    "end": 104.79,
                    "word": " twice",
                    "probability": 0.8408203125
                },
                {
                    "start": 104.79,
                    "end": 105.21,
                    "word": ".",
                    "probability": 0.748046875
                },
                {
                    "start": 105.21,
                    "end": 105.33,
                    "word": " You",
                    "probability": 0.77197265625
                },
                {
                    "start": 105.33,
                    "end": 105.45,
                    "word": " can",
                    "probability": 0.89501953125
                },
                {
                    "start": 105.45,
                    "end": 105.57,
                    "word": " see",
                    "probability": 0.79931640625
                },
                {
                    "start": 105.57,
                    "end": 105.69,
                    "word": " that",
                    "probability": 0.86328125
                },
                {
                    "start": 105.69,
                    "end": 105.87,
                    "word": " my",
                    "probability": 0.89892578125
                },
                {
                    "start": 105.87,
                    "end": 106.03,
                    "word": " hair",
                    "probability": 0.8740234375
                },
                {
                    "start": 106.03,
                    "end": 106.35,
                    "word": " hasn't",
                    "probability": 0.92333984375
                },
                {
                    "start": 106.35,
                    "end": 106.81,
                    "word": " changed",
                    "probability": 0.78076171875
                },
                {
                    "start": 106.81,
                    "end": 109.25,
                    "word": " basically",
                    "probability": 0.61669921875
                },
                {
                    "start": 109.25,
                    "end": 109.43,
                    "word": " at",
                    "probability": 0.88671875
                },
                {
                    "start": 109.43,
                    "end": 109.57,
                    "word": " all",
                    "probability": 0.8984375
                },
                {
                    "start": 109.57,
                    "end": 109.73,
                    "word": " in",
                    "probability": 0.904296875
                },
                {
                    "start": 109.73,
                    "end": 110.07,
                    "word": " years",
                    "probability": 0.86669921875
                },
                {
                    "start": 110.07,
                    "end": 110.55,
                    "word": ".",
                    "probability": 0.99560546875
                },
                {
                    "start": 110.89,
                    "end": 111.15,
                    "word": " So",
                    "probability": 0.85546875
                }
            ]
        },
        {
            "id": 16,
            "seek": 11322,
            "start": 111.15,
            "end": 116.95,
            "text": " before we get started, let's just answer the question, what is MLSat? Because I get asked",
            "tokens": [
                51524,
                949,
                321,
                483,
                1409,
                11,
                718,
                311,
                445,
                1867,
                264,
                1168,
                11,
                437,
                307,
                376,
                19198,
                267,
                30,
                1436,
                286,
                483,
                2351,
                51804
            ],
            "temperature": 0.0,
            "avg_logprob": -0.19430693718466427,
            "compression_ratio": 1.5368852459016393,
            "no_speech_prob": 0.00035691261291503906,
            "words": [
                {
                    "start": 111.15,
                    "end": 111.69,
                    "word": " before",
                    "probability": 0.59130859375
                },
                {
                    "start": 111.69,
                    "end": 111.89,
                    "word": " we",
                    "probability": 0.9013671875
                },
                {
                    "start": 111.89,
                    "end": 112.03,
                    "word": " get",
                    "probability": 0.9013671875
                },
                {
                    "start": 112.03,
                    "end": 112.39,
                    "word": " started",
                    "probability": 0.798828125
                },
                {
                    "start": 112.39,
                    "end": 112.57,
                    "word": ",",
                    "probability": 0.8916015625
                },
                {
                    "start": 112.57,
                    "end": 112.85,
                    "word": " let's",
                    "probability": 0.94921875
                },
                {
                    "start": 112.85,
                    "end": 113.33,
                    "word": " just",
                    "probability": 0.88916015625
                },
                {
                    "start": 113.33,
                    "end": 113.57,
                    "word": " answer",
                    "probability": 0.88623046875
                },
                {
                    "start": 113.57,
                    "end": 113.73,
                    "word": " the",
                    "probability": 0.8154296875
                },
                {
                    "start": 113.73,
                    "end": 114.01,
                    "word": " question",
                    "probability": 0.80078125
                },
                {
                    "start": 114.01,
                    "end": 114.15,
                    "word": ",",
                    "probability": 0.66796875
                },
                {
                    "start": 114.15,
                    "end": 114.31,
                    "word": " what",
                    "probability": 0.58837890625
                },
                {
                    "start": 114.31,
                    "end": 114.65,
                    "word": " is",
                    "probability": 0.90380859375
                },
                {
                    "start": 114.65,
                    "end": 115.47,
                    "word": " MLSat",
                    "probability": 0.3767695426940918
                },
                {
                    "start": 115.47,
                    "end": 115.67,
                    "word": "?",
                    "probability": 0.8115234375
                },
                {
                    "start": 115.69,
                    "end": 116.17,
                    "word": " Because",
                    "probability": 0.68359375
                },
                {
                    "start": 116.17,
                    "end": 116.35,
                    "word": " I",
                    "probability": 0.998046875
                },
                {
                    "start": 116.35,
                    "end": 116.57,
                    "word": " get",
                    "probability": 0.90380859375
                },
                {
                    "start": 116.57,
                    "end": 116.95,
                    "word": " asked",
                    "probability": 0.73974609375
                }
            ]
        },
        {
            "id": 17,
            "seek": 14084,
            "start": 116.95,
            "end": 123.97,
            "text": " to that a lot by practitioners in security and in machine learning, honestly. And mostly it's",
            "tokens": [
                50364,
                281,
                300,
                257,
                688,
                538,
                25742,
                294,
                3825,
                293,
                294,
                3479,
                2539,
                11,
                6095,
                13,
                400,
                5240,
                309,
                311,
                50712
            ],
            "temperature": 0.0,
            "avg_logprob": -0.187123492180583,
            "compression_ratio": 1.7442922374429224,
            "no_speech_prob": 0.00154876708984375,
            "words": [
                {
                    "start": 116.95,
                    "end": 117.11,
                    "word": " to",
                    "probability": 0.314453125
                },
                {
                    "start": 117.11,
                    "end": 117.37,
                    "word": " that",
                    "probability": 0.82470703125
                },
                {
                    "start": 117.65,
                    "end": 117.93,
                    "word": " a lot",
                    "probability": 0.896484375
                },
                {
                    "start": 117.93,
                    "end": 118.33,
                    "word": " by",
                    "probability": 0.84033203125
                },
                {
                    "start": 118.33,
                    "end": 119.13,
                    "word": " practitioners",
                    "probability": 0.76416015625
                },
                {
                    "start": 119.13,
                    "end": 119.69,
                    "word": " in",
                    "probability": 0.89453125
                },
                {
                    "start": 119.69,
                    "end": 120.21,
                    "word": " security",
                    "probability": 0.9228515625
                },
                {
                    "start": 120.21,
                    "end": 120.89,
                    "word": " and",
                    "probability": 0.865234375
                },
                {
                    "start": 120.89,
                    "end": 121.25,
                    "word": " in",
                    "probability": 0.86279296875
                },
                {
                    "start": 121.25,
                    "end": 121.51,
                    "word": " machine",
                    "probability": 0.6953125
                },
                {
                    "start": 121.51,
                    "end": 121.89,
                    "word": " learning",
                    "probability": 0.84814453125
                },
                {
                    "start": 121.89,
                    "end": 122.07,
                    "word": ",",
                    "probability": 0.7177734375
                },
                {
                    "start": 122.07,
                    "end": 122.33,
                    "word": " honestly",
                    "probability": 0.77490234375
                },
                {
                    "start": 122.33,
                    "end": 123.09,
                    "word": ".",
                    "probability": 0.78173828125
                },
                {
                    "start": 123.25,
                    "end": 123.41,
                    "word": " And",
                    "probability": 0.7822265625
                },
                {
                    "start": 123.41,
                    "end": 123.73,
                    "word": " mostly",
                    "probability": 0.81689453125
                },
                {
                    "start": 123.73,
                    "end": 123.97,
                    "word": " it's",
                    "probability": 0.8818359375
                }
            ]
        },
        {
            "id": 18,
            "seek": 14084,
            "start": 123.97,
            "end": 131.39,
            "text": " an argument. And it's an argument between three groups of people, as it usually is. And the first",
            "tokens": [
                50712,
                364,
                6770,
                13,
                400,
                309,
                311,
                364,
                6770,
                1296,
                1045,
                3935,
                295,
                561,
                11,
                382,
                309,
                2673,
                307,
                13,
                400,
                264,
                700,
                51084
            ],
            "temperature": 0.0,
            "avg_logprob": -0.187123492180583,
            "compression_ratio": 1.7442922374429224,
            "no_speech_prob": 0.00154876708984375,
            "words": [
                {
                    "start": 123.97,
                    "end": 124.13,
                    "word": " an",
                    "probability": 0.9130859375
                },
                {
                    "start": 124.13,
                    "end": 124.53,
                    "word": " argument",
                    "probability": 0.8486328125
                },
                {
                    "start": 124.53,
                    "end": 124.85,
                    "word": ".",
                    "probability": 0.435546875
                },
                {
                    "start": 125.15,
                    "end": 125.91,
                    "word": " And",
                    "probability": 0.5703125
                },
                {
                    "start": 125.91,
                    "end": 126.05,
                    "word": " it's",
                    "probability": 0.94921875
                },
                {
                    "start": 126.05,
                    "end": 126.19,
                    "word": " an",
                    "probability": 0.9111328125
                },
                {
                    "start": 126.19,
                    "end": 126.43,
                    "word": " argument",
                    "probability": 0.83740234375
                },
                {
                    "start": 126.43,
                    "end": 126.87,
                    "word": " between",
                    "probability": 0.78125
                },
                {
                    "start": 126.87,
                    "end": 127.33,
                    "word": " three",
                    "probability": 0.8427734375
                },
                {
                    "start": 127.33,
                    "end": 127.71,
                    "word": " groups",
                    "probability": 0.77392578125
                },
                {
                    "start": 127.71,
                    "end": 127.93,
                    "word": " of",
                    "probability": 0.9072265625
                },
                {
                    "start": 127.93,
                    "end": 128.31,
                    "word": " people",
                    "probability": 0.9296875
                },
                {
                    "start": 128.31,
                    "end": 128.59,
                    "word": ",",
                    "probability": 0.58935546875
                },
                {
                    "start": 128.59,
                    "end": 129.41,
                    "word": " as",
                    "probability": 0.8994140625
                },
                {
                    "start": 129.41,
                    "end": 129.55,
                    "word": " it",
                    "probability": 0.90087890625
                },
                {
                    "start": 129.55,
                    "end": 129.95,
                    "word": " usually",
                    "probability": 0.83251953125
                },
                {
                    "start": 129.95,
                    "end": 130.51,
                    "word": " is",
                    "probability": 0.9033203125
                },
                {
                    "start": 130.51,
                    "end": 130.89,
                    "word": ".",
                    "probability": 0.94677734375
                },
                {
                    "start": 130.89,
                    "end": 131.05,
                    "word": " And",
                    "probability": 0.8662109375
                },
                {
                    "start": 131.05,
                    "end": 131.17,
                    "word": " the",
                    "probability": 0.81494140625
                },
                {
                    "start": 131.17,
                    "end": 131.39,
                    "word": " first",
                    "probability": 0.814453125
                }
            ]
        },
        {
            "id": 19,
            "seek": 14084,
            "start": 131.39,
            "end": 137.21,
            "text": " group of people is people who think it's only applications of machine learning and artificial",
            "tokens": [
                51084,
                1594,
                295,
                561,
                307,
                561,
                567,
                519,
                309,
                311,
                787,
                5821,
                295,
                3479,
                2539,
                293,
                11677,
                51380
            ],
            "temperature": 0.0,
            "avg_logprob": -0.187123492180583,
            "compression_ratio": 1.7442922374429224,
            "no_speech_prob": 0.00154876708984375,
            "words": [
                {
                    "start": 131.39,
                    "end": 131.63,
                    "word": " group",
                    "probability": 0.85546875
                },
                {
                    "start": 131.63,
                    "end": 131.79,
                    "word": " of",
                    "probability": 0.9033203125
                },
                {
                    "start": 131.79,
                    "end": 132.11,
                    "word": " people",
                    "probability": 0.9296875
                },
                {
                    "start": 132.11,
                    "end": 132.75,
                    "word": " is",
                    "probability": 0.8603515625
                },
                {
                    "start": 132.75,
                    "end": 133.05,
                    "word": " people",
                    "probability": 0.90966796875
                },
                {
                    "start": 133.05,
                    "end": 133.23,
                    "word": " who",
                    "probability": 0.81005859375
                },
                {
                    "start": 133.23,
                    "end": 133.43,
                    "word": " think",
                    "probability": 0.82177734375
                },
                {
                    "start": 133.43,
                    "end": 133.71,
                    "word": " it's",
                    "probability": 0.946044921875
                },
                {
                    "start": 133.71,
                    "end": 134.17,
                    "word": " only",
                    "probability": 0.8583984375
                },
                {
                    "start": 134.17,
                    "end": 134.99,
                    "word": " applications",
                    "probability": 0.73095703125
                },
                {
                    "start": 134.99,
                    "end": 135.49,
                    "word": " of",
                    "probability": 0.90234375
                },
                {
                    "start": 135.49,
                    "end": 135.95,
                    "word": " machine",
                    "probability": 0.6953125
                },
                {
                    "start": 135.95,
                    "end": 136.31,
                    "word": " learning",
                    "probability": 0.84814453125
                },
                {
                    "start": 136.31,
                    "end": 136.77,
                    "word": " and",
                    "probability": 0.88720703125
                },
                {
                    "start": 136.77,
                    "end": 137.21,
                    "word": " artificial",
                    "probability": 0.81982421875
                }
            ]
        },
        {
            "id": 20,
            "seek": 14084,
            "start": 137.21,
            "end": 144.57,
            "text": " intelligence to security problems. So that's automated malware detection. That's next generation",
            "tokens": [
                51380,
                7599,
                281,
                3825,
                2740,
                13,
                407,
                300,
                311,
                18473,
                40747,
                17784,
                13,
                663,
                311,
                958,
                5125,
                51760
            ],
            "temperature": 0.0,
            "avg_logprob": -0.187123492180583,
            "compression_ratio": 1.7442922374429224,
            "no_speech_prob": 0.00154876708984375,
            "words": [
                {
                    "start": 137.21,
                    "end": 137.85,
                    "word": " intelligence",
                    "probability": 0.69775390625
                },
                {
                    "start": 137.85,
                    "end": 138.27,
                    "word": " to",
                    "probability": 0.8740234375
                },
                {
                    "start": 138.27,
                    "end": 138.75,
                    "word": " security",
                    "probability": 0.923828125
                },
                {
                    "start": 138.75,
                    "end": 139.35,
                    "word": " problems",
                    "probability": 0.68408203125
                },
                {
                    "start": 139.35,
                    "end": 139.67,
                    "word": ".",
                    "probability": 0.9599609375
                },
                {
                    "start": 139.67,
                    "end": 139.87,
                    "word": " So",
                    "probability": 0.88037109375
                },
                {
                    "start": 139.87,
                    "end": 140.29,
                    "word": " that's",
                    "probability": 0.896240234375
                },
                {
                    "start": 140.29,
                    "end": 140.79,
                    "word": " automated",
                    "probability": 0.8720703125
                },
                {
                    "start": 140.79,
                    "end": 141.21,
                    "word": " malware",
                    "probability": 0.8037109375
                },
                {
                    "start": 141.21,
                    "end": 141.75,
                    "word": " detection",
                    "probability": 0.74169921875
                },
                {
                    "start": 141.75,
                    "end": 142.23,
                    "word": ".",
                    "probability": 0.53564453125
                },
                {
                    "start": 142.23,
                    "end": 142.67,
                    "word": " That's",
                    "probability": 0.9033203125
                },
                {
                    "start": 142.67,
                    "end": 143.87,
                    "word": " next",
                    "probability": 0.80419921875
                },
                {
                    "start": 143.87,
                    "end": 144.57,
                    "word": " generation",
                    "probability": 0.62109375
                }
            ]
        },
        {
            "id": 21,
            "seek": 16252,
            "start": 144.57,
            "end": 151.73,
            "text": " SIMs. That's automated red teaming and that sort of stuff. And then there are people who think that",
            "tokens": [
                50380,
                24738,
                82,
                13,
                663,
                311,
                18473,
                2182,
                1469,
                278,
                293,
                300,
                1333,
                295,
                1507,
                13,
                400,
                550,
                456,
                366,
                561,
                567,
                519,
                300,
                50720
            ],
            "temperature": 0.0,
            "avg_logprob": -0.2260890117648876,
            "compression_ratio": 1.6264367816091954,
            "no_speech_prob": 0.00130462646484375,
            "words": [
                {
                    "start": 144.57,
                    "end": 145.63,
                    "word": " SIMs",
                    "probability": 0.19849395751953125
                },
                {
                    "start": 145.63,
                    "end": 145.83,
                    "word": ".",
                    "probability": 0.19140625
                },
                {
                    "start": 145.83,
                    "end": 146.27,
                    "word": " That's",
                    "probability": 0.87890625
                },
                {
                    "start": 146.27,
                    "end": 147.21,
                    "word": " automated",
                    "probability": 0.77880859375
                },
                {
                    "start": 147.21,
                    "end": 147.63,
                    "word": " red",
                    "probability": 0.6181640625
                },
                {
                    "start": 147.63,
                    "end": 148.23,
                    "word": " teaming",
                    "probability": 0.93505859375
                },
                {
                    "start": 148.23,
                    "end": 148.39,
                    "word": " and",
                    "probability": 0.75537109375
                },
                {
                    "start": 148.39,
                    "end": 148.57,
                    "word": " that",
                    "probability": 0.88134765625
                },
                {
                    "start": 148.57,
                    "end": 148.79,
                    "word": " sort",
                    "probability": 0.81884765625
                },
                {
                    "start": 148.79,
                    "end": 148.93,
                    "word": " of",
                    "probability": 0.91162109375
                },
                {
                    "start": 148.93,
                    "end": 149.23,
                    "word": " stuff",
                    "probability": 0.77978515625
                },
                {
                    "start": 149.23,
                    "end": 149.77,
                    "word": ".",
                    "probability": 0.96240234375
                },
                {
                    "start": 150.49,
                    "end": 150.69,
                    "word": " And",
                    "probability": 0.81787109375
                },
                {
                    "start": 150.69,
                    "end": 150.79,
                    "word": " then",
                    "probability": 0.7177734375
                },
                {
                    "start": 150.79,
                    "end": 150.93,
                    "word": " there",
                    "probability": 0.83447265625
                },
                {
                    "start": 150.93,
                    "end": 151.01,
                    "word": " are",
                    "probability": 0.8857421875
                },
                {
                    "start": 151.01,
                    "end": 151.23,
                    "word": " people",
                    "probability": 0.923828125
                },
                {
                    "start": 151.23,
                    "end": 151.41,
                    "word": " who",
                    "probability": 0.80615234375
                },
                {
                    "start": 151.41,
                    "end": 151.59,
                    "word": " think",
                    "probability": 0.81103515625
                },
                {
                    "start": 151.59,
                    "end": 151.73,
                    "word": " that",
                    "probability": 0.84619140625
                }
            ]
        },
        {
            "id": 22,
            "seek": 16252,
            "start": 151.73,
            "end": 158.79,
            "text": " it's just securing artificial intelligence in machine learning systems. And then there are",
            "tokens": [
                50720,
                309,
                311,
                445,
                33640,
                11677,
                7599,
                294,
                3479,
                2539,
                3652,
                13,
                400,
                550,
                456,
                366,
                51072
            ],
            "temperature": 0.0,
            "avg_logprob": -0.2260890117648876,
            "compression_ratio": 1.6264367816091954,
            "no_speech_prob": 0.00130462646484375,
            "words": [
                {
                    "start": 151.73,
                    "end": 151.99,
                    "word": " it's",
                    "probability": 0.948486328125
                },
                {
                    "start": 151.99,
                    "end": 152.73,
                    "word": " just",
                    "probability": 0.87353515625
                },
                {
                    "start": 152.73,
                    "end": 153.65,
                    "word": " securing",
                    "probability": 0.89453125
                },
                {
                    "start": 153.65,
                    "end": 154.17,
                    "word": " artificial",
                    "probability": 0.79931640625
                },
                {
                    "start": 154.17,
                    "end": 154.95,
                    "word": " intelligence",
                    "probability": 0.71337890625
                },
                {
                    "start": 154.95,
                    "end": 155.19,
                    "word": " in",
                    "probability": 0.2032470703125
                },
                {
                    "start": 155.19,
                    "end": 155.51,
                    "word": " machine",
                    "probability": 0.70166015625
                },
                {
                    "start": 155.51,
                    "end": 155.85,
                    "word": " learning",
                    "probability": 0.83447265625
                },
                {
                    "start": 155.85,
                    "end": 156.49,
                    "word": " systems",
                    "probability": 0.8017578125
                },
                {
                    "start": 156.49,
                    "end": 157.03,
                    "word": ".",
                    "probability": 0.90380859375
                },
                {
                    "start": 157.69,
                    "end": 158.23,
                    "word": " And",
                    "probability": 0.85205078125
                },
                {
                    "start": 158.23,
                    "end": 158.39,
                    "word": " then",
                    "probability": 0.72412109375
                },
                {
                    "start": 158.39,
                    "end": 158.61,
                    "word": " there",
                    "probability": 0.84375
                },
                {
                    "start": 158.61,
                    "end": 158.79,
                    "word": " are",
                    "probability": 0.88818359375
                }
            ]
        },
        {
            "id": 23,
            "seek": 16252,
            "start": 158.79,
            "end": 166.25,
            "text": " those of us who are correct who think it's both. It's kind of any intersection of artificial",
            "tokens": [
                51072,
                729,
                295,
                505,
                567,
                366,
                3006,
                567,
                519,
                309,
                311,
                1293,
                13,
                467,
                311,
                733,
                295,
                604,
                15236,
                295,
                11677,
                51452
            ],
            "temperature": 0.0,
            "avg_logprob": -0.2260890117648876,
            "compression_ratio": 1.6264367816091954,
            "no_speech_prob": 0.00130462646484375,
            "words": [
                {
                    "start": 158.79,
                    "end": 159.05,
                    "word": " those",
                    "probability": 0.853515625
                },
                {
                    "start": 159.05,
                    "end": 159.23,
                    "word": " of",
                    "probability": 0.9033203125
                },
                {
                    "start": 159.23,
                    "end": 159.47,
                    "word": " us",
                    "probability": 0.904296875
                },
                {
                    "start": 159.47,
                    "end": 159.67,
                    "word": " who",
                    "probability": 0.81884765625
                },
                {
                    "start": 159.67,
                    "end": 159.89,
                    "word": " are",
                    "probability": 0.8935546875
                },
                {
                    "start": 159.89,
                    "end": 160.57,
                    "word": " correct",
                    "probability": 0.875
                },
                {
                    "start": 160.57,
                    "end": 161.25,
                    "word": " who",
                    "probability": 0.390380859375
                },
                {
                    "start": 161.25,
                    "end": 161.53,
                    "word": " think",
                    "probability": 0.80908203125
                },
                {
                    "start": 161.53,
                    "end": 161.77,
                    "word": " it's",
                    "probability": 0.94970703125
                },
                {
                    "start": 161.77,
                    "end": 162.37,
                    "word": " both",
                    "probability": 0.5693359375
                },
                {
                    "start": 162.37,
                    "end": 162.85,
                    "word": ".",
                    "probability": 0.8115234375
                },
                {
                    "start": 162.85,
                    "end": 163.19,
                    "word": " It's",
                    "probability": 0.95166015625
                },
                {
                    "start": 163.19,
                    "end": 163.39,
                    "word": " kind",
                    "probability": 0.69384765625
                },
                {
                    "start": 163.39,
                    "end": 163.49,
                    "word": " of",
                    "probability": 0.908203125
                },
                {
                    "start": 163.49,
                    "end": 163.95,
                    "word": " any",
                    "probability": 0.80029296875
                },
                {
                    "start": 163.95,
                    "end": 165.17,
                    "word": " intersection",
                    "probability": 0.77587890625
                },
                {
                    "start": 165.17,
                    "end": 165.73,
                    "word": " of",
                    "probability": 0.90625
                },
                {
                    "start": 165.73,
                    "end": 166.25,
                    "word": " artificial",
                    "probability": 0.818359375
                }
            ]
        },
        {
            "id": 24,
            "seek": 18754,
            "start": 166.25,
            "end": 174.35,
            "text": " intelligence, right? That broad category of machine learning slash AI and security.",
            "tokens": [
                50368,
                7599,
                11,
                558,
                30,
                663,
                4152,
                7719,
                295,
                3479,
                2539,
                17330,
                7318,
                293,
                3825,
                13,
                50728
            ],
            "temperature": 0.0,
            "avg_logprob": -0.1964738177689346,
            "compression_ratio": 1.4489795918367347,
            "no_speech_prob": 0.2449951171875,
            "words": [
                {
                    "start": 166.25,
                    "end": 166.97,
                    "word": " intelligence",
                    "probability": 0.354736328125
                },
                {
                    "start": 166.97,
                    "end": 167.97,
                    "word": ",",
                    "probability": 0.64111328125
                },
                {
                    "start": 167.97,
                    "end": 168.67,
                    "word": " right",
                    "probability": 0.3984375
                },
                {
                    "start": 168.67,
                    "end": 168.77,
                    "word": "?",
                    "probability": 0.2247314453125
                },
                {
                    "start": 168.77,
                    "end": 168.93,
                    "word": " That",
                    "probability": 0.7939453125
                },
                {
                    "start": 168.93,
                    "end": 169.39,
                    "word": " broad",
                    "probability": 0.912109375
                },
                {
                    "start": 169.39,
                    "end": 169.95,
                    "word": " category",
                    "probability": 0.8271484375
                },
                {
                    "start": 169.95,
                    "end": 170.31,
                    "word": " of",
                    "probability": 0.90087890625
                },
                {
                    "start": 170.31,
                    "end": 170.85,
                    "word": " machine",
                    "probability": 0.70654296875
                },
                {
                    "start": 170.85,
                    "end": 171.27,
                    "word": " learning",
                    "probability": 0.84765625
                },
                {
                    "start": 171.27,
                    "end": 171.77,
                    "word": " slash",
                    "probability": 0.263916015625
                },
                {
                    "start": 171.77,
                    "end": 172.23,
                    "word": " AI",
                    "probability": 0.85400390625
                },
                {
                    "start": 172.23,
                    "end": 172.99,
                    "word": " and",
                    "probability": 0.525390625
                },
                {
                    "start": 172.99,
                    "end": 173.53,
                    "word": " security",
                    "probability": 0.927734375
                },
                {
                    "start": 173.53,
                    "end": 174.35,
                    "word": ".",
                    "probability": 0.833984375
                }
            ]
        },
        {
            "id": 25,
            "seek": 18754,
            "start": 174.47,
            "end": 184.79,
            "text": " And since this is my talk, and since I am the Mao Zedong of MLSat, I get to decide that it's both.",
            "tokens": [
                50792,
                400,
                1670,
                341,
                307,
                452,
                751,
                11,
                293,
                1670,
                286,
                669,
                264,
                38030,
                1176,
                292,
                556,
                295,
                376,
                19198,
                267,
                11,
                286,
                483,
                281,
                4536,
                300,
                309,
                311,
                1293,
                13,
                51284
            ],
            "temperature": 0.0,
            "avg_logprob": -0.1964738177689346,
            "compression_ratio": 1.4489795918367347,
            "no_speech_prob": 0.2449951171875,
            "words": [
                {
                    "start": 174.47,
                    "end": 175.21,
                    "word": " And",
                    "probability": 0.80615234375
                },
                {
                    "start": 175.21,
                    "end": 175.57,
                    "word": " since",
                    "probability": 0.79833984375
                },
                {
                    "start": 175.57,
                    "end": 175.79,
                    "word": " this",
                    "probability": 0.884765625
                },
                {
                    "start": 175.79,
                    "end": 175.95,
                    "word": " is",
                    "probability": 0.90625
                },
                {
                    "start": 175.95,
                    "end": 176.25,
                    "word": " my",
                    "probability": 0.90771484375
                },
                {
                    "start": 176.25,
                    "end": 176.75,
                    "word": " talk",
                    "probability": 0.81884765625
                },
                {
                    "start": 176.75,
                    "end": 177.09,
                    "word": ",",
                    "probability": 0.56494140625
                },
                {
                    "start": 177.09,
                    "end": 177.39,
                    "word": " and",
                    "probability": 0.8857421875
                },
                {
                    "start": 177.39,
                    "end": 177.99,
                    "word": " since",
                    "probability": 0.81298828125
                },
                {
                    "start": 177.99,
                    "end": 178.47,
                    "word": " I",
                    "probability": 0.9931640625
                },
                {
                    "start": 178.47,
                    "end": 178.67,
                    "word": " am",
                    "probability": 0.89794921875
                },
                {
                    "start": 178.67,
                    "end": 178.89,
                    "word": " the",
                    "probability": 0.7998046875
                },
                {
                    "start": 178.89,
                    "end": 179.13,
                    "word": " Mao",
                    "probability": 0.6708984375
                },
                {
                    "start": 179.13,
                    "end": 179.59,
                    "word": " Zedong",
                    "probability": 0.9443359375
                },
                {
                    "start": 179.59,
                    "end": 180.15,
                    "word": " of",
                    "probability": 0.8974609375
                },
                {
                    "start": 180.15,
                    "end": 180.99,
                    "word": " MLSat",
                    "probability": 0.40529123942057294
                },
                {
                    "start": 180.99,
                    "end": 181.25,
                    "word": ",",
                    "probability": 0.7900390625
                },
                {
                    "start": 181.25,
                    "end": 182.23,
                    "word": " I",
                    "probability": 0.986328125
                },
                {
                    "start": 182.23,
                    "end": 182.57,
                    "word": " get",
                    "probability": 0.90087890625
                },
                {
                    "start": 182.57,
                    "end": 183.15,
                    "word": " to",
                    "probability": 0.90283203125
                },
                {
                    "start": 183.15,
                    "end": 183.61,
                    "word": " decide",
                    "probability": 0.89453125
                },
                {
                    "start": 183.61,
                    "end": 184.07,
                    "word": " that",
                    "probability": 0.8564453125
                },
                {
                    "start": 184.07,
                    "end": 184.31,
                    "word": " it's",
                    "probability": 0.942626953125
                },
                {
                    "start": 184.31,
                    "end": 184.63,
                    "word": " both",
                    "probability": 0.7998046875
                },
                {
                    "start": 184.63,
                    "end": 184.79,
                    "word": ".",
                    "probability": 0.7802734375
                }
            ]
        },
        {
            "id": 26,
            "seek": 18754,
            "start": 184.79,
            "end": 191.27,
            "text": " And if anybody argues with you or you hear anybody having this totally asinine argument, you can just",
            "tokens": [
                51284,
                400,
                498,
                4472,
                38218,
                365,
                291,
                420,
                291,
                1568,
                4472,
                1419,
                341,
                3879,
                382,
                259,
                533,
                6770,
                11,
                291,
                393,
                445,
                51620
            ],
            "temperature": 0.0,
            "avg_logprob": -0.1964738177689346,
            "compression_ratio": 1.4489795918367347,
            "no_speech_prob": 0.2449951171875,
            "words": [
                {
                    "start": 184.79,
                    "end": 184.91,
                    "word": " And",
                    "probability": 0.873046875
                },
                {
                    "start": 184.91,
                    "end": 185.05,
                    "word": " if",
                    "probability": 0.9033203125
                },
                {
                    "start": 185.05,
                    "end": 185.45,
                    "word": " anybody",
                    "probability": 0.76513671875
                },
                {
                    "start": 185.45,
                    "end": 186.11,
                    "word": " argues",
                    "probability": 0.90283203125
                },
                {
                    "start": 186.11,
                    "end": 186.39,
                    "word": " with",
                    "probability": 0.8125
                },
                {
                    "start": 186.39,
                    "end": 186.65,
                    "word": " you",
                    "probability": 0.89892578125
                },
                {
                    "start": 186.65,
                    "end": 186.89,
                    "word": " or",
                    "probability": 0.52099609375
                },
                {
                    "start": 186.89,
                    "end": 187.09,
                    "word": " you",
                    "probability": 0.89794921875
                },
                {
                    "start": 187.09,
                    "end": 187.35,
                    "word": " hear",
                    "probability": 0.8623046875
                },
                {
                    "start": 187.35,
                    "end": 187.77,
                    "word": " anybody",
                    "probability": 0.73046875
                },
                {
                    "start": 187.77,
                    "end": 188.15,
                    "word": " having",
                    "probability": 0.81298828125
                },
                {
                    "start": 188.15,
                    "end": 188.57,
                    "word": " this",
                    "probability": 0.87890625
                },
                {
                    "start": 188.57,
                    "end": 189.39,
                    "word": " totally",
                    "probability": 0.767578125
                },
                {
                    "start": 189.39,
                    "end": 190.05,
                    "word": " asinine",
                    "probability": 0.9420572916666666
                },
                {
                    "start": 190.05,
                    "end": 190.51,
                    "word": " argument",
                    "probability": 0.84521484375
                },
                {
                    "start": 190.51,
                    "end": 190.51,
                    "word": ",",
                    "probability": 0.91064453125
                },
                {
                    "start": 190.51,
                    "end": 190.83,
                    "word": " you",
                    "probability": 0.8955078125
                },
                {
                    "start": 190.83,
                    "end": 191.01,
                    "word": " can",
                    "probability": 0.896484375
                },
                {
                    "start": 191.01,
                    "end": 191.27,
                    "word": " just",
                    "probability": 0.88623046875
                }
            ]
        },
        {
            "id": 27,
            "seek": 21168,
            "start": 191.27,
            "end": 196.47,
            "text": " tell them that I said it's both. And then the argument can be over and everybody can do more",
            "tokens": [
                50368,
                980,
                552,
                300,
                286,
                848,
                309,
                311,
                1293,
                13,
                400,
                550,
                264,
                6770,
                393,
                312,
                670,
                293,
                2201,
                393,
                360,
                544,
                50620
            ],
            "temperature": 0.0,
            "avg_logprob": -0.2110164204631189,
            "compression_ratio": 1.6422413793103448,
            "no_speech_prob": 0.00014889240264892578,
            "words": [
                {
                    "start": 191.27,
                    "end": 191.65,
                    "word": " tell",
                    "probability": 0.443603515625
                },
                {
                    "start": 191.65,
                    "end": 191.95,
                    "word": " them",
                    "probability": 0.73291015625
                },
                {
                    "start": 191.95,
                    "end": 192.17,
                    "word": " that",
                    "probability": 0.8515625
                },
                {
                    "start": 192.17,
                    "end": 192.29,
                    "word": " I",
                    "probability": 0.9697265625
                },
                {
                    "start": 192.29,
                    "end": 192.51,
                    "word": " said",
                    "probability": 0.86376953125
                },
                {
                    "start": 192.51,
                    "end": 192.75,
                    "word": " it's",
                    "probability": 0.916015625
                },
                {
                    "start": 192.75,
                    "end": 193.05,
                    "word": " both",
                    "probability": 0.78662109375
                },
                {
                    "start": 193.05,
                    "end": 193.53,
                    "word": ".",
                    "probability": 0.20703125
                },
                {
                    "start": 193.61,
                    "end": 193.97,
                    "word": " And",
                    "probability": 0.68017578125
                },
                {
                    "start": 193.97,
                    "end": 194.13,
                    "word": " then",
                    "probability": 0.72705078125
                },
                {
                    "start": 194.13,
                    "end": 194.29,
                    "word": " the",
                    "probability": 0.79736328125
                },
                {
                    "start": 194.29,
                    "end": 194.55,
                    "word": " argument",
                    "probability": 0.85302734375
                },
                {
                    "start": 194.55,
                    "end": 194.77,
                    "word": " can",
                    "probability": 0.88720703125
                },
                {
                    "start": 194.77,
                    "end": 194.91,
                    "word": " be",
                    "probability": 0.8984375
                },
                {
                    "start": 194.91,
                    "end": 195.15,
                    "word": " over",
                    "probability": 0.880859375
                },
                {
                    "start": 195.15,
                    "end": 195.41,
                    "word": " and",
                    "probability": 0.7099609375
                },
                {
                    "start": 195.41,
                    "end": 195.69,
                    "word": " everybody",
                    "probability": 0.69580078125
                },
                {
                    "start": 195.69,
                    "end": 195.97,
                    "word": " can",
                    "probability": 0.90087890625
                },
                {
                    "start": 195.97,
                    "end": 196.17,
                    "word": " do",
                    "probability": 0.91064453125
                },
                {
                    "start": 196.17,
                    "end": 196.47,
                    "word": " more",
                    "probability": 0.87939453125
                }
            ]
        },
        {
            "id": 28,
            "seek": 21168,
            "start": 196.47,
            "end": 204.19,
            "text": " productive things with their lives. So we'll start with deep fakes, right? So D is for deep fakes.",
            "tokens": [
                50620,
                13304,
                721,
                365,
                641,
                2909,
                13,
                407,
                321,
                603,
                722,
                365,
                2452,
                283,
                3419,
                11,
                558,
                30,
                407,
                413,
                307,
                337,
                2452,
                283,
                3419,
                13,
                51000
            ],
            "temperature": 0.0,
            "avg_logprob": -0.2110164204631189,
            "compression_ratio": 1.6422413793103448,
            "no_speech_prob": 0.00014889240264892578,
            "words": [
                {
                    "start": 196.47,
                    "end": 196.83,
                    "word": " productive",
                    "probability": 0.7177734375
                },
                {
                    "start": 196.83,
                    "end": 197.15,
                    "word": " things",
                    "probability": 0.72607421875
                },
                {
                    "start": 197.15,
                    "end": 197.31,
                    "word": " with",
                    "probability": 0.818359375
                },
                {
                    "start": 197.31,
                    "end": 197.45,
                    "word": " their",
                    "probability": 0.90087890625
                },
                {
                    "start": 197.45,
                    "end": 197.83,
                    "word": " lives",
                    "probability": 0.74072265625
                },
                {
                    "start": 197.83,
                    "end": 198.31,
                    "word": ".",
                    "probability": 0.97998046875
                },
                {
                    "start": 198.65,
                    "end": 198.95,
                    "word": " So",
                    "probability": 0.83642578125
                },
                {
                    "start": 198.95,
                    "end": 199.89,
                    "word": " we'll",
                    "probability": 0.7431640625
                },
                {
                    "start": 199.89,
                    "end": 200.09,
                    "word": " start",
                    "probability": 0.865234375
                },
                {
                    "start": 200.09,
                    "end": 200.35,
                    "word": " with",
                    "probability": 0.82373046875
                },
                {
                    "start": 200.35,
                    "end": 200.81,
                    "word": " deep",
                    "probability": 0.4677734375
                },
                {
                    "start": 200.81,
                    "end": 201.19,
                    "word": " fakes",
                    "probability": 0.689208984375
                },
                {
                    "start": 201.19,
                    "end": 201.41,
                    "word": ",",
                    "probability": 0.2401123046875
                },
                {
                    "start": 201.41,
                    "end": 201.85,
                    "word": " right",
                    "probability": 0.434814453125
                },
                {
                    "start": 201.85,
                    "end": 202.01,
                    "word": "?",
                    "probability": 0.70556640625
                },
                {
                    "start": 202.01,
                    "end": 202.11,
                    "word": " So",
                    "probability": 0.80419921875
                },
                {
                    "start": 202.11,
                    "end": 202.55,
                    "word": " D",
                    "probability": 0.6572265625
                },
                {
                    "start": 202.55,
                    "end": 202.77,
                    "word": " is",
                    "probability": 0.91162109375
                },
                {
                    "start": 202.77,
                    "end": 202.97,
                    "word": " for",
                    "probability": 0.904296875
                },
                {
                    "start": 202.97,
                    "end": 203.23,
                    "word": " deep",
                    "probability": 0.7958984375
                },
                {
                    "start": 203.23,
                    "end": 203.71,
                    "word": " fakes",
                    "probability": 0.8798828125
                },
                {
                    "start": 203.71,
                    "end": 204.19,
                    "word": ".",
                    "probability": 0.7392578125
                }
            ]
        },
        {
            "id": 29,
            "seek": 21168,
            "start": 204.19,
            "end": 211.11,
            "text": " And this man on the right, if I didn't have a speaker video, I probably could have just put",
            "tokens": [
                51000,
                400,
                341,
                587,
                322,
                264,
                558,
                11,
                498,
                286,
                994,
                380,
                362,
                257,
                8145,
                960,
                11,
                286,
                1391,
                727,
                362,
                445,
                829,
                51352
            ],
            "temperature": 0.0,
            "avg_logprob": -0.2110164204631189,
            "compression_ratio": 1.6422413793103448,
            "no_speech_prob": 0.00014889240264892578,
            "words": [
                {
                    "start": 204.19,
                    "end": 204.51,
                    "word": " And",
                    "probability": 0.8115234375
                },
                {
                    "start": 204.51,
                    "end": 205.95,
                    "word": " this",
                    "probability": 0.796875
                },
                {
                    "start": 205.95,
                    "end": 206.27,
                    "word": " man",
                    "probability": 0.89892578125
                },
                {
                    "start": 206.27,
                    "end": 206.51,
                    "word": " on",
                    "probability": 0.8916015625
                },
                {
                    "start": 206.51,
                    "end": 206.67,
                    "word": " the",
                    "probability": 0.81396484375
                },
                {
                    "start": 206.67,
                    "end": 206.93,
                    "word": " right",
                    "probability": 0.8798828125
                },
                {
                    "start": 206.93,
                    "end": 207.15,
                    "word": ",",
                    "probability": 0.89404296875
                },
                {
                    "start": 207.15,
                    "end": 207.39,
                    "word": " if",
                    "probability": 0.900390625
                },
                {
                    "start": 207.39,
                    "end": 207.53,
                    "word": " I",
                    "probability": 0.99658203125
                },
                {
                    "start": 207.53,
                    "end": 207.75,
                    "word": " didn't",
                    "probability": 0.9267578125
                },
                {
                    "start": 207.75,
                    "end": 207.91,
                    "word": " have",
                    "probability": 0.88623046875
                },
                {
                    "start": 208.01,
                    "end": 208.29,
                    "word": " a speaker",
                    "probability": 0.75
                },
                {
                    "start": 208.29,
                    "end": 208.75,
                    "word": " video",
                    "probability": 0.78271484375
                },
                {
                    "start": 208.75,
                    "end": 209.91,
                    "word": ",",
                    "probability": 0.87060546875
                },
                {
                    "start": 209.91,
                    "end": 210.11,
                    "word": " I",
                    "probability": 0.9873046875
                },
                {
                    "start": 210.11,
                    "end": 210.43,
                    "word": " probably",
                    "probability": 0.9091796875
                },
                {
                    "start": 210.43,
                    "end": 210.67,
                    "word": " could",
                    "probability": 0.81298828125
                },
                {
                    "start": 210.67,
                    "end": 210.75,
                    "word": " have",
                    "probability": 0.65673828125
                },
                {
                    "start": 210.75,
                    "end": 210.93,
                    "word": " just",
                    "probability": 0.87548828125
                },
                {
                    "start": 210.93,
                    "end": 211.11,
                    "word": " put",
                    "probability": 0.89990234375
                }
            ]
        },
        {
            "id": 30,
            "seek": 21168,
            "start": 211.11,
            "end": 215.41,
            "text": " that picture up and said it was me. And everybody would have believed it because he's got kind of",
            "tokens": [
                51352,
                300,
                3036,
                493,
                293,
                848,
                309,
                390,
                385,
                13,
                400,
                2201,
                576,
                362,
                7847,
                309,
                570,
                415,
                311,
                658,
                733,
                295,
                51576
            ],
            "temperature": 0.0,
            "avg_logprob": -0.2110164204631189,
            "compression_ratio": 1.6422413793103448,
            "no_speech_prob": 0.00014889240264892578,
            "words": [
                {
                    "start": 211.11,
                    "end": 211.31,
                    "word": " that",
                    "probability": 0.880859375
                },
                {
                    "start": 211.31,
                    "end": 211.63,
                    "word": " picture",
                    "probability": 0.7548828125
                },
                {
                    "start": 211.63,
                    "end": 211.87,
                    "word": " up",
                    "probability": 0.90234375
                },
                {
                    "start": 211.87,
                    "end": 212.03,
                    "word": " and",
                    "probability": 0.8818359375
                },
                {
                    "start": 212.03,
                    "end": 212.15,
                    "word": " said",
                    "probability": 0.8876953125
                },
                {
                    "start": 212.15,
                    "end": 212.27,
                    "word": " it",
                    "probability": 0.873046875
                },
                {
                    "start": 212.27,
                    "end": 212.41,
                    "word": " was",
                    "probability": 0.89794921875
                },
                {
                    "start": 212.41,
                    "end": 212.65,
                    "word": " me",
                    "probability": 0.9130859375
                },
                {
                    "start": 212.65,
                    "end": 212.89,
                    "word": ".",
                    "probability": 0.69580078125
                },
                {
                    "start": 212.89,
                    "end": 213.27,
                    "word": " And",
                    "probability": 0.89208984375
                },
                {
                    "start": 213.27,
                    "end": 213.57,
                    "word": " everybody",
                    "probability": 0.7548828125
                },
                {
                    "start": 213.57,
                    "end": 213.77,
                    "word": " would",
                    "probability": 0.8427734375
                },
                {
                    "start": 213.77,
                    "end": 213.87,
                    "word": " have",
                    "probability": 0.8076171875
                },
                {
                    "start": 213.87,
                    "end": 214.13,
                    "word": " believed",
                    "probability": 0.66064453125
                },
                {
                    "start": 214.13,
                    "end": 214.43,
                    "word": " it",
                    "probability": 0.9013671875
                },
                {
                    "start": 214.43,
                    "end": 214.63,
                    "word": " because",
                    "probability": 0.3125
                },
                {
                    "start": 214.63,
                    "end": 214.83,
                    "word": " he's",
                    "probability": 0.93212890625
                },
                {
                    "start": 214.83,
                    "end": 214.99,
                    "word": " got",
                    "probability": 0.8818359375
                },
                {
                    "start": 214.99,
                    "end": 215.25,
                    "word": " kind",
                    "probability": 0.681640625
                },
                {
                    "start": 215.25,
                    "end": 215.41,
                    "word": " of",
                    "probability": 0.89306640625
                }
            ]
        },
        {
            "id": 31,
            "seek": 24138,
            "start": 215.41,
            "end": 223.96,
            "text": " he looks like the kind of guy, right? And I'm allowed to say that because he's not real.",
            "tokens": [
                50412,
                415,
                1542,
                411,
                264,
                733,
                295,
                2146,
                11,
                558,
                30,
                400,
                286,
                478,
                4350,
                281,
                584,
                300,
                570,
                415,
                311,
                406,
                957,
                13,
                50776
            ],
            "temperature": 0.0,
            "avg_logprob": -0.2049731246886715,
            "compression_ratio": 1.566820276497696,
            "no_speech_prob": 0.0015010833740234375,
            "words": [
                {
                    "start": 215.41,
                    "end": 216.25,
                    "word": " he",
                    "probability": 0.0215301513671875
                },
                {
                    "start": 216.25,
                    "end": 216.83,
                    "word": " looks",
                    "probability": 0.66259765625
                },
                {
                    "start": 216.83,
                    "end": 217.19,
                    "word": " like",
                    "probability": 0.828125
                },
                {
                    "start": 217.19,
                    "end": 218.11,
                    "word": " the",
                    "probability": 0.67626953125
                },
                {
                    "start": 218.11,
                    "end": 218.33,
                    "word": " kind",
                    "probability": 0.63623046875
                },
                {
                    "start": 218.33,
                    "end": 218.41,
                    "word": " of",
                    "probability": 0.8994140625
                },
                {
                    "start": 218.41,
                    "end": 218.71,
                    "word": " guy",
                    "probability": 0.8056640625
                },
                {
                    "start": 218.71,
                    "end": 218.91,
                    "word": ",",
                    "probability": 0.41796875
                },
                {
                    "start": 218.91,
                    "end": 219.17,
                    "word": " right",
                    "probability": 0.84912109375
                },
                {
                    "start": 219.17,
                    "end": 219.79,
                    "word": "?",
                    "probability": 0.7490234375
                },
                {
                    "start": 220.79,
                    "end": 221.41,
                    "word": " And",
                    "probability": 0.78271484375
                },
                {
                    "start": 221.41,
                    "end": 221.57,
                    "word": " I'm",
                    "probability": 0.982666015625
                },
                {
                    "start": 221.57,
                    "end": 221.73,
                    "word": " allowed",
                    "probability": 0.79541015625
                },
                {
                    "start": 221.73,
                    "end": 221.93,
                    "word": " to",
                    "probability": 0.90380859375
                },
                {
                    "start": 221.93,
                    "end": 222.07,
                    "word": " say",
                    "probability": 0.89794921875
                },
                {
                    "start": 222.07,
                    "end": 222.39,
                    "word": " that",
                    "probability": 0.8828125
                },
                {
                    "start": 222.39,
                    "end": 222.91,
                    "word": " because",
                    "probability": 0.62646484375
                },
                {
                    "start": 222.91,
                    "end": 223.15,
                    "word": " he's",
                    "probability": 0.947509765625
                },
                {
                    "start": 223.15,
                    "end": 223.33,
                    "word": " not",
                    "probability": 0.896484375
                },
                {
                    "start": 223.33,
                    "end": 223.65,
                    "word": " real",
                    "probability": 0.8955078125
                },
                {
                    "start": 223.65,
                    "end": 223.96,
                    "word": ".",
                    "probability": 0.87890625
                }
            ]
        },
        {
            "id": 32,
            "seek": 24138,
            "start": 224.19,
            "end": 229.63,
            "text": " This is not a real person. This is completely an AI-generated image.",
            "tokens": [
                50820,
                639,
                307,
                406,
                257,
                957,
                954,
                13,
                639,
                307,
                2584,
                364,
                7318,
                12,
                21848,
                770,
                3256,
                13,
                51040
            ],
            "temperature": 0.0,
            "avg_logprob": -0.2049731246886715,
            "compression_ratio": 1.566820276497696,
            "no_speech_prob": 0.0015010833740234375,
            "words": [
                {
                    "start": 224.19,
                    "end": 224.81,
                    "word": " This",
                    "probability": 0.80859375
                },
                {
                    "start": 224.81,
                    "end": 224.97,
                    "word": " is",
                    "probability": 0.90869140625
                },
                {
                    "start": 224.97,
                    "end": 225.23,
                    "word": " not",
                    "probability": 0.8916015625
                },
                {
                    "start": 225.51,
                    "end": 225.69,
                    "word": " a real",
                    "probability": 0.89599609375
                },
                {
                    "start": 225.69,
                    "end": 226.11,
                    "word": " person",
                    "probability": 0.79638671875
                },
                {
                    "start": 226.11,
                    "end": 226.31,
                    "word": ".",
                    "probability": 0.83642578125
                },
                {
                    "start": 226.31,
                    "end": 226.45,
                    "word": " This",
                    "probability": 0.8125
                },
                {
                    "start": 226.45,
                    "end": 226.57,
                    "word": " is",
                    "probability": 0.908203125
                },
                {
                    "start": 226.57,
                    "end": 227.19,
                    "word": " completely",
                    "probability": 0.63720703125
                },
                {
                    "start": 227.19,
                    "end": 227.53,
                    "word": " an",
                    "probability": 0.90673828125
                },
                {
                    "start": 227.53,
                    "end": 227.89,
                    "word": " AI",
                    "probability": 0.90576171875
                },
                {
                    "start": 227.89,
                    "end": 228.51,
                    "word": "-generated",
                    "probability": 0.6563313802083334
                },
                {
                    "start": 228.51,
                    "end": 229.01,
                    "word": " image",
                    "probability": 0.826171875
                },
                {
                    "start": 229.01,
                    "end": 229.63,
                    "word": ".",
                    "probability": 0.9287109375
                }
            ]
        },
        {
            "id": 33,
            "seek": 24138,
            "start": 230.49,
            "end": 236.89,
            "text": " And so, you know, to give deep fakes a real definition, they are a convincing synthetic",
            "tokens": [
                51128,
                400,
                370,
                11,
                291,
                458,
                11,
                281,
                976,
                2452,
                283,
                3419,
                257,
                957,
                7123,
                11,
                436,
                366,
                257,
                24823,
                23420,
                51456
            ],
            "temperature": 0.0,
            "avg_logprob": -0.2049731246886715,
            "compression_ratio": 1.566820276497696,
            "no_speech_prob": 0.0015010833740234375,
            "words": [
                {
                    "start": 230.49,
                    "end": 230.97,
                    "word": " And",
                    "probability": 0.8310546875
                },
                {
                    "start": 230.97,
                    "end": 231.35,
                    "word": " so",
                    "probability": 0.89013671875
                },
                {
                    "start": 231.35,
                    "end": 231.61,
                    "word": ",",
                    "probability": 0.6806640625
                },
                {
                    "start": 231.61,
                    "end": 231.75,
                    "word": " you",
                    "probability": 0.583984375
                },
                {
                    "start": 231.75,
                    "end": 231.83,
                    "word": " know",
                    "probability": 0.8193359375
                },
                {
                    "start": 231.83,
                    "end": 231.89,
                    "word": ",",
                    "probability": 0.96533203125
                },
                {
                    "start": 231.89,
                    "end": 231.99,
                    "word": " to",
                    "probability": 0.9033203125
                },
                {
                    "start": 231.99,
                    "end": 232.19,
                    "word": " give",
                    "probability": 0.8291015625
                },
                {
                    "start": 232.19,
                    "end": 232.73,
                    "word": " deep",
                    "probability": 0.90771484375
                },
                {
                    "start": 232.73,
                    "end": 233.01,
                    "word": " fakes",
                    "probability": 0.630615234375
                },
                {
                    "start": 233.15,
                    "end": 233.37,
                    "word": " a real",
                    "probability": 0.8974609375
                },
                {
                    "start": 233.37,
                    "end": 233.99,
                    "word": " definition",
                    "probability": 0.84130859375
                },
                {
                    "start": 233.99,
                    "end": 234.67,
                    "word": ",",
                    "probability": 0.8779296875
                },
                {
                    "start": 234.67,
                    "end": 234.89,
                    "word": " they",
                    "probability": 0.66064453125
                },
                {
                    "start": 234.89,
                    "end": 235.11,
                    "word": " are",
                    "probability": 0.390869140625
                },
                {
                    "start": 235.39,
                    "end": 236.13,
                    "word": " a convincing",
                    "probability": 0.75390625
                },
                {
                    "start": 236.13,
                    "end": 236.89,
                    "word": " synthetic",
                    "probability": 0.6435546875
                }
            ]
        },
        {
            "id": 34,
            "seek": 24138,
            "start": 236.89,
            "end": 245.11,
            "text": " image, video, or audio recording which purports to be real. So, you know, I'm pretty convinced",
            "tokens": [
                51456,
                3256,
                11,
                960,
                11,
                420,
                6278,
                6613,
                597,
                1864,
                17845,
                281,
                312,
                957,
                13,
                407,
                11,
                291,
                458,
                11,
                286,
                478,
                1238,
                12561,
                51852
            ],
            "temperature": 0.0,
            "avg_logprob": -0.2049731246886715,
            "compression_ratio": 1.566820276497696,
            "no_speech_prob": 0.0015010833740234375,
            "words": [
                {
                    "start": 236.89,
                    "end": 237.73,
                    "word": " image",
                    "probability": 0.8173828125
                },
                {
                    "start": 237.73,
                    "end": 237.89,
                    "word": ",",
                    "probability": 0.81982421875
                },
                {
                    "start": 237.89,
                    "end": 238.39,
                    "word": " video",
                    "probability": 0.78662109375
                },
                {
                    "start": 238.39,
                    "end": 238.61,
                    "word": ",",
                    "probability": 0.65771484375
                },
                {
                    "start": 238.61,
                    "end": 238.81,
                    "word": " or",
                    "probability": 0.91357421875
                },
                {
                    "start": 238.81,
                    "end": 239.21,
                    "word": " audio",
                    "probability": 0.93798828125
                },
                {
                    "start": 239.21,
                    "end": 239.69,
                    "word": " recording",
                    "probability": 0.8212890625
                },
                {
                    "start": 239.69,
                    "end": 240.11,
                    "word": " which",
                    "probability": 0.515625
                },
                {
                    "start": 240.11,
                    "end": 240.81,
                    "word": " purports",
                    "probability": 0.90380859375
                },
                {
                    "start": 240.81,
                    "end": 241.09,
                    "word": " to",
                    "probability": 0.904296875
                },
                {
                    "start": 241.09,
                    "end": 241.23,
                    "word": " be",
                    "probability": 0.90966796875
                },
                {
                    "start": 241.23,
                    "end": 241.63,
                    "word": " real",
                    "probability": 0.86474609375
                },
                {
                    "start": 241.63,
                    "end": 242.11,
                    "word": ".",
                    "probability": 0.958984375
                },
                {
                    "start": 242.11,
                    "end": 242.61,
                    "word": " So",
                    "probability": 0.8779296875
                },
                {
                    "start": 242.61,
                    "end": 243.31,
                    "word": ",",
                    "probability": 0.86669921875
                },
                {
                    "start": 243.31,
                    "end": 243.45,
                    "word": " you",
                    "probability": 0.83349609375
                },
                {
                    "start": 243.45,
                    "end": 243.59,
                    "word": " know",
                    "probability": 0.81689453125
                },
                {
                    "start": 243.59,
                    "end": 243.97,
                    "word": ",",
                    "probability": 0.9873046875
                },
                {
                    "start": 243.97,
                    "end": 244.29,
                    "word": " I'm",
                    "probability": 0.999267578125
                },
                {
                    "start": 244.29,
                    "end": 244.51,
                    "word": " pretty",
                    "probability": 0.89599609375
                },
                {
                    "start": 244.51,
                    "end": 245.11,
                    "word": " convinced",
                    "probability": 0.751953125
                }
            ]
        },
        {
            "id": 35,
            "seek": 26874,
            "start": 245.11,
            "end": 251.75,
            "text": " by this picture. If I saw it in the wild, right, if somebody put it as like their Twitter avatar",
            "tokens": [
                50364,
                538,
                341,
                3036,
                13,
                759,
                286,
                1866,
                309,
                294,
                264,
                4868,
                11,
                558,
                11,
                498,
                2618,
                829,
                309,
                382,
                411,
                641,
                5794,
                36205,
                50728
            ],
            "temperature": 0.0,
            "avg_logprob": -0.2365196037526224,
            "compression_ratio": 1.6130434782608696,
            "no_speech_prob": 0.0003981590270996094,
            "words": [
                {
                    "start": 245.11,
                    "end": 245.35,
                    "word": " by",
                    "probability": 0.896484375
                },
                {
                    "start": 245.35,
                    "end": 245.57,
                    "word": " this",
                    "probability": 0.88232421875
                },
                {
                    "start": 245.57,
                    "end": 245.97,
                    "word": " picture",
                    "probability": 0.7666015625
                },
                {
                    "start": 245.97,
                    "end": 246.45,
                    "word": ".",
                    "probability": 0.54541015625
                },
                {
                    "start": 246.71,
                    "end": 247.19,
                    "word": " If",
                    "probability": 0.833984375
                },
                {
                    "start": 247.19,
                    "end": 247.39,
                    "word": " I",
                    "probability": 0.982421875
                },
                {
                    "start": 247.39,
                    "end": 247.79,
                    "word": " saw",
                    "probability": 0.82373046875
                },
                {
                    "start": 247.79,
                    "end": 248.05,
                    "word": " it",
                    "probability": 0.89697265625
                },
                {
                    "start": 248.05,
                    "end": 248.83,
                    "word": " in",
                    "probability": 0.87060546875
                },
                {
                    "start": 248.83,
                    "end": 248.97,
                    "word": " the",
                    "probability": 0.81298828125
                },
                {
                    "start": 248.97,
                    "end": 249.37,
                    "word": " wild",
                    "probability": 0.8544921875
                },
                {
                    "start": 249.37,
                    "end": 249.53,
                    "word": ",",
                    "probability": 0.7890625
                },
                {
                    "start": 249.53,
                    "end": 249.67,
                    "word": " right",
                    "probability": 0.619140625
                },
                {
                    "start": 249.67,
                    "end": 249.69,
                    "word": ",",
                    "probability": 0.4970703125
                },
                {
                    "start": 249.69,
                    "end": 249.87,
                    "word": " if",
                    "probability": 0.8037109375
                },
                {
                    "start": 249.87,
                    "end": 250.19,
                    "word": " somebody",
                    "probability": 0.78271484375
                },
                {
                    "start": 250.19,
                    "end": 250.41,
                    "word": " put",
                    "probability": 0.9013671875
                },
                {
                    "start": 250.41,
                    "end": 250.53,
                    "word": " it",
                    "probability": 0.8974609375
                },
                {
                    "start": 250.53,
                    "end": 250.67,
                    "word": " as",
                    "probability": 0.8828125
                },
                {
                    "start": 250.67,
                    "end": 250.81,
                    "word": " like",
                    "probability": 0.54345703125
                },
                {
                    "start": 250.81,
                    "end": 250.97,
                    "word": " their",
                    "probability": 0.8740234375
                },
                {
                    "start": 250.97,
                    "end": 251.33,
                    "word": " Twitter",
                    "probability": 0.771484375
                },
                {
                    "start": 251.33,
                    "end": 251.75,
                    "word": " avatar",
                    "probability": 0.91015625
                }
            ]
        },
        {
            "id": 36,
            "seek": 26874,
            "start": 252.37,
            "end": 259.03,
            "text": " and was like, yeah, I am a software engineer at Google, I could buy that. I could buy that",
            "tokens": [
                50728,
                293,
                390,
                411,
                11,
                1338,
                11,
                286,
                669,
                257,
                4722,
                11403,
                412,
                3329,
                11,
                286,
                727,
                2256,
                300,
                13,
                286,
                727,
                2256,
                300,
                51056
            ],
            "temperature": 0.0,
            "avg_logprob": -0.2365196037526224,
            "compression_ratio": 1.6130434782608696,
            "no_speech_prob": 0.0003981590270996094,
            "words": [
                {
                    "start": 252.37,
                    "end": 252.61,
                    "word": " and",
                    "probability": 0.62353515625
                },
                {
                    "start": 252.61,
                    "end": 252.83,
                    "word": " was",
                    "probability": 0.87548828125
                },
                {
                    "start": 252.83,
                    "end": 253.05,
                    "word": " like",
                    "probability": 0.86962890625
                },
                {
                    "start": 253.05,
                    "end": 253.09,
                    "word": ",",
                    "probability": 0.67041015625
                },
                {
                    "start": 253.09,
                    "end": 253.31,
                    "word": " yeah",
                    "probability": 0.46630859375
                },
                {
                    "start": 253.31,
                    "end": 253.37,
                    "word": ",",
                    "probability": 0.8076171875
                },
                {
                    "start": 253.37,
                    "end": 253.75,
                    "word": " I",
                    "probability": 0.9833984375
                },
                {
                    "start": 253.75,
                    "end": 254.31,
                    "word": " am",
                    "probability": 0.90478515625
                },
                {
                    "start": 255.03,
                    "end": 255.81,
                    "word": " a software",
                    "probability": 0.8291015625
                },
                {
                    "start": 255.81,
                    "end": 256.23,
                    "word": " engineer",
                    "probability": 0.71142578125
                },
                {
                    "start": 256.23,
                    "end": 256.53,
                    "word": " at",
                    "probability": 0.9013671875
                },
                {
                    "start": 256.53,
                    "end": 256.87,
                    "word": " Google",
                    "probability": 0.86181640625
                },
                {
                    "start": 256.87,
                    "end": 257.23,
                    "word": ",",
                    "probability": 0.4814453125
                },
                {
                    "start": 257.23,
                    "end": 257.51,
                    "word": " I",
                    "probability": 0.99755859375
                },
                {
                    "start": 257.51,
                    "end": 257.65,
                    "word": " could",
                    "probability": 0.8212890625
                },
                {
                    "start": 257.65,
                    "end": 257.83,
                    "word": " buy",
                    "probability": 0.81005859375
                },
                {
                    "start": 257.83,
                    "end": 258.15,
                    "word": " that",
                    "probability": 0.87939453125
                },
                {
                    "start": 258.15,
                    "end": 258.37,
                    "word": ".",
                    "probability": 0.759765625
                },
                {
                    "start": 258.37,
                    "end": 258.63,
                    "word": " I",
                    "probability": 0.99658203125
                },
                {
                    "start": 258.63,
                    "end": 258.73,
                    "word": " could",
                    "probability": 0.80322265625
                },
                {
                    "start": 258.73,
                    "end": 258.87,
                    "word": " buy",
                    "probability": 0.82958984375
                },
                {
                    "start": 258.87,
                    "end": 259.03,
                    "word": " that",
                    "probability": 0.87890625
                }
            ]
        },
        {
            "id": 37,
            "seek": 26874,
            "start": 259.03,
            "end": 264.81,
            "text": " based on this image, right? You know? We have our stereotypes, right? The thin-rimmed glasses,",
            "tokens": [
                51056,
                2361,
                322,
                341,
                3256,
                11,
                558,
                30,
                509,
                458,
                30,
                492,
                362,
                527,
                30853,
                11,
                558,
                30,
                440,
                5862,
                12,
                5565,
                1912,
                10812,
                11,
                51340
            ],
            "temperature": 0.0,
            "avg_logprob": -0.2365196037526224,
            "compression_ratio": 1.6130434782608696,
            "no_speech_prob": 0.0003981590270996094,
            "words": [
                {
                    "start": 259.03,
                    "end": 259.27,
                    "word": " based",
                    "probability": 0.83935546875
                },
                {
                    "start": 259.27,
                    "end": 259.43,
                    "word": " on",
                    "probability": 0.90673828125
                },
                {
                    "start": 259.43,
                    "end": 259.61,
                    "word": " this",
                    "probability": 0.8798828125
                },
                {
                    "start": 259.61,
                    "end": 259.91,
                    "word": " image",
                    "probability": 0.83544921875
                },
                {
                    "start": 259.91,
                    "end": 259.91,
                    "word": ",",
                    "probability": 0.78271484375
                },
                {
                    "start": 259.91,
                    "end": 260.15,
                    "word": " right",
                    "probability": 0.88037109375
                },
                {
                    "start": 260.15,
                    "end": 260.63,
                    "word": "?",
                    "probability": 0.71484375
                },
                {
                    "start": 260.95,
                    "end": 261.13,
                    "word": " You",
                    "probability": 0.75537109375
                },
                {
                    "start": 261.13,
                    "end": 261.27,
                    "word": " know",
                    "probability": 0.8173828125
                },
                {
                    "start": 261.27,
                    "end": 261.47,
                    "word": "?",
                    "probability": 0.10821533203125
                },
                {
                    "start": 261.47,
                    "end": 261.67,
                    "word": " We",
                    "probability": 0.900390625
                },
                {
                    "start": 261.67,
                    "end": 261.93,
                    "word": " have",
                    "probability": 0.7978515625
                },
                {
                    "start": 261.93,
                    "end": 262.11,
                    "word": " our",
                    "probability": 0.8125
                },
                {
                    "start": 262.11,
                    "end": 262.79,
                    "word": " stereotypes",
                    "probability": 0.876953125
                },
                {
                    "start": 262.79,
                    "end": 262.95,
                    "word": ",",
                    "probability": 0.9072265625
                },
                {
                    "start": 262.95,
                    "end": 263.03,
                    "word": " right",
                    "probability": 0.8486328125
                },
                {
                    "start": 263.03,
                    "end": 263.07,
                    "word": "?",
                    "probability": 0.845703125
                },
                {
                    "start": 263.07,
                    "end": 263.23,
                    "word": " The",
                    "probability": 0.82177734375
                },
                {
                    "start": 263.23,
                    "end": 263.45,
                    "word": " thin",
                    "probability": 0.8876953125
                },
                {
                    "start": 263.45,
                    "end": 263.83,
                    "word": "-rimmed",
                    "probability": 0.7736002604166666
                },
                {
                    "start": 263.83,
                    "end": 264.31,
                    "word": " glasses",
                    "probability": 0.755859375
                },
                {
                    "start": 264.31,
                    "end": 264.81,
                    "word": ",",
                    "probability": 0.8525390625
                }
            ]
        },
        {
            "id": 38,
            "seek": 26874,
            "start": 264.81,
            "end": 272.47,
            "text": " the scruff. And it's super convincing. It's super convincing. And I think that's really,",
            "tokens": [
                51340,
                264,
                795,
                50154,
                13,
                400,
                309,
                311,
                1687,
                24823,
                13,
                467,
                311,
                1687,
                24823,
                13,
                400,
                286,
                519,
                300,
                311,
                534,
                11,
                51724
            ],
            "temperature": 0.0,
            "avg_logprob": -0.2365196037526224,
            "compression_ratio": 1.6130434782608696,
            "no_speech_prob": 0.0003981590270996094,
            "words": [
                {
                    "start": 264.81,
                    "end": 264.97,
                    "word": " the",
                    "probability": 0.79150390625
                },
                {
                    "start": 264.97,
                    "end": 265.55,
                    "word": " scruff",
                    "probability": 0.640869140625
                },
                {
                    "start": 265.55,
                    "end": 266.03,
                    "word": ".",
                    "probability": 0.5615234375
                },
                {
                    "start": 266.51,
                    "end": 266.99,
                    "word": " And",
                    "probability": 0.619140625
                },
                {
                    "start": 266.99,
                    "end": 268.33,
                    "word": " it's",
                    "probability": 0.88720703125
                },
                {
                    "start": 268.33,
                    "end": 268.57,
                    "word": " super",
                    "probability": 0.8896484375
                },
                {
                    "start": 268.57,
                    "end": 268.97,
                    "word": " convincing",
                    "probability": 0.79833984375
                },
                {
                    "start": 268.97,
                    "end": 269.45,
                    "word": ".",
                    "probability": 0.79736328125
                },
                {
                    "start": 269.45,
                    "end": 269.77,
                    "word": " It's",
                    "probability": 0.79931640625
                },
                {
                    "start": 269.77,
                    "end": 270.01,
                    "word": " super",
                    "probability": 0.869140625
                },
                {
                    "start": 270.01,
                    "end": 270.35,
                    "word": " convincing",
                    "probability": 0.83251953125
                },
                {
                    "start": 270.35,
                    "end": 270.71,
                    "word": ".",
                    "probability": 0.81201171875
                },
                {
                    "start": 270.71,
                    "end": 270.99,
                    "word": " And",
                    "probability": 0.87744140625
                },
                {
                    "start": 270.99,
                    "end": 271.73,
                    "word": " I",
                    "probability": 0.98681640625
                },
                {
                    "start": 271.73,
                    "end": 271.85,
                    "word": " think",
                    "probability": 0.8115234375
                },
                {
                    "start": 271.85,
                    "end": 272.13,
                    "word": " that's",
                    "probability": 0.9375
                },
                {
                    "start": 272.13,
                    "end": 272.35,
                    "word": " really",
                    "probability": 0.7666015625
                },
                {
                    "start": 272.35,
                    "end": 272.47,
                    "word": ",",
                    "probability": 0.397705078125
                }
            ]
        },
        {
            "id": 39,
            "seek": 29296,
            "start": 272.47,
            "end": 278.79,
            "text": " when we think about impact, that's what it is. It's that they're convincing to users. So, we also",
            "tokens": [
                50364,
                562,
                321,
                519,
                466,
                2712,
                11,
                300,
                311,
                437,
                309,
                307,
                13,
                467,
                311,
                300,
                436,
                434,
                24823,
                281,
                5022,
                13,
                407,
                11,
                321,
                611,
                50680
            ],
            "temperature": 0.0,
            "avg_logprob": -0.19098395914644808,
            "compression_ratio": 1.553763440860215,
            "no_speech_prob": 0.0002378225326538086,
            "words": [
                {
                    "start": 272.47,
                    "end": 272.63,
                    "word": " when",
                    "probability": 0.0631103515625
                },
                {
                    "start": 272.63,
                    "end": 272.81,
                    "word": " we",
                    "probability": 0.896484375
                },
                {
                    "start": 272.81,
                    "end": 272.97,
                    "word": " think",
                    "probability": 0.81201171875
                },
                {
                    "start": 272.97,
                    "end": 273.17,
                    "word": " about",
                    "probability": 0.80322265625
                },
                {
                    "start": 273.17,
                    "end": 273.89,
                    "word": " impact",
                    "probability": 0.76904296875
                },
                {
                    "start": 273.89,
                    "end": 274.57,
                    "word": ",",
                    "probability": 0.6884765625
                },
                {
                    "start": 274.57,
                    "end": 275.39,
                    "word": " that's",
                    "probability": 0.902099609375
                },
                {
                    "start": 275.39,
                    "end": 275.53,
                    "word": " what",
                    "probability": 0.88525390625
                },
                {
                    "start": 275.53,
                    "end": 275.65,
                    "word": " it",
                    "probability": 0.90087890625
                },
                {
                    "start": 275.65,
                    "end": 275.89,
                    "word": " is",
                    "probability": 0.90234375
                },
                {
                    "start": 275.89,
                    "end": 276.05,
                    "word": ".",
                    "probability": 0.5595703125
                },
                {
                    "start": 276.05,
                    "end": 276.05,
                    "word": " It's",
                    "probability": 0.833740234375
                },
                {
                    "start": 276.05,
                    "end": 276.21,
                    "word": " that",
                    "probability": 0.8134765625
                },
                {
                    "start": 276.21,
                    "end": 276.43,
                    "word": " they're",
                    "probability": 0.83056640625
                },
                {
                    "start": 276.43,
                    "end": 277.03,
                    "word": " convincing",
                    "probability": 0.7724609375
                },
                {
                    "start": 277.03,
                    "end": 277.37,
                    "word": " to",
                    "probability": 0.89990234375
                },
                {
                    "start": 277.37,
                    "end": 277.81,
                    "word": " users",
                    "probability": 0.82958984375
                },
                {
                    "start": 277.81,
                    "end": 278.33,
                    "word": ".",
                    "probability": 0.93310546875
                },
                {
                    "start": 278.33,
                    "end": 278.43,
                    "word": " So",
                    "probability": 0.83837890625
                },
                {
                    "start": 278.43,
                    "end": 278.51,
                    "word": ",",
                    "probability": 0.1641845703125
                },
                {
                    "start": 278.51,
                    "end": 278.59,
                    "word": " we",
                    "probability": 0.90625
                },
                {
                    "start": 278.59,
                    "end": 278.79,
                    "word": " also",
                    "probability": 0.82421875
                }
            ]
        },
        {
            "id": 40,
            "seek": 29296,
            "start": 278.79,
            "end": 287.33,
            "text": " see deep fakes in the text generation space where it purports to be a real email. Or we see deep",
            "tokens": [
                50680,
                536,
                2452,
                283,
                3419,
                294,
                264,
                2487,
                5125,
                1901,
                689,
                309,
                1864,
                17845,
                281,
                312,
                257,
                957,
                3796,
                13,
                1610,
                321,
                536,
                2452,
                51108
            ],
            "temperature": 0.0,
            "avg_logprob": -0.19098395914644808,
            "compression_ratio": 1.553763440860215,
            "no_speech_prob": 0.0002378225326538086,
            "words": [
                {
                    "start": 278.79,
                    "end": 279.11,
                    "word": " see",
                    "probability": 0.80615234375
                },
                {
                    "start": 279.11,
                    "end": 279.39,
                    "word": " deep",
                    "probability": 0.91259765625
                },
                {
                    "start": 279.39,
                    "end": 279.75,
                    "word": " fakes",
                    "probability": 0.708740234375
                },
                {
                    "start": 279.75,
                    "end": 280.05,
                    "word": " in",
                    "probability": 0.89892578125
                },
                {
                    "start": 280.05,
                    "end": 280.21,
                    "word": " the",
                    "probability": 0.8125
                },
                {
                    "start": 280.21,
                    "end": 280.55,
                    "word": " text",
                    "probability": 0.91748046875
                },
                {
                    "start": 280.55,
                    "end": 281.19,
                    "word": " generation",
                    "probability": 0.72607421875
                },
                {
                    "start": 281.19,
                    "end": 281.89,
                    "word": " space",
                    "probability": 0.8125
                },
                {
                    "start": 281.89,
                    "end": 282.21,
                    "word": " where",
                    "probability": 0.55322265625
                },
                {
                    "start": 282.21,
                    "end": 282.43,
                    "word": " it",
                    "probability": 0.86083984375
                },
                {
                    "start": 282.43,
                    "end": 282.89,
                    "word": " purports",
                    "probability": 0.888671875
                },
                {
                    "start": 282.89,
                    "end": 283.15,
                    "word": " to",
                    "probability": 0.90625
                },
                {
                    "start": 283.15,
                    "end": 283.43,
                    "word": " be",
                    "probability": 0.90673828125
                },
                {
                    "start": 283.67,
                    "end": 284.17,
                    "word": " a real",
                    "probability": 0.86865234375
                },
                {
                    "start": 284.17,
                    "end": 284.71,
                    "word": " email",
                    "probability": 0.84765625
                },
                {
                    "start": 284.71,
                    "end": 285.27,
                    "word": ".",
                    "probability": 0.185791015625
                },
                {
                    "start": 285.27,
                    "end": 285.49,
                    "word": " Or",
                    "probability": 0.87060546875
                },
                {
                    "start": 285.49,
                    "end": 286.47,
                    "word": " we",
                    "probability": 0.70849609375
                },
                {
                    "start": 286.47,
                    "end": 286.83,
                    "word": " see",
                    "probability": 0.80859375
                },
                {
                    "start": 286.83,
                    "end": 287.33,
                    "word": " deep",
                    "probability": 0.90283203125
                }
            ]
        },
        {
            "id": 41,
            "seek": 29296,
            "start": 287.33,
            "end": 296.69,
            "text": " fake writing which mimics the style of famous authors. We see deep fake audio recordings which",
            "tokens": [
                51108,
                7592,
                3579,
                597,
                12247,
                1167,
                264,
                3758,
                295,
                4618,
                16552,
                13,
                492,
                536,
                2452,
                7592,
                6278,
                25162,
                597,
                51572
            ],
            "temperature": 0.0,
            "avg_logprob": -0.19098395914644808,
            "compression_ratio": 1.553763440860215,
            "no_speech_prob": 0.0002378225326538086,
            "words": [
                {
                    "start": 287.33,
                    "end": 287.73,
                    "word": " fake",
                    "probability": 0.6748046875
                },
                {
                    "start": 287.73,
                    "end": 288.43,
                    "word": " writing",
                    "probability": 0.87744140625
                },
                {
                    "start": 288.43,
                    "end": 288.93,
                    "word": " which",
                    "probability": 0.34423828125
                },
                {
                    "start": 288.93,
                    "end": 289.53,
                    "word": " mimics",
                    "probability": 0.953369140625
                },
                {
                    "start": 289.53,
                    "end": 289.85,
                    "word": " the",
                    "probability": 0.81787109375
                },
                {
                    "start": 289.85,
                    "end": 290.43,
                    "word": " style",
                    "probability": 0.78955078125
                },
                {
                    "start": 290.43,
                    "end": 291.27,
                    "word": " of",
                    "probability": 0.89990234375
                },
                {
                    "start": 291.27,
                    "end": 293.21,
                    "word": " famous",
                    "probability": 0.85888671875
                },
                {
                    "start": 293.21,
                    "end": 293.63,
                    "word": " authors",
                    "probability": 0.77294921875
                },
                {
                    "start": 293.63,
                    "end": 293.99,
                    "word": ".",
                    "probability": 0.98095703125
                },
                {
                    "start": 293.99,
                    "end": 294.07,
                    "word": " We",
                    "probability": 0.9033203125
                },
                {
                    "start": 294.07,
                    "end": 294.35,
                    "word": " see",
                    "probability": 0.81103515625
                },
                {
                    "start": 294.35,
                    "end": 294.77,
                    "word": " deep",
                    "probability": 0.92333984375
                },
                {
                    "start": 294.77,
                    "end": 295.19,
                    "word": " fake",
                    "probability": 0.7763671875
                },
                {
                    "start": 295.19,
                    "end": 295.65,
                    "word": " audio",
                    "probability": 0.9189453125
                },
                {
                    "start": 295.65,
                    "end": 296.25,
                    "word": " recordings",
                    "probability": 0.8896484375
                },
                {
                    "start": 296.25,
                    "end": 296.69,
                    "word": " which",
                    "probability": 0.77978515625
                }
            ]
        },
        {
            "id": 42,
            "seek": 32068,
            "start": 297.25,
            "end": 302.77,
            "text": " sound like the person who they are trained to sound like. Now, there's a piece of commercial",
            "tokens": [
                50392,
                1626,
                411,
                264,
                954,
                567,
                436,
                366,
                8895,
                281,
                1626,
                411,
                13,
                823,
                11,
                456,
                311,
                257,
                2522,
                295,
                6841,
                50668
            ],
            "temperature": 0.0,
            "avg_logprob": -0.20519770985963393,
            "compression_ratio": 1.6422413793103448,
            "no_speech_prob": 0.00453948974609375,
            "words": [
                {
                    "start": 297.25,
                    "end": 297.63,
                    "word": " sound",
                    "probability": 0.3837890625
                },
                {
                    "start": 297.63,
                    "end": 298.19,
                    "word": " like",
                    "probability": 0.8740234375
                },
                {
                    "start": 298.19,
                    "end": 298.39,
                    "word": " the",
                    "probability": 0.81884765625
                },
                {
                    "start": 298.39,
                    "end": 298.87,
                    "word": " person",
                    "probability": 0.80126953125
                },
                {
                    "start": 298.87,
                    "end": 299.25,
                    "word": " who",
                    "probability": 0.77490234375
                },
                {
                    "start": 299.25,
                    "end": 299.45,
                    "word": " they",
                    "probability": 0.736328125
                },
                {
                    "start": 299.45,
                    "end": 299.67,
                    "word": " are",
                    "probability": 0.89599609375
                },
                {
                    "start": 299.67,
                    "end": 300.15,
                    "word": " trained",
                    "probability": 0.82275390625
                },
                {
                    "start": 300.15,
                    "end": 300.47,
                    "word": " to",
                    "probability": 0.90673828125
                },
                {
                    "start": 300.47,
                    "end": 300.75,
                    "word": " sound",
                    "probability": 0.8076171875
                },
                {
                    "start": 300.75,
                    "end": 301.19,
                    "word": " like",
                    "probability": 0.8759765625
                },
                {
                    "start": 301.19,
                    "end": 301.67,
                    "word": ".",
                    "probability": 0.9345703125
                },
                {
                    "start": 301.67,
                    "end": 301.81,
                    "word": " Now",
                    "probability": 0.1982421875
                },
                {
                    "start": 301.81,
                    "end": 301.87,
                    "word": ",",
                    "probability": 0.40625
                },
                {
                    "start": 301.87,
                    "end": 302.01,
                    "word": " there's",
                    "probability": 0.913818359375
                },
                {
                    "start": 302.11,
                    "end": 302.29,
                    "word": " a piece",
                    "probability": 0.828125
                },
                {
                    "start": 302.29,
                    "end": 302.43,
                    "word": " of",
                    "probability": 0.90576171875
                },
                {
                    "start": 302.43,
                    "end": 302.77,
                    "word": " commercial",
                    "probability": 0.83837890625
                }
            ]
        },
        {
            "id": 43,
            "seek": 32068,
            "start": 302.77,
            "end": 308.37,
            "text": " software out there called Liarbird which does this really well. And so, you know, they're very",
            "tokens": [
                50668,
                4722,
                484,
                456,
                1219,
                8349,
                289,
                18080,
                597,
                775,
                341,
                534,
                731,
                13,
                400,
                370,
                11,
                291,
                458,
                11,
                436,
                434,
                588,
                50944
            ],
            "temperature": 0.0,
            "avg_logprob": -0.20519770985963393,
            "compression_ratio": 1.6422413793103448,
            "no_speech_prob": 0.00453948974609375,
            "words": [
                {
                    "start": 302.77,
                    "end": 303.21,
                    "word": " software",
                    "probability": 0.8505859375
                },
                {
                    "start": 303.21,
                    "end": 303.39,
                    "word": " out",
                    "probability": 0.8203125
                },
                {
                    "start": 303.39,
                    "end": 303.51,
                    "word": " there",
                    "probability": 0.85107421875
                },
                {
                    "start": 303.51,
                    "end": 303.77,
                    "word": " called",
                    "probability": 0.79248046875
                },
                {
                    "start": 303.77,
                    "end": 304.41,
                    "word": " Liarbird",
                    "probability": 0.6525065104166666
                },
                {
                    "start": 304.41,
                    "end": 305.29,
                    "word": " which",
                    "probability": 0.2369384765625
                },
                {
                    "start": 305.29,
                    "end": 305.53,
                    "word": " does",
                    "probability": 0.92333984375
                },
                {
                    "start": 305.53,
                    "end": 305.75,
                    "word": " this",
                    "probability": 0.8798828125
                },
                {
                    "start": 305.75,
                    "end": 306.03,
                    "word": " really",
                    "probability": 0.77685546875
                },
                {
                    "start": 306.03,
                    "end": 306.37,
                    "word": " well",
                    "probability": 0.8837890625
                },
                {
                    "start": 306.37,
                    "end": 306.71,
                    "word": ".",
                    "probability": 0.720703125
                },
                {
                    "start": 306.71,
                    "end": 306.81,
                    "word": " And",
                    "probability": 0.7802734375
                },
                {
                    "start": 306.81,
                    "end": 307.11,
                    "word": " so",
                    "probability": 0.88427734375
                },
                {
                    "start": 307.11,
                    "end": 307.31,
                    "word": ",",
                    "probability": 0.66748046875
                },
                {
                    "start": 307.31,
                    "end": 307.47,
                    "word": " you",
                    "probability": 0.478515625
                },
                {
                    "start": 307.47,
                    "end": 307.57,
                    "word": " know",
                    "probability": 0.81591796875
                },
                {
                    "start": 307.57,
                    "end": 307.67,
                    "word": ",",
                    "probability": 0.9697265625
                },
                {
                    "start": 307.67,
                    "end": 307.93,
                    "word": " they're",
                    "probability": 0.828125
                },
                {
                    "start": 307.93,
                    "end": 308.37,
                    "word": " very",
                    "probability": 0.73681640625
                }
            ]
        },
        {
            "id": 44,
            "seek": 32068,
            "start": 308.37,
            "end": 316.71,
            "text": " convincing to users. And as a result, you know, they get used widely and they're easy to make,",
            "tokens": [
                50944,
                24823,
                281,
                5022,
                13,
                400,
                382,
                257,
                1874,
                11,
                291,
                458,
                11,
                436,
                483,
                1143,
                13371,
                293,
                436,
                434,
                1858,
                281,
                652,
                11,
                51364
            ],
            "temperature": 0.0,
            "avg_logprob": -0.20519770985963393,
            "compression_ratio": 1.6422413793103448,
            "no_speech_prob": 0.00453948974609375,
            "words": [
                {
                    "start": 308.37,
                    "end": 308.91,
                    "word": " convincing",
                    "probability": 0.79248046875
                },
                {
                    "start": 308.91,
                    "end": 309.21,
                    "word": " to",
                    "probability": 0.908203125
                },
                {
                    "start": 309.21,
                    "end": 309.63,
                    "word": " users",
                    "probability": 0.837890625
                },
                {
                    "start": 309.63,
                    "end": 310.23,
                    "word": ".",
                    "probability": 0.447265625
                },
                {
                    "start": 310.63,
                    "end": 310.99,
                    "word": " And",
                    "probability": 0.86376953125
                },
                {
                    "start": 310.99,
                    "end": 312.01,
                    "word": " as",
                    "probability": 0.79638671875
                },
                {
                    "start": 312.15,
                    "end": 312.49,
                    "word": " a result",
                    "probability": 0.82568359375
                },
                {
                    "start": 312.49,
                    "end": 312.65,
                    "word": ",",
                    "probability": 0.9541015625
                },
                {
                    "start": 312.65,
                    "end": 312.71,
                    "word": " you",
                    "probability": 0.6298828125
                },
                {
                    "start": 312.71,
                    "end": 312.89,
                    "word": " know",
                    "probability": 0.81591796875
                },
                {
                    "start": 312.89,
                    "end": 313.33,
                    "word": ",",
                    "probability": 0.97021484375
                },
                {
                    "start": 313.33,
                    "end": 313.65,
                    "word": " they",
                    "probability": 0.72119140625
                },
                {
                    "start": 313.65,
                    "end": 313.87,
                    "word": " get",
                    "probability": 0.904296875
                },
                {
                    "start": 313.87,
                    "end": 314.21,
                    "word": " used",
                    "probability": 0.7958984375
                },
                {
                    "start": 314.21,
                    "end": 315.09,
                    "word": " widely",
                    "probability": 0.86376953125
                },
                {
                    "start": 315.09,
                    "end": 315.39,
                    "word": " and",
                    "probability": 0.8037109375
                },
                {
                    "start": 315.39,
                    "end": 315.55,
                    "word": " they're",
                    "probability": 0.839599609375
                },
                {
                    "start": 315.55,
                    "end": 315.81,
                    "word": " easy",
                    "probability": 0.80908203125
                },
                {
                    "start": 315.81,
                    "end": 316.01,
                    "word": " to",
                    "probability": 0.90380859375
                },
                {
                    "start": 316.01,
                    "end": 316.45,
                    "word": " make",
                    "probability": 0.8740234375
                },
                {
                    "start": 316.45,
                    "end": 316.71,
                    "word": ",",
                    "probability": 0.83154296875
                }
            ]
        },
        {
            "id": 45,
            "seek": 32068,
            "start": 316.71,
            "end": 324.41,
            "text": " right? That's the other thing, is they don't take as much time as finding and crafting an image or",
            "tokens": [
                51364,
                558,
                30,
                663,
                311,
                264,
                661,
                551,
                11,
                307,
                436,
                500,
                380,
                747,
                382,
                709,
                565,
                382,
                5006,
                293,
                29048,
                364,
                3256,
                420,
                51740
            ],
            "temperature": 0.0,
            "avg_logprob": -0.20519770985963393,
            "compression_ratio": 1.6422413793103448,
            "no_speech_prob": 0.00453948974609375,
            "words": [
                {
                    "start": 316.71,
                    "end": 317.03,
                    "word": " right",
                    "probability": 0.88427734375
                },
                {
                    "start": 317.03,
                    "end": 317.63,
                    "word": "?",
                    "probability": 0.81884765625
                },
                {
                    "start": 317.63,
                    "end": 317.91,
                    "word": " That's",
                    "probability": 0.906494140625
                },
                {
                    "start": 317.91,
                    "end": 318.01,
                    "word": " the",
                    "probability": 0.81591796875
                },
                {
                    "start": 318.01,
                    "end": 318.19,
                    "word": " other",
                    "probability": 0.7783203125
                },
                {
                    "start": 318.19,
                    "end": 318.51,
                    "word": " thing",
                    "probability": 0.8203125
                },
                {
                    "start": 318.51,
                    "end": 318.59,
                    "word": ",",
                    "probability": 0.378173828125
                },
                {
                    "start": 318.59,
                    "end": 318.75,
                    "word": " is",
                    "probability": 0.88671875
                },
                {
                    "start": 318.75,
                    "end": 318.89,
                    "word": " they",
                    "probability": 0.72998046875
                },
                {
                    "start": 318.89,
                    "end": 319.07,
                    "word": " don't",
                    "probability": 0.943603515625
                },
                {
                    "start": 319.07,
                    "end": 319.59,
                    "word": " take",
                    "probability": 0.8046875
                },
                {
                    "start": 319.59,
                    "end": 320.13,
                    "word": " as",
                    "probability": 0.90625
                },
                {
                    "start": 320.13,
                    "end": 320.41,
                    "word": " much",
                    "probability": 0.8115234375
                },
                {
                    "start": 320.41,
                    "end": 320.91,
                    "word": " time",
                    "probability": 0.80517578125
                },
                {
                    "start": 320.91,
                    "end": 321.17,
                    "word": " as",
                    "probability": 0.89306640625
                },
                {
                    "start": 321.17,
                    "end": 322.15,
                    "word": " finding",
                    "probability": 0.732421875
                },
                {
                    "start": 322.15,
                    "end": 322.57,
                    "word": " and",
                    "probability": 0.89453125
                },
                {
                    "start": 322.57,
                    "end": 323.13,
                    "word": " crafting",
                    "probability": 0.755859375
                },
                {
                    "start": 323.13,
                    "end": 323.49,
                    "word": " an",
                    "probability": 0.89892578125
                },
                {
                    "start": 323.49,
                    "end": 323.93,
                    "word": " image",
                    "probability": 0.845703125
                },
                {
                    "start": 323.93,
                    "end": 324.41,
                    "word": " or",
                    "probability": 0.79296875
                }
            ]
        },
        {
            "id": 46,
            "seek": 34972,
            "start": 325.43,
            "end": 332.17,
            "text": " carefully learning a style or mimicking a voice, right? You just kind of train the model and push",
            "tokens": [
                50412,
                7500,
                2539,
                257,
                3758,
                420,
                12247,
                10401,
                257,
                3177,
                11,
                558,
                30,
                509,
                445,
                733,
                295,
                3847,
                264,
                2316,
                293,
                2944,
                50748
            ],
            "temperature": 0.0,
            "avg_logprob": -0.18989158649833834,
            "compression_ratio": 1.56198347107438,
            "no_speech_prob": 2.6285648345947266e-05,
            "words": [
                {
                    "start": 325.43,
                    "end": 325.95,
                    "word": " carefully",
                    "probability": 0.266845703125
                },
                {
                    "start": 325.95,
                    "end": 326.47,
                    "word": " learning",
                    "probability": 0.85302734375
                },
                {
                    "start": 326.77,
                    "end": 327.35,
                    "word": " a style",
                    "probability": 0.806640625
                },
                {
                    "start": 327.35,
                    "end": 327.99,
                    "word": " or",
                    "probability": 0.8095703125
                },
                {
                    "start": 327.99,
                    "end": 328.91,
                    "word": " mimicking",
                    "probability": 0.882080078125
                },
                {
                    "start": 329.13,
                    "end": 329.49,
                    "word": " a voice",
                    "probability": 0.8759765625
                },
                {
                    "start": 329.49,
                    "end": 329.63,
                    "word": ",",
                    "probability": 0.72802734375
                },
                {
                    "start": 329.63,
                    "end": 329.77,
                    "word": " right",
                    "probability": 0.666015625
                },
                {
                    "start": 329.77,
                    "end": 329.81,
                    "word": "?",
                    "probability": 0.73779296875
                },
                {
                    "start": 329.81,
                    "end": 329.93,
                    "word": " You",
                    "probability": 0.89404296875
                },
                {
                    "start": 329.93,
                    "end": 330.11,
                    "word": " just",
                    "probability": 0.875
                },
                {
                    "start": 330.11,
                    "end": 330.31,
                    "word": " kind",
                    "probability": 0.5029296875
                },
                {
                    "start": 330.31,
                    "end": 330.71,
                    "word": " of",
                    "probability": 0.91064453125
                },
                {
                    "start": 330.71,
                    "end": 331.17,
                    "word": " train",
                    "probability": 0.78564453125
                },
                {
                    "start": 331.17,
                    "end": 331.35,
                    "word": " the",
                    "probability": 0.81298828125
                },
                {
                    "start": 331.35,
                    "end": 331.67,
                    "word": " model",
                    "probability": 0.8193359375
                },
                {
                    "start": 331.67,
                    "end": 331.87,
                    "word": " and",
                    "probability": 0.89208984375
                },
                {
                    "start": 331.87,
                    "end": 332.17,
                    "word": " push",
                    "probability": 0.81103515625
                }
            ]
        },
        {
            "id": 47,
            "seek": 34972,
            "start": 332.17,
            "end": 338.41,
            "text": " them out. So, they're easy to make and they're convincing to users. Which means that they can",
            "tokens": [
                50748,
                552,
                484,
                13,
                407,
                11,
                436,
                434,
                1858,
                281,
                652,
                293,
                436,
                434,
                24823,
                281,
                5022,
                13,
                3013,
                1355,
                300,
                436,
                393,
                51060
            ],
            "temperature": 0.0,
            "avg_logprob": -0.18989158649833834,
            "compression_ratio": 1.56198347107438,
            "no_speech_prob": 2.6285648345947266e-05,
            "words": [
                {
                    "start": 332.17,
                    "end": 332.33,
                    "word": " them",
                    "probability": 0.71826171875
                },
                {
                    "start": 332.33,
                    "end": 332.55,
                    "word": " out",
                    "probability": 0.81103515625
                },
                {
                    "start": 332.55,
                    "end": 332.67,
                    "word": ".",
                    "probability": 0.71240234375
                },
                {
                    "start": 332.67,
                    "end": 332.77,
                    "word": " So",
                    "probability": 0.90087890625
                },
                {
                    "start": 332.77,
                    "end": 332.85,
                    "word": ",",
                    "probability": 0.05609130859375
                },
                {
                    "start": 332.85,
                    "end": 332.91,
                    "word": " they're",
                    "probability": 0.82958984375
                },
                {
                    "start": 332.91,
                    "end": 333.23,
                    "word": " easy",
                    "probability": 0.78369140625
                },
                {
                    "start": 333.23,
                    "end": 333.37,
                    "word": " to",
                    "probability": 0.89111328125
                },
                {
                    "start": 333.37,
                    "end": 333.57,
                    "word": " make",
                    "probability": 0.87939453125
                },
                {
                    "start": 333.57,
                    "end": 333.69,
                    "word": " and",
                    "probability": 0.7333984375
                },
                {
                    "start": 333.69,
                    "end": 333.85,
                    "word": " they're",
                    "probability": 0.8330078125
                },
                {
                    "start": 333.85,
                    "end": 334.21,
                    "word": " convincing",
                    "probability": 0.75537109375
                },
                {
                    "start": 334.21,
                    "end": 334.47,
                    "word": " to",
                    "probability": 0.89306640625
                },
                {
                    "start": 334.47,
                    "end": 334.79,
                    "word": " users",
                    "probability": 0.83447265625
                },
                {
                    "start": 334.79,
                    "end": 335.05,
                    "word": ".",
                    "probability": 0.1611328125
                },
                {
                    "start": 335.05,
                    "end": 335.49,
                    "word": " Which",
                    "probability": 0.83935546875
                },
                {
                    "start": 335.49,
                    "end": 335.83,
                    "word": " means",
                    "probability": 0.8798828125
                },
                {
                    "start": 335.83,
                    "end": 336.17,
                    "word": " that",
                    "probability": 0.84130859375
                },
                {
                    "start": 336.17,
                    "end": 338.25,
                    "word": " they",
                    "probability": 0.65966796875
                },
                {
                    "start": 338.25,
                    "end": 338.41,
                    "word": " can",
                    "probability": 0.89599609375
                }
            ]
        },
        {
            "id": 48,
            "seek": 34972,
            "start": 338.41,
            "end": 345.87,
            "text": " have real impact. But that impact is mostly societal, right? So, there was a famous, you know,",
            "tokens": [
                51060,
                362,
                957,
                2712,
                13,
                583,
                300,
                2712,
                307,
                5240,
                33472,
                11,
                558,
                30,
                407,
                11,
                456,
                390,
                257,
                4618,
                11,
                291,
                458,
                11,
                51436
            ],
            "temperature": 0.0,
            "avg_logprob": -0.18989158649833834,
            "compression_ratio": 1.56198347107438,
            "no_speech_prob": 2.6285648345947266e-05,
            "words": [
                {
                    "start": 338.41,
                    "end": 338.61,
                    "word": " have",
                    "probability": 0.88037109375
                },
                {
                    "start": 338.61,
                    "end": 338.99,
                    "word": " real",
                    "probability": 0.7919921875
                },
                {
                    "start": 338.99,
                    "end": 339.61,
                    "word": " impact",
                    "probability": 0.8740234375
                },
                {
                    "start": 339.61,
                    "end": 339.69,
                    "word": ".",
                    "probability": 0.302490234375
                },
                {
                    "start": 339.69,
                    "end": 339.87,
                    "word": " But",
                    "probability": 0.89111328125
                },
                {
                    "start": 339.87,
                    "end": 340.05,
                    "word": " that",
                    "probability": 0.81298828125
                },
                {
                    "start": 340.05,
                    "end": 340.51,
                    "word": " impact",
                    "probability": 0.86474609375
                },
                {
                    "start": 340.51,
                    "end": 340.73,
                    "word": " is",
                    "probability": 0.90869140625
                },
                {
                    "start": 340.73,
                    "end": 341.29,
                    "word": " mostly",
                    "probability": 0.830078125
                },
                {
                    "start": 341.29,
                    "end": 342.63,
                    "word": " societal",
                    "probability": 0.81298828125
                },
                {
                    "start": 342.63,
                    "end": 343.13,
                    "word": ",",
                    "probability": 0.8427734375
                },
                {
                    "start": 343.13,
                    "end": 343.37,
                    "word": " right",
                    "probability": 0.88037109375
                },
                {
                    "start": 343.37,
                    "end": 343.51,
                    "word": "?",
                    "probability": 0.97509765625
                },
                {
                    "start": 343.51,
                    "end": 343.75,
                    "word": " So",
                    "probability": 0.8740234375
                },
                {
                    "start": 343.75,
                    "end": 343.81,
                    "word": ",",
                    "probability": 0.7802734375
                },
                {
                    "start": 343.81,
                    "end": 343.91,
                    "word": " there",
                    "probability": 0.84716796875
                },
                {
                    "start": 343.91,
                    "end": 344.05,
                    "word": " was",
                    "probability": 0.89208984375
                },
                {
                    "start": 344.27,
                    "end": 344.73,
                    "word": " a famous",
                    "probability": 0.86669921875
                },
                {
                    "start": 344.73,
                    "end": 344.95,
                    "word": ",",
                    "probability": 0.468994140625
                },
                {
                    "start": 344.95,
                    "end": 345.45,
                    "word": " you",
                    "probability": 0.75830078125
                },
                {
                    "start": 345.45,
                    "end": 345.59,
                    "word": " know",
                    "probability": 0.8134765625
                },
                {
                    "start": 345.59,
                    "end": 345.87,
                    "word": ",",
                    "probability": 0.98046875
                }
            ]
        },
        {
            "id": 49,
            "seek": 34972,
            "start": 345.87,
            "end": 353.45,
            "text": " semi-viral video of Jordan Peele doing a deep fake where it was an image of President Obama",
            "tokens": [
                51436,
                12909,
                12,
                85,
                35083,
                960,
                295,
                10979,
                430,
                1653,
                306,
                884,
                257,
                2452,
                7592,
                689,
                309,
                390,
                364,
                3256,
                295,
                3117,
                9560,
                51816
            ],
            "temperature": 0.0,
            "avg_logprob": -0.18989158649833834,
            "compression_ratio": 1.56198347107438,
            "no_speech_prob": 2.6285648345947266e-05,
            "words": [
                {
                    "start": 345.87,
                    "end": 346.21,
                    "word": " semi",
                    "probability": 0.60107421875
                },
                {
                    "start": 346.21,
                    "end": 346.67,
                    "word": "-viral",
                    "probability": 0.8631184895833334
                },
                {
                    "start": 346.67,
                    "end": 347.11,
                    "word": " video",
                    "probability": 0.7841796875
                },
                {
                    "start": 347.11,
                    "end": 347.57,
                    "word": " of",
                    "probability": 0.904296875
                },
                {
                    "start": 347.57,
                    "end": 347.99,
                    "word": " Jordan",
                    "probability": 0.8134765625
                },
                {
                    "start": 347.99,
                    "end": 348.51,
                    "word": " Peele",
                    "probability": 0.8943684895833334
                },
                {
                    "start": 348.51,
                    "end": 349.83,
                    "word": " doing",
                    "probability": 0.8427734375
                },
                {
                    "start": 350.03,
                    "end": 350.29,
                    "word": " a deep",
                    "probability": 0.931640625
                },
                {
                    "start": 350.29,
                    "end": 350.79,
                    "word": " fake",
                    "probability": 0.55810546875
                },
                {
                    "start": 350.79,
                    "end": 351.23,
                    "word": " where",
                    "probability": 0.7685546875
                },
                {
                    "start": 351.23,
                    "end": 351.63,
                    "word": " it",
                    "probability": 0.896484375
                },
                {
                    "start": 351.63,
                    "end": 351.81,
                    "word": " was",
                    "probability": 0.89892578125
                },
                {
                    "start": 351.81,
                    "end": 351.99,
                    "word": " an",
                    "probability": 0.9033203125
                },
                {
                    "start": 351.99,
                    "end": 352.23,
                    "word": " image",
                    "probability": 0.83544921875
                },
                {
                    "start": 352.23,
                    "end": 352.37,
                    "word": " of",
                    "probability": 0.90673828125
                },
                {
                    "start": 352.37,
                    "end": 352.81,
                    "word": " President",
                    "probability": 0.79296875
                },
                {
                    "start": 352.81,
                    "end": 353.45,
                    "word": " Obama",
                    "probability": 0.92041015625
                }
            ]
        },
        {
            "id": 50,
            "seek": 37080,
            "start": 354.25,
            "end": 365.03,
            "text": " speaking. And you can look it up. It's really easy to find. And so, like, the risk there is really",
            "tokens": [
                50404,
                4124,
                13,
                400,
                291,
                393,
                574,
                309,
                493,
                13,
                467,
                311,
                534,
                1858,
                281,
                915,
                13,
                400,
                370,
                11,
                411,
                11,
                264,
                3148,
                456,
                307,
                534,
                50940
            ],
            "temperature": 0.0,
            "avg_logprob": -0.1886160671710968,
            "compression_ratio": 1.443609022556391,
            "no_speech_prob": 0.00028252601623535156,
            "words": [
                {
                    "start": 354.25,
                    "end": 354.73,
                    "word": " speaking",
                    "probability": 0.242431640625
                },
                {
                    "start": 354.73,
                    "end": 355.37,
                    "word": ".",
                    "probability": 0.56640625
                },
                {
                    "start": 355.89,
                    "end": 356.41,
                    "word": " And",
                    "probability": 0.802734375
                },
                {
                    "start": 356.41,
                    "end": 356.53,
                    "word": " you",
                    "probability": 0.875
                },
                {
                    "start": 356.53,
                    "end": 356.69,
                    "word": " can",
                    "probability": 0.896484375
                },
                {
                    "start": 356.69,
                    "end": 356.87,
                    "word": " look",
                    "probability": 0.91650390625
                },
                {
                    "start": 356.87,
                    "end": 356.97,
                    "word": " it",
                    "probability": 0.9013671875
                },
                {
                    "start": 356.97,
                    "end": 357.11,
                    "word": " up",
                    "probability": 0.9091796875
                },
                {
                    "start": 357.11,
                    "end": 357.27,
                    "word": ".",
                    "probability": 0.320068359375
                },
                {
                    "start": 357.27,
                    "end": 357.37,
                    "word": " It's",
                    "probability": 0.949951171875
                },
                {
                    "start": 357.37,
                    "end": 357.79,
                    "word": " really",
                    "probability": 0.7275390625
                },
                {
                    "start": 357.79,
                    "end": 358.05,
                    "word": " easy",
                    "probability": 0.8076171875
                },
                {
                    "start": 358.05,
                    "end": 358.23,
                    "word": " to",
                    "probability": 0.90673828125
                },
                {
                    "start": 358.23,
                    "end": 358.51,
                    "word": " find",
                    "probability": 0.8046875
                },
                {
                    "start": 358.51,
                    "end": 359.15,
                    "word": ".",
                    "probability": 0.95947265625
                },
                {
                    "start": 360.85,
                    "end": 361.15,
                    "word": " And",
                    "probability": 0.77880859375
                },
                {
                    "start": 361.15,
                    "end": 361.59,
                    "word": " so",
                    "probability": 0.87158203125
                },
                {
                    "start": 361.59,
                    "end": 361.73,
                    "word": ",",
                    "probability": 0.292724609375
                },
                {
                    "start": 361.73,
                    "end": 361.99,
                    "word": " like",
                    "probability": 0.63037109375
                },
                {
                    "start": 361.99,
                    "end": 362.45,
                    "word": ",",
                    "probability": 0.900390625
                },
                {
                    "start": 362.45,
                    "end": 362.63,
                    "word": " the",
                    "probability": 0.81494140625
                },
                {
                    "start": 362.63,
                    "end": 363.33,
                    "word": " risk",
                    "probability": 0.81298828125
                },
                {
                    "start": 363.33,
                    "end": 363.85,
                    "word": " there",
                    "probability": 0.85693359375
                },
                {
                    "start": 363.85,
                    "end": 364.57,
                    "word": " is",
                    "probability": 0.8759765625
                },
                {
                    "start": 364.57,
                    "end": 365.03,
                    "word": " really",
                    "probability": 0.76025390625
                }
            ]
        },
        {
            "id": 51,
            "seek": 37080,
            "start": 365.69,
            "end": 374.53,
            "text": " that broad swaths of people will believe that a doctored or deep faked video is real. And so,",
            "tokens": [
                50976,
                300,
                4152,
                1693,
                998,
                82,
                295,
                561,
                486,
                1697,
                300,
                257,
                17112,
                2769,
                420,
                2452,
                283,
                7301,
                960,
                307,
                957,
                13,
                400,
                370,
                11,
                51408
            ],
            "temperature": 0.0,
            "avg_logprob": -0.1886160671710968,
            "compression_ratio": 1.443609022556391,
            "no_speech_prob": 0.00028252601623535156,
            "words": [
                {
                    "start": 365.69,
                    "end": 366.15,
                    "word": " that",
                    "probability": 0.869140625
                },
                {
                    "start": 366.15,
                    "end": 367.59,
                    "word": " broad",
                    "probability": 0.76904296875
                },
                {
                    "start": 367.59,
                    "end": 368.33,
                    "word": " swaths",
                    "probability": 0.8816731770833334
                },
                {
                    "start": 368.33,
                    "end": 368.49,
                    "word": " of",
                    "probability": 0.9033203125
                },
                {
                    "start": 368.49,
                    "end": 368.81,
                    "word": " people",
                    "probability": 0.92724609375
                },
                {
                    "start": 368.81,
                    "end": 369.09,
                    "word": " will",
                    "probability": 0.8134765625
                },
                {
                    "start": 369.09,
                    "end": 369.69,
                    "word": " believe",
                    "probability": 0.73291015625
                },
                {
                    "start": 369.69,
                    "end": 370.01,
                    "word": " that",
                    "probability": 0.87158203125
                },
                {
                    "start": 370.31,
                    "end": 371.01,
                    "word": " a doctored",
                    "probability": 0.920654296875
                },
                {
                    "start": 371.01,
                    "end": 371.33,
                    "word": " or",
                    "probability": 0.86279296875
                },
                {
                    "start": 371.33,
                    "end": 371.77,
                    "word": " deep",
                    "probability": 0.87939453125
                },
                {
                    "start": 371.77,
                    "end": 372.15,
                    "word": " faked",
                    "probability": 0.536407470703125
                },
                {
                    "start": 372.15,
                    "end": 372.57,
                    "word": " video",
                    "probability": 0.791015625
                },
                {
                    "start": 372.57,
                    "end": 372.97,
                    "word": " is",
                    "probability": 0.90625
                },
                {
                    "start": 372.97,
                    "end": 373.41,
                    "word": " real",
                    "probability": 0.892578125
                },
                {
                    "start": 373.41,
                    "end": 373.99,
                    "word": ".",
                    "probability": 0.8154296875
                },
                {
                    "start": 373.99,
                    "end": 374.15,
                    "word": " And",
                    "probability": 0.8623046875
                },
                {
                    "start": 374.15,
                    "end": 374.39,
                    "word": " so",
                    "probability": 0.89794921875
                },
                {
                    "start": 374.39,
                    "end": 374.53,
                    "word": ",",
                    "probability": 0.2978515625
                }
            ]
        },
        {
            "id": 52,
            "seek": 39954,
            "start": 374.53,
            "end": 385.37,
            "text": " there's something like President Donald J. Trump, you know, declaring war on North Korea",
            "tokens": [
                50364,
                456,
                311,
                746,
                411,
                3117,
                8632,
                508,
                13,
                3899,
                11,
                291,
                458,
                11,
                40374,
                1516,
                322,
                4067,
                6307,
                50916
            ],
            "temperature": 0.0,
            "avg_logprob": -0.19797584583813493,
            "compression_ratio": 1.6,
            "no_speech_prob": 0.0390625,
            "words": [
                {
                    "start": 374.53,
                    "end": 374.89,
                    "word": " there's",
                    "probability": 0.43927001953125
                },
                {
                    "start": 374.89,
                    "end": 375.23,
                    "word": " something",
                    "probability": 0.75244140625
                },
                {
                    "start": 375.23,
                    "end": 375.89,
                    "word": " like",
                    "probability": 0.8671875
                },
                {
                    "start": 375.89,
                    "end": 378.31,
                    "word": " President",
                    "probability": 0.5400390625
                },
                {
                    "start": 378.31,
                    "end": 379.05,
                    "word": " Donald",
                    "probability": 0.84716796875
                },
                {
                    "start": 379.05,
                    "end": 379.61,
                    "word": " J",
                    "probability": 0.98095703125
                },
                {
                    "start": 379.61,
                    "end": 379.81,
                    "word": ".",
                    "probability": 0.68896484375
                },
                {
                    "start": 379.81,
                    "end": 380.07,
                    "word": " Trump",
                    "probability": 0.83935546875
                },
                {
                    "start": 380.07,
                    "end": 380.67,
                    "word": ",",
                    "probability": 0.337646484375
                },
                {
                    "start": 380.67,
                    "end": 382.41,
                    "word": " you",
                    "probability": 0.76416015625
                },
                {
                    "start": 382.41,
                    "end": 382.65,
                    "word": " know",
                    "probability": 0.81298828125
                },
                {
                    "start": 382.65,
                    "end": 383.27,
                    "word": ",",
                    "probability": 0.95556640625
                },
                {
                    "start": 383.27,
                    "end": 383.87,
                    "word": " declaring",
                    "probability": 0.7734375
                },
                {
                    "start": 383.87,
                    "end": 384.29,
                    "word": " war",
                    "probability": 0.8896484375
                },
                {
                    "start": 384.29,
                    "end": 384.53,
                    "word": " on",
                    "probability": 0.9111328125
                },
                {
                    "start": 384.53,
                    "end": 384.83,
                    "word": " North",
                    "probability": 0.8720703125
                },
                {
                    "start": 384.83,
                    "end": 385.37,
                    "word": " Korea",
                    "probability": 0.80908203125
                }
            ]
        },
        {
            "id": 53,
            "seek": 39954,
            "start": 385.37,
            "end": 391.37,
            "text": " on a hot mic or suggesting that he might invade North Korea on a hot mic. That could be really",
            "tokens": [
                50916,
                322,
                257,
                2368,
                3123,
                420,
                18094,
                300,
                415,
                1062,
                39171,
                4067,
                6307,
                322,
                257,
                2368,
                3123,
                13,
                663,
                727,
                312,
                534,
                51200
            ],
            "temperature": 0.0,
            "avg_logprob": -0.19797584583813493,
            "compression_ratio": 1.6,
            "no_speech_prob": 0.0390625,
            "words": [
                {
                    "start": 385.37,
                    "end": 385.83,
                    "word": " on",
                    "probability": 0.8779296875
                },
                {
                    "start": 385.97,
                    "end": 386.21,
                    "word": " a hot",
                    "probability": 0.85693359375
                },
                {
                    "start": 386.21,
                    "end": 386.55,
                    "word": " mic",
                    "probability": 0.8759765625
                },
                {
                    "start": 386.55,
                    "end": 386.91,
                    "word": " or",
                    "probability": 0.5712890625
                },
                {
                    "start": 386.91,
                    "end": 387.41,
                    "word": " suggesting",
                    "probability": 0.59130859375
                },
                {
                    "start": 387.41,
                    "end": 387.65,
                    "word": " that",
                    "probability": 0.87939453125
                },
                {
                    "start": 387.65,
                    "end": 387.81,
                    "word": " he",
                    "probability": 0.90625
                },
                {
                    "start": 387.81,
                    "end": 388.11,
                    "word": " might",
                    "probability": 0.806640625
                },
                {
                    "start": 388.11,
                    "end": 388.63,
                    "word": " invade",
                    "probability": 0.91552734375
                },
                {
                    "start": 388.63,
                    "end": 388.91,
                    "word": " North",
                    "probability": 0.8662109375
                },
                {
                    "start": 388.91,
                    "end": 389.19,
                    "word": " Korea",
                    "probability": 0.78125
                },
                {
                    "start": 389.19,
                    "end": 389.49,
                    "word": " on",
                    "probability": 0.90087890625
                },
                {
                    "start": 389.59,
                    "end": 389.75,
                    "word": " a hot",
                    "probability": 0.9052734375
                },
                {
                    "start": 389.75,
                    "end": 390.07,
                    "word": " mic",
                    "probability": 0.8896484375
                },
                {
                    "start": 390.07,
                    "end": 390.53,
                    "word": ".",
                    "probability": 0.3447265625
                },
                {
                    "start": 390.53,
                    "end": 390.85,
                    "word": " That",
                    "probability": 0.7919921875
                },
                {
                    "start": 390.85,
                    "end": 390.99,
                    "word": " could",
                    "probability": 0.78759765625
                },
                {
                    "start": 390.99,
                    "end": 391.11,
                    "word": " be",
                    "probability": 0.90234375
                },
                {
                    "start": 391.11,
                    "end": 391.37,
                    "word": " really",
                    "probability": 0.77392578125
                }
            ]
        },
        {
            "id": 54,
            "seek": 39954,
            "start": 391.37,
            "end": 396.73,
            "text": " convincing to people. And so, that's the real risk of deep fakes, is that they are a powerful",
            "tokens": [
                51200,
                24823,
                281,
                561,
                13,
                400,
                370,
                11,
                300,
                311,
                264,
                957,
                3148,
                295,
                2452,
                283,
                3419,
                11,
                307,
                300,
                436,
                366,
                257,
                4005,
                51484
            ],
            "temperature": 0.0,
            "avg_logprob": -0.19797584583813493,
            "compression_ratio": 1.6,
            "no_speech_prob": 0.0390625,
            "words": [
                {
                    "start": 391.37,
                    "end": 391.75,
                    "word": " convincing",
                    "probability": 0.79345703125
                },
                {
                    "start": 391.75,
                    "end": 391.97,
                    "word": " to",
                    "probability": 0.9052734375
                },
                {
                    "start": 391.97,
                    "end": 392.27,
                    "word": " people",
                    "probability": 0.92578125
                },
                {
                    "start": 392.27,
                    "end": 392.45,
                    "word": ".",
                    "probability": 0.66357421875
                },
                {
                    "start": 392.45,
                    "end": 392.49,
                    "word": " And",
                    "probability": 0.8193359375
                },
                {
                    "start": 392.49,
                    "end": 392.65,
                    "word": " so",
                    "probability": 0.89453125
                },
                {
                    "start": 392.65,
                    "end": 392.69,
                    "word": ",",
                    "probability": 0.09246826171875
                },
                {
                    "start": 392.69,
                    "end": 392.91,
                    "word": " that's",
                    "probability": 0.937255859375
                },
                {
                    "start": 392.91,
                    "end": 393.03,
                    "word": " the",
                    "probability": 0.81689453125
                },
                {
                    "start": 393.03,
                    "end": 393.33,
                    "word": " real",
                    "probability": 0.86376953125
                },
                {
                    "start": 393.33,
                    "end": 393.87,
                    "word": " risk",
                    "probability": 0.8203125
                },
                {
                    "start": 393.87,
                    "end": 394.07,
                    "word": " of",
                    "probability": 0.90185546875
                },
                {
                    "start": 394.07,
                    "end": 394.35,
                    "word": " deep",
                    "probability": 0.88232421875
                },
                {
                    "start": 394.35,
                    "end": 394.83,
                    "word": " fakes",
                    "probability": 0.634765625
                },
                {
                    "start": 394.83,
                    "end": 394.93,
                    "word": ",",
                    "probability": 0.3291015625
                },
                {
                    "start": 394.93,
                    "end": 395.11,
                    "word": " is",
                    "probability": 0.85400390625
                },
                {
                    "start": 395.11,
                    "end": 395.33,
                    "word": " that",
                    "probability": 0.87255859375
                },
                {
                    "start": 395.33,
                    "end": 395.51,
                    "word": " they",
                    "probability": 0.73095703125
                },
                {
                    "start": 395.51,
                    "end": 395.65,
                    "word": " are",
                    "probability": 0.8701171875
                },
                {
                    "start": 395.95,
                    "end": 396.73,
                    "word": " a powerful",
                    "probability": 0.84716796875
                }
            ]
        },
        {
            "id": 55,
            "seek": 39954,
            "start": 396.73,
            "end": 403.27,
            "text": " disinformation tool which is powered by artificial intelligence. And so, you know,",
            "tokens": [
                51484,
                717,
                20941,
                2290,
                597,
                307,
                17786,
                538,
                11677,
                7599,
                13,
                400,
                370,
                11,
                291,
                458,
                11,
                51796
            ],
            "temperature": 0.0,
            "avg_logprob": -0.19797584583813493,
            "compression_ratio": 1.6,
            "no_speech_prob": 0.0390625,
            "words": [
                {
                    "start": 396.73,
                    "end": 397.87,
                    "word": " disinformation",
                    "probability": 0.864501953125
                },
                {
                    "start": 397.87,
                    "end": 398.71,
                    "word": " tool",
                    "probability": 0.88232421875
                },
                {
                    "start": 398.71,
                    "end": 399.41,
                    "word": " which",
                    "probability": 0.475341796875
                },
                {
                    "start": 399.41,
                    "end": 399.61,
                    "word": " is",
                    "probability": 0.904296875
                },
                {
                    "start": 399.61,
                    "end": 399.99,
                    "word": " powered",
                    "probability": 0.65380859375
                },
                {
                    "start": 399.99,
                    "end": 400.27,
                    "word": " by",
                    "probability": 0.904296875
                },
                {
                    "start": 400.27,
                    "end": 400.73,
                    "word": " artificial",
                    "probability": 0.796875
                },
                {
                    "start": 400.73,
                    "end": 401.37,
                    "word": " intelligence",
                    "probability": 0.69873046875
                },
                {
                    "start": 401.37,
                    "end": 401.97,
                    "word": ".",
                    "probability": 0.9384765625
                },
                {
                    "start": 402.33,
                    "end": 402.47,
                    "word": " And",
                    "probability": 0.873046875
                },
                {
                    "start": 402.47,
                    "end": 402.83,
                    "word": " so",
                    "probability": 0.89013671875
                },
                {
                    "start": 402.83,
                    "end": 402.99,
                    "word": ",",
                    "probability": 0.93212890625
                },
                {
                    "start": 402.99,
                    "end": 403.17,
                    "word": " you",
                    "probability": 0.84228515625
                },
                {
                    "start": 403.17,
                    "end": 403.23,
                    "word": " know",
                    "probability": 0.81640625
                },
                {
                    "start": 403.23,
                    "end": 403.27,
                    "word": ",",
                    "probability": 0.96533203125
                }
            ]
        },
        {
            "id": 56,
            "seek": 42256,
            "start": 403.27,
            "end": 411.91,
            "text": " our mitigation against deep fakes is kind of, eh. Like, there's very little we can do about deep",
            "tokens": [
                50364,
                527,
                32649,
                1970,
                2452,
                283,
                3419,
                307,
                733,
                295,
                11,
                7670,
                13,
                1743,
                11,
                456,
                311,
                588,
                707,
                321,
                393,
                360,
                466,
                2452,
                50796
            ],
            "temperature": 0.0,
            "avg_logprob": -0.19467906110190056,
            "compression_ratio": 1.5319148936170213,
            "no_speech_prob": 0.00017130374908447266,
            "words": [
                {
                    "start": 403.27,
                    "end": 403.43,
                    "word": " our",
                    "probability": 0.2451171875
                },
                {
                    "start": 403.43,
                    "end": 403.95,
                    "word": " mitigation",
                    "probability": 0.77099609375
                },
                {
                    "start": 403.95,
                    "end": 404.45,
                    "word": " against",
                    "probability": 0.85205078125
                },
                {
                    "start": 404.45,
                    "end": 404.75,
                    "word": " deep",
                    "probability": 0.89404296875
                },
                {
                    "start": 404.75,
                    "end": 405.13,
                    "word": " fakes",
                    "probability": 0.7000732421875
                },
                {
                    "start": 405.13,
                    "end": 405.43,
                    "word": " is",
                    "probability": 0.8720703125
                },
                {
                    "start": 405.43,
                    "end": 405.75,
                    "word": " kind",
                    "probability": 0.73779296875
                },
                {
                    "start": 405.75,
                    "end": 405.93,
                    "word": " of",
                    "probability": 0.90576171875
                },
                {
                    "start": 405.93,
                    "end": 406.51,
                    "word": ",",
                    "probability": 0.267333984375
                },
                {
                    "start": 406.51,
                    "end": 407.39,
                    "word": " eh",
                    "probability": 0.671875
                },
                {
                    "start": 407.39,
                    "end": 408.03,
                    "word": ".",
                    "probability": 0.53662109375
                },
                {
                    "start": 408.51,
                    "end": 408.79,
                    "word": " Like",
                    "probability": 0.619140625
                },
                {
                    "start": 408.79,
                    "end": 409.01,
                    "word": ",",
                    "probability": 0.42333984375
                },
                {
                    "start": 409.01,
                    "end": 409.61,
                    "word": " there's",
                    "probability": 0.914306640625
                },
                {
                    "start": 409.61,
                    "end": 409.79,
                    "word": " very",
                    "probability": 0.7236328125
                },
                {
                    "start": 409.79,
                    "end": 410.03,
                    "word": " little",
                    "probability": 0.810546875
                },
                {
                    "start": 410.03,
                    "end": 410.27,
                    "word": " we",
                    "probability": 0.8994140625
                },
                {
                    "start": 410.27,
                    "end": 410.47,
                    "word": " can",
                    "probability": 0.9013671875
                },
                {
                    "start": 410.47,
                    "end": 410.99,
                    "word": " do",
                    "probability": 0.912109375
                },
                {
                    "start": 410.99,
                    "end": 411.55,
                    "word": " about",
                    "probability": 0.81689453125
                },
                {
                    "start": 411.55,
                    "end": 411.91,
                    "word": " deep",
                    "probability": 0.916015625
                }
            ]
        },
        {
            "id": 57,
            "seek": 42256,
            "start": 411.91,
            "end": 417.97,
            "text": " fakes. They're difficult to detect. They're very convincing to users. They're easy to make. And so,",
            "tokens": [
                50796,
                283,
                3419,
                13,
                814,
                434,
                2252,
                281,
                5531,
                13,
                814,
                434,
                588,
                24823,
                281,
                5022,
                13,
                814,
                434,
                1858,
                281,
                652,
                13,
                400,
                370,
                11,
                51096
            ],
            "temperature": 0.0,
            "avg_logprob": -0.19467906110190056,
            "compression_ratio": 1.5319148936170213,
            "no_speech_prob": 0.00017130374908447266,
            "words": [
                {
                    "start": 411.91,
                    "end": 412.37,
                    "word": " fakes",
                    "probability": 0.8701171875
                },
                {
                    "start": 412.37,
                    "end": 412.67,
                    "word": ".",
                    "probability": 0.93701171875
                },
                {
                    "start": 412.67,
                    "end": 412.93,
                    "word": " They're",
                    "probability": 0.842041015625
                },
                {
                    "start": 412.93,
                    "end": 413.35,
                    "word": " difficult",
                    "probability": 0.85986328125
                },
                {
                    "start": 413.35,
                    "end": 413.61,
                    "word": " to",
                    "probability": 0.912109375
                },
                {
                    "start": 413.61,
                    "end": 414.07,
                    "word": " detect",
                    "probability": 0.91015625
                },
                {
                    "start": 414.07,
                    "end": 414.27,
                    "word": ".",
                    "probability": 0.55126953125
                },
                {
                    "start": 414.27,
                    "end": 414.37,
                    "word": " They're",
                    "probability": 0.845703125
                },
                {
                    "start": 414.37,
                    "end": 414.67,
                    "word": " very",
                    "probability": 0.73876953125
                },
                {
                    "start": 414.67,
                    "end": 415.11,
                    "word": " convincing",
                    "probability": 0.79296875
                },
                {
                    "start": 415.11,
                    "end": 415.45,
                    "word": " to",
                    "probability": 0.9140625
                },
                {
                    "start": 415.45,
                    "end": 415.91,
                    "word": " users",
                    "probability": 0.8349609375
                },
                {
                    "start": 415.91,
                    "end": 416.09,
                    "word": ".",
                    "probability": 0.7744140625
                },
                {
                    "start": 416.09,
                    "end": 416.25,
                    "word": " They're",
                    "probability": 0.85498046875
                },
                {
                    "start": 416.25,
                    "end": 416.63,
                    "word": " easy",
                    "probability": 0.8046875
                },
                {
                    "start": 416.63,
                    "end": 416.83,
                    "word": " to",
                    "probability": 0.9013671875
                },
                {
                    "start": 416.83,
                    "end": 417.15,
                    "word": " make",
                    "probability": 0.88330078125
                },
                {
                    "start": 417.15,
                    "end": 417.47,
                    "word": ".",
                    "probability": 0.87646484375
                },
                {
                    "start": 417.47,
                    "end": 417.67,
                    "word": " And",
                    "probability": 0.8984375
                },
                {
                    "start": 417.67,
                    "end": 417.85,
                    "word": " so",
                    "probability": 0.89892578125
                },
                {
                    "start": 417.85,
                    "end": 417.97,
                    "word": ",",
                    "probability": 0.135498046875
                }
            ]
        },
        {
            "id": 58,
            "seek": 42256,
            "start": 417.97,
            "end": 426.29,
            "text": " really, it's a broader social problem that we need to tackle of having people be inherently",
            "tokens": [
                51096,
                534,
                11,
                309,
                311,
                257,
                13227,
                2093,
                1154,
                300,
                321,
                643,
                281,
                14896,
                295,
                1419,
                561,
                312,
                27993,
                51540
            ],
            "temperature": 0.0,
            "avg_logprob": -0.19467906110190056,
            "compression_ratio": 1.5319148936170213,
            "no_speech_prob": 0.00017130374908447266,
            "words": [
                {
                    "start": 417.97,
                    "end": 418.31,
                    "word": " really",
                    "probability": 0.77783203125
                },
                {
                    "start": 418.31,
                    "end": 418.57,
                    "word": ",",
                    "probability": 0.806640625
                },
                {
                    "start": 418.57,
                    "end": 419.11,
                    "word": " it's",
                    "probability": 0.951904296875
                },
                {
                    "start": 419.31,
                    "end": 419.75,
                    "word": " a broader",
                    "probability": 0.7490234375
                },
                {
                    "start": 419.75,
                    "end": 420.45,
                    "word": " social",
                    "probability": 0.79541015625
                },
                {
                    "start": 420.45,
                    "end": 421.09,
                    "word": " problem",
                    "probability": 0.771484375
                },
                {
                    "start": 421.09,
                    "end": 421.33,
                    "word": " that",
                    "probability": 0.87646484375
                },
                {
                    "start": 421.33,
                    "end": 421.45,
                    "word": " we",
                    "probability": 0.904296875
                },
                {
                    "start": 421.45,
                    "end": 421.65,
                    "word": " need",
                    "probability": 0.822265625
                },
                {
                    "start": 421.65,
                    "end": 421.81,
                    "word": " to",
                    "probability": 0.90625
                },
                {
                    "start": 421.81,
                    "end": 422.23,
                    "word": " tackle",
                    "probability": 0.783203125
                },
                {
                    "start": 422.23,
                    "end": 422.99,
                    "word": " of",
                    "probability": 0.76611328125
                },
                {
                    "start": 422.99,
                    "end": 423.65,
                    "word": " having",
                    "probability": 0.79541015625
                },
                {
                    "start": 423.65,
                    "end": 424.19,
                    "word": " people",
                    "probability": 0.92578125
                },
                {
                    "start": 424.19,
                    "end": 425.51,
                    "word": " be",
                    "probability": 0.90234375
                },
                {
                    "start": 425.51,
                    "end": 426.29,
                    "word": " inherently",
                    "probability": 0.82861328125
                }
            ]
        },
        {
            "id": 59,
            "seek": 44852,
            "start": 426.29,
            "end": 435.09,
            "text": " less trustful of information sources and really vet where they get their information from,",
            "tokens": [
                50388,
                1570,
                3361,
                906,
                295,
                1589,
                7139,
                293,
                534,
                12423,
                689,
                436,
                483,
                641,
                1589,
                490,
                11,
                50780
            ],
            "temperature": 0.0,
            "avg_logprob": -0.1845703125,
            "compression_ratio": 1.4734042553191489,
            "no_speech_prob": 0.0002868175506591797,
            "words": [
                {
                    "start": 426.29,
                    "end": 426.97,
                    "word": " less",
                    "probability": 0.208740234375
                },
                {
                    "start": 426.97,
                    "end": 428.27,
                    "word": " trustful",
                    "probability": 0.782470703125
                },
                {
                    "start": 428.27,
                    "end": 429.09,
                    "word": " of",
                    "probability": 0.89892578125
                },
                {
                    "start": 429.09,
                    "end": 429.81,
                    "word": " information",
                    "probability": 0.712890625
                },
                {
                    "start": 429.81,
                    "end": 430.43,
                    "word": " sources",
                    "probability": 0.74560546875
                },
                {
                    "start": 430.43,
                    "end": 431.41,
                    "word": " and",
                    "probability": 0.7353515625
                },
                {
                    "start": 431.41,
                    "end": 431.81,
                    "word": " really",
                    "probability": 0.78125
                },
                {
                    "start": 431.81,
                    "end": 432.45,
                    "word": " vet",
                    "probability": 0.80517578125
                },
                {
                    "start": 432.45,
                    "end": 432.99,
                    "word": " where",
                    "probability": 0.90771484375
                },
                {
                    "start": 432.99,
                    "end": 433.17,
                    "word": " they",
                    "probability": 0.7353515625
                },
                {
                    "start": 433.17,
                    "end": 433.37,
                    "word": " get",
                    "probability": 0.89697265625
                },
                {
                    "start": 433.37,
                    "end": 433.57,
                    "word": " their",
                    "probability": 0.890625
                },
                {
                    "start": 433.57,
                    "end": 434.19,
                    "word": " information",
                    "probability": 0.73193359375
                },
                {
                    "start": 434.19,
                    "end": 434.75,
                    "word": " from",
                    "probability": 0.81640625
                },
                {
                    "start": 434.75,
                    "end": 435.09,
                    "word": ",",
                    "probability": 0.58642578125
                }
            ]
        },
        {
            "id": 60,
            "seek": 44852,
            "start": 435.09,
            "end": 443.37,
            "text": " which is a hard problem. But this isn't an enterprise security threat. This is a social",
            "tokens": [
                50828,
                597,
                307,
                257,
                1152,
                1154,
                13,
                583,
                341,
                1943,
                380,
                364,
                14132,
                3825,
                4734,
                13,
                639,
                307,
                257,
                2093,
                51232
            ],
            "temperature": 0.0,
            "avg_logprob": -0.1845703125,
            "compression_ratio": 1.4734042553191489,
            "no_speech_prob": 0.0002868175506591797,
            "words": [
                {
                    "start": 435.09,
                    "end": 435.89,
                    "word": " which",
                    "probability": 0.912109375
                },
                {
                    "start": 435.89,
                    "end": 436.05,
                    "word": " is",
                    "probability": 0.90380859375
                },
                {
                    "start": 436.27,
                    "end": 436.61,
                    "word": " a hard",
                    "probability": 0.82421875
                },
                {
                    "start": 436.61,
                    "end": 437.19,
                    "word": " problem",
                    "probability": 0.77099609375
                },
                {
                    "start": 437.19,
                    "end": 437.39,
                    "word": ".",
                    "probability": 0.446044921875
                },
                {
                    "start": 437.39,
                    "end": 437.69,
                    "word": " But",
                    "probability": 0.8896484375
                },
                {
                    "start": 437.69,
                    "end": 437.89,
                    "word": " this",
                    "probability": 0.87646484375
                },
                {
                    "start": 437.89,
                    "end": 438.53,
                    "word": " isn't",
                    "probability": 0.933349609375
                },
                {
                    "start": 438.53,
                    "end": 438.75,
                    "word": " an",
                    "probability": 0.8427734375
                },
                {
                    "start": 438.75,
                    "end": 439.51,
                    "word": " enterprise",
                    "probability": 0.78564453125
                },
                {
                    "start": 439.51,
                    "end": 440.43,
                    "word": " security",
                    "probability": 0.919921875
                },
                {
                    "start": 440.43,
                    "end": 441.01,
                    "word": " threat",
                    "probability": 0.86083984375
                },
                {
                    "start": 441.01,
                    "end": 441.39,
                    "word": ".",
                    "probability": 0.80322265625
                },
                {
                    "start": 441.39,
                    "end": 442.13,
                    "word": " This",
                    "probability": 0.81591796875
                },
                {
                    "start": 442.13,
                    "end": 442.29,
                    "word": " is",
                    "probability": 0.9091796875
                },
                {
                    "start": 442.69,
                    "end": 443.37,
                    "word": " a social",
                    "probability": 0.814453125
                }
            ]
        },
        {
            "id": 61,
            "seek": 44852,
            "start": 443.37,
            "end": 452.25,
            "text": " threat. So, moving back to thinking about deployed models, we'll look at adversarial examples. And",
            "tokens": [
                51232,
                4734,
                13,
                407,
                11,
                2684,
                646,
                281,
                1953,
                466,
                17826,
                5245,
                11,
                321,
                603,
                574,
                412,
                17641,
                44745,
                5110,
                13,
                400,
                51656
            ],
            "temperature": 0.0,
            "avg_logprob": -0.1845703125,
            "compression_ratio": 1.4734042553191489,
            "no_speech_prob": 0.0002868175506591797,
            "words": [
                {
                    "start": 443.37,
                    "end": 444.13,
                    "word": " threat",
                    "probability": 0.8408203125
                },
                {
                    "start": 444.13,
                    "end": 445.03,
                    "word": ".",
                    "probability": 0.99658203125
                },
                {
                    "start": 445.03,
                    "end": 445.73,
                    "word": " So",
                    "probability": 0.87548828125
                },
                {
                    "start": 445.73,
                    "end": 445.79,
                    "word": ",",
                    "probability": 0.06890869140625
                },
                {
                    "start": 445.79,
                    "end": 446.01,
                    "word": " moving",
                    "probability": 0.85498046875
                },
                {
                    "start": 446.01,
                    "end": 446.51,
                    "word": " back",
                    "probability": 0.8125
                },
                {
                    "start": 446.51,
                    "end": 447.05,
                    "word": " to",
                    "probability": 0.9033203125
                },
                {
                    "start": 447.05,
                    "end": 447.31,
                    "word": " thinking",
                    "probability": 0.75830078125
                },
                {
                    "start": 447.31,
                    "end": 447.71,
                    "word": " about",
                    "probability": 0.83447265625
                },
                {
                    "start": 447.71,
                    "end": 448.45,
                    "word": " deployed",
                    "probability": 0.748046875
                },
                {
                    "start": 448.45,
                    "end": 449.01,
                    "word": " models",
                    "probability": 0.8056640625
                },
                {
                    "start": 449.01,
                    "end": 449.59,
                    "word": ",",
                    "probability": 0.80908203125
                },
                {
                    "start": 449.59,
                    "end": 450.29,
                    "word": " we'll",
                    "probability": 0.9208984375
                },
                {
                    "start": 450.29,
                    "end": 450.43,
                    "word": " look",
                    "probability": 0.921875
                },
                {
                    "start": 450.43,
                    "end": 450.61,
                    "word": " at",
                    "probability": 0.91064453125
                },
                {
                    "start": 450.61,
                    "end": 451.27,
                    "word": " adversarial",
                    "probability": 0.895263671875
                },
                {
                    "start": 451.27,
                    "end": 451.83,
                    "word": " examples",
                    "probability": 0.79638671875
                },
                {
                    "start": 451.83,
                    "end": 452.15,
                    "word": ".",
                    "probability": 0.7314453125
                },
                {
                    "start": 452.15,
                    "end": 452.25,
                    "word": " And",
                    "probability": 0.857421875
                }
            ]
        },
        {
            "id": 62,
            "seek": 47520,
            "start": 452.25,
            "end": 459.23,
            "text": " this gif over here on the right is from a pretty well-known example that was really,",
            "tokens": [
                50364,
                341,
                290,
                351,
                670,
                510,
                322,
                264,
                558,
                307,
                490,
                257,
                1238,
                731,
                12,
                6861,
                1365,
                300,
                390,
                534,
                11,
                50704
            ],
            "temperature": 0.0,
            "avg_logprob": -0.2512571969251523,
            "compression_ratio": 1.5745614035087718,
            "no_speech_prob": 0.006191253662109375,
            "words": [
                {
                    "start": 452.25,
                    "end": 452.45,
                    "word": " this",
                    "probability": 0.10870361328125
                },
                {
                    "start": 452.45,
                    "end": 452.97,
                    "word": " gif",
                    "probability": 0.5802001953125
                },
                {
                    "start": 452.97,
                    "end": 453.21,
                    "word": " over",
                    "probability": 0.62548828125
                },
                {
                    "start": 453.21,
                    "end": 453.43,
                    "word": " here",
                    "probability": 0.7978515625
                },
                {
                    "start": 453.43,
                    "end": 453.59,
                    "word": " on",
                    "probability": 0.88330078125
                },
                {
                    "start": 453.59,
                    "end": 453.71,
                    "word": " the",
                    "probability": 0.81689453125
                },
                {
                    "start": 453.71,
                    "end": 453.99,
                    "word": " right",
                    "probability": 0.8681640625
                },
                {
                    "start": 453.99,
                    "end": 454.45,
                    "word": " is",
                    "probability": 0.857421875
                },
                {
                    "start": 454.45,
                    "end": 454.83,
                    "word": " from",
                    "probability": 0.80712890625
                },
                {
                    "start": 455.27,
                    "end": 455.69,
                    "word": " a pretty",
                    "probability": 0.84912109375
                },
                {
                    "start": 455.69,
                    "end": 455.95,
                    "word": " well",
                    "probability": 0.87890625
                },
                {
                    "start": 455.95,
                    "end": 456.23,
                    "word": "-known",
                    "probability": 0.65869140625
                },
                {
                    "start": 456.23,
                    "end": 456.81,
                    "word": " example",
                    "probability": 0.92431640625
                },
                {
                    "start": 456.81,
                    "end": 458.65,
                    "word": " that",
                    "probability": 0.6005859375
                },
                {
                    "start": 458.65,
                    "end": 458.77,
                    "word": " was",
                    "probability": 0.884765625
                },
                {
                    "start": 458.77,
                    "end": 459.07,
                    "word": " really",
                    "probability": 0.77294921875
                },
                {
                    "start": 459.07,
                    "end": 459.23,
                    "word": ",",
                    "probability": 0.650390625
                }
            ]
        },
        {
            "id": 63,
            "seek": 47520,
            "start": 459.23,
            "end": 465.27,
            "text": " really hyped up. And it was the turtle rifle paper. Because the Google computer vision API",
            "tokens": [
                50704,
                534,
                43172,
                493,
                13,
                400,
                309,
                390,
                264,
                22866,
                18008,
                3035,
                13,
                1436,
                264,
                3329,
                3820,
                5201,
                9362,
                51016
            ],
            "temperature": 0.0,
            "avg_logprob": -0.2512571969251523,
            "compression_ratio": 1.5745614035087718,
            "no_speech_prob": 0.006191253662109375,
            "words": [
                {
                    "start": 459.23,
                    "end": 459.33,
                    "word": " really",
                    "probability": 0.77978515625
                },
                {
                    "start": 459.33,
                    "end": 459.73,
                    "word": " hyped",
                    "probability": 0.8203125
                },
                {
                    "start": 459.73,
                    "end": 460.09,
                    "word": " up",
                    "probability": 0.8984375
                },
                {
                    "start": 460.09,
                    "end": 460.17,
                    "word": ".",
                    "probability": 0.2646484375
                },
                {
                    "start": 460.17,
                    "end": 460.19,
                    "word": " And",
                    "probability": 0.6337890625
                },
                {
                    "start": 460.19,
                    "end": 460.29,
                    "word": " it",
                    "probability": 0.89501953125
                },
                {
                    "start": 460.29,
                    "end": 460.41,
                    "word": " was",
                    "probability": 0.89453125
                },
                {
                    "start": 460.41,
                    "end": 460.63,
                    "word": " the",
                    "probability": 0.80810546875
                },
                {
                    "start": 460.63,
                    "end": 461.09,
                    "word": " turtle",
                    "probability": 0.4970703125
                },
                {
                    "start": 461.09,
                    "end": 461.51,
                    "word": " rifle",
                    "probability": 0.77294921875
                },
                {
                    "start": 461.51,
                    "end": 462.05,
                    "word": " paper",
                    "probability": 0.796875
                },
                {
                    "start": 462.05,
                    "end": 462.43,
                    "word": ".",
                    "probability": 0.169189453125
                },
                {
                    "start": 462.43,
                    "end": 462.59,
                    "word": " Because",
                    "probability": 0.71728515625
                },
                {
                    "start": 462.59,
                    "end": 463.09,
                    "word": " the",
                    "probability": 0.80615234375
                },
                {
                    "start": 463.09,
                    "end": 463.59,
                    "word": " Google",
                    "probability": 0.8701171875
                },
                {
                    "start": 463.59,
                    "end": 464.13,
                    "word": " computer",
                    "probability": 0.39306640625
                },
                {
                    "start": 464.13,
                    "end": 464.63,
                    "word": " vision",
                    "probability": 0.79345703125
                },
                {
                    "start": 464.63,
                    "end": 465.27,
                    "word": " API",
                    "probability": 0.783203125
                }
            ]
        },
        {
            "id": 64,
            "seek": 47520,
            "start": 466.15,
            "end": 472.57,
            "text": " is what they attacked. And they 3D printed a turtle. And this is pretty obviously a turtle",
            "tokens": [
                51080,
                307,
                437,
                436,
                12692,
                13,
                400,
                436,
                805,
                35,
                13567,
                257,
                22866,
                13,
                400,
                341,
                307,
                1238,
                2745,
                257,
                22866,
                51384
            ],
            "temperature": 0.0,
            "avg_logprob": -0.2512571969251523,
            "compression_ratio": 1.5745614035087718,
            "no_speech_prob": 0.006191253662109375,
            "words": [
                {
                    "start": 466.15,
                    "end": 466.75,
                    "word": " is",
                    "probability": 0.86376953125
                },
                {
                    "start": 466.75,
                    "end": 466.95,
                    "word": " what",
                    "probability": 0.8857421875
                },
                {
                    "start": 466.95,
                    "end": 467.19,
                    "word": " they",
                    "probability": 0.73193359375
                },
                {
                    "start": 467.19,
                    "end": 467.89,
                    "word": " attacked",
                    "probability": 0.64794921875
                },
                {
                    "start": 467.89,
                    "end": 468.27,
                    "word": ".",
                    "probability": 0.7060546875
                },
                {
                    "start": 468.27,
                    "end": 468.49,
                    "word": " And",
                    "probability": 0.88134765625
                },
                {
                    "start": 468.49,
                    "end": 468.65,
                    "word": " they",
                    "probability": 0.7109375
                },
                {
                    "start": 468.65,
                    "end": 469.31,
                    "word": " 3D",
                    "probability": 0.862548828125
                },
                {
                    "start": 469.31,
                    "end": 469.81,
                    "word": " printed",
                    "probability": 0.74560546875
                },
                {
                    "start": 470.57,
                    "end": 470.87,
                    "word": " a turtle",
                    "probability": 0.79541015625
                },
                {
                    "start": 470.87,
                    "end": 471.11,
                    "word": ".",
                    "probability": 0.89013671875
                },
                {
                    "start": 471.11,
                    "end": 471.21,
                    "word": " And",
                    "probability": 0.80322265625
                },
                {
                    "start": 471.21,
                    "end": 471.35,
                    "word": " this",
                    "probability": 0.85791015625
                },
                {
                    "start": 471.35,
                    "end": 471.47,
                    "word": " is",
                    "probability": 0.88720703125
                },
                {
                    "start": 471.47,
                    "end": 471.63,
                    "word": " pretty",
                    "probability": 0.82470703125
                },
                {
                    "start": 471.63,
                    "end": 472.11,
                    "word": " obviously",
                    "probability": 0.73486328125
                },
                {
                    "start": 472.33,
                    "end": 472.57,
                    "word": " a turtle",
                    "probability": 0.796875
                }
            ]
        },
        {
            "id": 65,
            "seek": 47520,
            "start": 472.57,
            "end": 478.93,
            "text": " to anybody watching who isn't powered by the Google Vision API. And if you are an artificial",
            "tokens": [
                51384,
                281,
                4472,
                1976,
                567,
                1943,
                380,
                17786,
                538,
                264,
                3329,
                25170,
                9362,
                13,
                400,
                498,
                291,
                366,
                364,
                11677,
                51708
            ],
            "temperature": 0.0,
            "avg_logprob": -0.2512571969251523,
            "compression_ratio": 1.5745614035087718,
            "no_speech_prob": 0.006191253662109375,
            "words": [
                {
                    "start": 472.57,
                    "end": 472.83,
                    "word": " to",
                    "probability": 0.89990234375
                },
                {
                    "start": 472.83,
                    "end": 473.37,
                    "word": " anybody",
                    "probability": 0.7734375
                },
                {
                    "start": 473.37,
                    "end": 474.29,
                    "word": " watching",
                    "probability": 0.72802734375
                },
                {
                    "start": 474.29,
                    "end": 474.75,
                    "word": " who",
                    "probability": 0.78076171875
                },
                {
                    "start": 474.75,
                    "end": 475.23,
                    "word": " isn't",
                    "probability": 0.9306640625
                },
                {
                    "start": 475.23,
                    "end": 475.61,
                    "word": " powered",
                    "probability": 0.67431640625
                },
                {
                    "start": 475.61,
                    "end": 475.89,
                    "word": " by",
                    "probability": 0.90087890625
                },
                {
                    "start": 475.89,
                    "end": 476.05,
                    "word": " the",
                    "probability": 0.802734375
                },
                {
                    "start": 476.05,
                    "end": 476.25,
                    "word": " Google",
                    "probability": 0.84765625
                },
                {
                    "start": 476.25,
                    "end": 476.63,
                    "word": " Vision",
                    "probability": 0.3046875
                },
                {
                    "start": 476.63,
                    "end": 477.03,
                    "word": " API",
                    "probability": 0.82763671875
                },
                {
                    "start": 477.03,
                    "end": 477.55,
                    "word": ".",
                    "probability": 0.94189453125
                },
                {
                    "start": 477.55,
                    "end": 477.99,
                    "word": " And",
                    "probability": 0.88232421875
                },
                {
                    "start": 477.99,
                    "end": 478.15,
                    "word": " if",
                    "probability": 0.90087890625
                },
                {
                    "start": 478.15,
                    "end": 478.29,
                    "word": " you",
                    "probability": 0.8935546875
                },
                {
                    "start": 478.29,
                    "end": 478.47,
                    "word": " are",
                    "probability": 0.88720703125
                },
                {
                    "start": 478.47,
                    "end": 478.59,
                    "word": " an",
                    "probability": 0.63134765625
                },
                {
                    "start": 478.59,
                    "end": 478.93,
                    "word": " artificial",
                    "probability": 0.80322265625
                }
            ]
        },
        {
            "id": 66,
            "seek": 49848,
            "start": 478.93,
            "end": 485.55,
            "text": " intelligence powered by the Google Vision API, I'm so sorry for the rest of this talk. Because",
            "tokens": [
                50372,
                7599,
                17786,
                538,
                264,
                3329,
                25170,
                9362,
                11,
                286,
                478,
                370,
                2597,
                337,
                264,
                1472,
                295,
                341,
                751,
                13,
                1436,
                50688
            ],
            "temperature": 0.0,
            "avg_logprob": -0.1712328701803129,
            "compression_ratio": 1.4256410256410257,
            "no_speech_prob": 0.011505126953125,
            "words": [
                {
                    "start": 478.93,
                    "end": 479.59,
                    "word": " intelligence",
                    "probability": 0.0281829833984375
                },
                {
                    "start": 479.59,
                    "end": 481.25,
                    "word": " powered",
                    "probability": 0.437255859375
                },
                {
                    "start": 481.25,
                    "end": 481.47,
                    "word": " by",
                    "probability": 0.89794921875
                },
                {
                    "start": 481.47,
                    "end": 481.61,
                    "word": " the",
                    "probability": 0.8037109375
                },
                {
                    "start": 481.61,
                    "end": 481.83,
                    "word": " Google",
                    "probability": 0.77978515625
                },
                {
                    "start": 481.83,
                    "end": 482.17,
                    "word": " Vision",
                    "probability": 0.67919921875
                },
                {
                    "start": 482.17,
                    "end": 482.61,
                    "word": " API",
                    "probability": 0.8212890625
                },
                {
                    "start": 482.61,
                    "end": 482.89,
                    "word": ",",
                    "probability": 0.306396484375
                },
                {
                    "start": 482.89,
                    "end": 483.05,
                    "word": " I'm",
                    "probability": 0.975341796875
                },
                {
                    "start": 483.05,
                    "end": 483.29,
                    "word": " so",
                    "probability": 0.89111328125
                },
                {
                    "start": 483.29,
                    "end": 483.59,
                    "word": " sorry",
                    "probability": 0.73828125
                },
                {
                    "start": 483.59,
                    "end": 483.79,
                    "word": " for",
                    "probability": 0.89111328125
                },
                {
                    "start": 483.79,
                    "end": 483.91,
                    "word": " the",
                    "probability": 0.810546875
                },
                {
                    "start": 483.91,
                    "end": 484.13,
                    "word": " rest",
                    "probability": 0.80078125
                },
                {
                    "start": 484.13,
                    "end": 484.25,
                    "word": " of",
                    "probability": 0.9033203125
                },
                {
                    "start": 484.25,
                    "end": 484.45,
                    "word": " this",
                    "probability": 0.88232421875
                },
                {
                    "start": 484.45,
                    "end": 484.85,
                    "word": " talk",
                    "probability": 0.80419921875
                },
                {
                    "start": 484.85,
                    "end": 485.29,
                    "word": ".",
                    "probability": 0.0916748046875
                },
                {
                    "start": 485.29,
                    "end": 485.55,
                    "word": " Because",
                    "probability": 0.432373046875
                }
            ]
        },
        {
            "id": 67,
            "seek": 49848,
            "start": 485.55,
            "end": 492.47,
            "text": " you're not gonna like it. That, you know, it's an input to the classifier that's specifically",
            "tokens": [
                50688,
                291,
                434,
                406,
                799,
                411,
                309,
                13,
                663,
                11,
                291,
                458,
                11,
                309,
                311,
                364,
                4846,
                281,
                264,
                1508,
                9902,
                300,
                311,
                4682,
                51052
            ],
            "temperature": 0.0,
            "avg_logprob": -0.1712328701803129,
            "compression_ratio": 1.4256410256410257,
            "no_speech_prob": 0.011505126953125,
            "words": [
                {
                    "start": 485.55,
                    "end": 485.65,
                    "word": " you're",
                    "probability": 0.912109375
                },
                {
                    "start": 485.65,
                    "end": 485.79,
                    "word": " not",
                    "probability": 0.89306640625
                },
                {
                    "start": 485.79,
                    "end": 485.95,
                    "word": " gonna",
                    "probability": 0.197265625
                },
                {
                    "start": 485.95,
                    "end": 486.25,
                    "word": " like",
                    "probability": 0.8759765625
                },
                {
                    "start": 486.25,
                    "end": 486.59,
                    "word": " it",
                    "probability": 0.90283203125
                },
                {
                    "start": 486.59,
                    "end": 487.15,
                    "word": ".",
                    "probability": 0.8798828125
                },
                {
                    "start": 488.67,
                    "end": 489.23,
                    "word": " That",
                    "probability": 0.451904296875
                },
                {
                    "start": 489.23,
                    "end": 489.45,
                    "word": ",",
                    "probability": 0.60205078125
                },
                {
                    "start": 489.45,
                    "end": 489.63,
                    "word": " you",
                    "probability": 0.82177734375
                },
                {
                    "start": 489.63,
                    "end": 489.71,
                    "word": " know",
                    "probability": 0.81689453125
                },
                {
                    "start": 489.71,
                    "end": 489.81,
                    "word": ",",
                    "probability": 0.95947265625
                },
                {
                    "start": 489.81,
                    "end": 490.07,
                    "word": " it's",
                    "probability": 0.946044921875
                },
                {
                    "start": 490.07,
                    "end": 490.25,
                    "word": " an",
                    "probability": 0.89013671875
                },
                {
                    "start": 490.25,
                    "end": 490.59,
                    "word": " input",
                    "probability": 0.87646484375
                },
                {
                    "start": 490.59,
                    "end": 490.79,
                    "word": " to",
                    "probability": 0.90185546875
                },
                {
                    "start": 490.79,
                    "end": 490.91,
                    "word": " the",
                    "probability": 0.81103515625
                },
                {
                    "start": 490.91,
                    "end": 491.43,
                    "word": " classifier",
                    "probability": 0.950927734375
                },
                {
                    "start": 491.43,
                    "end": 491.81,
                    "word": " that's",
                    "probability": 0.878662109375
                },
                {
                    "start": 491.81,
                    "end": 492.47,
                    "word": " specifically",
                    "probability": 0.71240234375
                }
            ]
        },
        {
            "id": 68,
            "seek": 49848,
            "start": 492.47,
            "end": 502.21,
            "text": " crafted to force misclassification. And so, this object as a turtle, just a little, like,",
            "tokens": [
                51052,
                36213,
                281,
                3464,
                3346,
                11665,
                3774,
                13,
                400,
                370,
                11,
                341,
                2657,
                382,
                257,
                22866,
                11,
                445,
                257,
                707,
                11,
                411,
                11,
                51524
            ],
            "temperature": 0.0,
            "avg_logprob": -0.1712328701803129,
            "compression_ratio": 1.4256410256410257,
            "no_speech_prob": 0.011505126953125,
            "words": [
                {
                    "start": 492.47,
                    "end": 493.29,
                    "word": " crafted",
                    "probability": 0.87255859375
                },
                {
                    "start": 493.29,
                    "end": 493.89,
                    "word": " to",
                    "probability": 0.9033203125
                },
                {
                    "start": 493.89,
                    "end": 494.29,
                    "word": " force",
                    "probability": 0.92578125
                },
                {
                    "start": 494.29,
                    "end": 495.43,
                    "word": " misclassification",
                    "probability": 0.9539388020833334
                },
                {
                    "start": 495.43,
                    "end": 495.99,
                    "word": ".",
                    "probability": 0.82763671875
                },
                {
                    "start": 496.33,
                    "end": 496.49,
                    "word": " And",
                    "probability": 0.8779296875
                },
                {
                    "start": 496.49,
                    "end": 496.91,
                    "word": " so",
                    "probability": 0.8896484375
                },
                {
                    "start": 496.91,
                    "end": 497.73,
                    "word": ",",
                    "probability": 0.384033203125
                },
                {
                    "start": 497.73,
                    "end": 498.27,
                    "word": " this",
                    "probability": 0.880859375
                },
                {
                    "start": 498.27,
                    "end": 499.33,
                    "word": " object",
                    "probability": 0.88134765625
                },
                {
                    "start": 499.33,
                    "end": 499.93,
                    "word": " as",
                    "probability": 0.68701171875
                },
                {
                    "start": 500.13,
                    "end": 500.41,
                    "word": " a turtle",
                    "probability": 0.802734375
                },
                {
                    "start": 500.41,
                    "end": 500.79,
                    "word": ",",
                    "probability": 0.81982421875
                },
                {
                    "start": 500.79,
                    "end": 501.11,
                    "word": " just",
                    "probability": 0.82421875
                },
                {
                    "start": 501.29,
                    "end": 501.53,
                    "word": " a little",
                    "probability": 0.8056640625
                },
                {
                    "start": 501.53,
                    "end": 501.63,
                    "word": ",",
                    "probability": 0.356689453125
                },
                {
                    "start": 501.63,
                    "end": 501.87,
                    "word": " like",
                    "probability": 0.84521484375
                },
                {
                    "start": 501.87,
                    "end": 502.21,
                    "word": ",",
                    "probability": 0.9169921875
                }
            ]
        },
        {
            "id": 69,
            "seek": 52400,
            "start": 502.21,
            "end": 509.71,
            "text": " toy turtle, right? Being misclassified as a rifle, the risk there is if you have a CCTV system",
            "tokens": [
                50364,
                12058,
                22866,
                11,
                558,
                30,
                8891,
                3346,
                11665,
                2587,
                382,
                257,
                18008,
                11,
                264,
                3148,
                456,
                307,
                498,
                291,
                362,
                257,
                44838,
                1185,
                50740
            ],
            "temperature": 0.0,
            "avg_logprob": -0.20605468874176344,
            "compression_ratio": 1.41,
            "no_speech_prob": 0.04534912109375,
            "words": [
                {
                    "start": 502.21,
                    "end": 502.51,
                    "word": " toy",
                    "probability": 0.0965576171875
                },
                {
                    "start": 502.51,
                    "end": 502.99,
                    "word": " turtle",
                    "probability": 0.7392578125
                },
                {
                    "start": 502.99,
                    "end": 503.21,
                    "word": ",",
                    "probability": 0.60009765625
                },
                {
                    "start": 503.21,
                    "end": 503.43,
                    "word": " right",
                    "probability": 0.78662109375
                },
                {
                    "start": 503.43,
                    "end": 504.13,
                    "word": "?",
                    "probability": 0.83251953125
                },
                {
                    "start": 504.43,
                    "end": 505.13,
                    "word": " Being",
                    "probability": 0.73828125
                },
                {
                    "start": 505.13,
                    "end": 505.83,
                    "word": " misclassified",
                    "probability": 0.9324544270833334
                },
                {
                    "start": 505.83,
                    "end": 505.99,
                    "word": " as",
                    "probability": 0.89794921875
                },
                {
                    "start": 506.11,
                    "end": 506.43,
                    "word": " a rifle",
                    "probability": 0.81005859375
                },
                {
                    "start": 506.43,
                    "end": 506.79,
                    "word": ",",
                    "probability": 0.716796875
                },
                {
                    "start": 506.79,
                    "end": 506.91,
                    "word": " the",
                    "probability": 0.8076171875
                },
                {
                    "start": 506.91,
                    "end": 507.25,
                    "word": " risk",
                    "probability": 0.81494140625
                },
                {
                    "start": 507.25,
                    "end": 507.59,
                    "word": " there",
                    "probability": 0.8525390625
                },
                {
                    "start": 507.59,
                    "end": 507.81,
                    "word": " is",
                    "probability": 0.873046875
                },
                {
                    "start": 507.81,
                    "end": 507.95,
                    "word": " if",
                    "probability": 0.84521484375
                },
                {
                    "start": 507.95,
                    "end": 508.03,
                    "word": " you",
                    "probability": 0.89306640625
                },
                {
                    "start": 508.03,
                    "end": 508.15,
                    "word": " have",
                    "probability": 0.876953125
                },
                {
                    "start": 508.39,
                    "end": 508.87,
                    "word": " a CCTV",
                    "probability": 0.86376953125
                },
                {
                    "start": 508.87,
                    "end": 509.71,
                    "word": " system",
                    "probability": 0.87646484375
                }
            ]
        },
        {
            "id": 70,
            "seek": 52400,
            "start": 509.71,
            "end": 519.81,
            "text": " that's powered by Google Vision API, which seems reasonable to license, that's looking for",
            "tokens": [
                50740,
                300,
                311,
                17786,
                538,
                3329,
                25170,
                9362,
                11,
                597,
                2544,
                10585,
                281,
                10476,
                11,
                300,
                311,
                1237,
                337,
                51240
            ],
            "temperature": 0.0,
            "avg_logprob": -0.20605468874176344,
            "compression_ratio": 1.41,
            "no_speech_prob": 0.04534912109375,
            "words": [
                {
                    "start": 509.71,
                    "end": 510.09,
                    "word": " that's",
                    "probability": 0.908203125
                },
                {
                    "start": 510.09,
                    "end": 510.47,
                    "word": " powered",
                    "probability": 0.69091796875
                },
                {
                    "start": 510.47,
                    "end": 511.11,
                    "word": " by",
                    "probability": 0.89453125
                },
                {
                    "start": 511.11,
                    "end": 515.07,
                    "word": " Google",
                    "probability": 0.433837890625
                },
                {
                    "start": 515.07,
                    "end": 515.45,
                    "word": " Vision",
                    "probability": 0.83349609375
                },
                {
                    "start": 515.45,
                    "end": 515.89,
                    "word": " API",
                    "probability": 0.822265625
                },
                {
                    "start": 515.89,
                    "end": 516.31,
                    "word": ",",
                    "probability": 0.8173828125
                },
                {
                    "start": 516.31,
                    "end": 516.49,
                    "word": " which",
                    "probability": 0.916015625
                },
                {
                    "start": 516.49,
                    "end": 517.05,
                    "word": " seems",
                    "probability": 0.76171875
                },
                {
                    "start": 517.05,
                    "end": 517.49,
                    "word": " reasonable",
                    "probability": 0.77783203125
                },
                {
                    "start": 517.49,
                    "end": 517.79,
                    "word": " to",
                    "probability": 0.900390625
                },
                {
                    "start": 517.79,
                    "end": 518.21,
                    "word": " license",
                    "probability": 0.78515625
                },
                {
                    "start": 518.21,
                    "end": 518.79,
                    "word": ",",
                    "probability": 0.8349609375
                },
                {
                    "start": 518.79,
                    "end": 519.11,
                    "word": " that's",
                    "probability": 0.93359375
                },
                {
                    "start": 519.11,
                    "end": 519.47,
                    "word": " looking",
                    "probability": 0.7890625
                },
                {
                    "start": 519.47,
                    "end": 519.81,
                    "word": " for",
                    "probability": 0.90478515625
                }
            ]
        },
        {
            "id": 71,
            "seek": 52400,
            "start": 519.81,
            "end": 527.73,
            "text": " potential security threats. And, you know, you don't input all of that, like, racist physiognomy",
            "tokens": [
                51240,
                3995,
                3825,
                14909,
                13,
                400,
                11,
                291,
                458,
                11,
                291,
                500,
                380,
                4846,
                439,
                295,
                300,
                11,
                411,
                11,
                16419,
                21265,
                2912,
                8488,
                51640
            ],
            "temperature": 0.0,
            "avg_logprob": -0.20605468874176344,
            "compression_ratio": 1.41,
            "no_speech_prob": 0.04534912109375,
            "words": [
                {
                    "start": 519.81,
                    "end": 520.21,
                    "word": " potential",
                    "probability": 0.83740234375
                },
                {
                    "start": 520.21,
                    "end": 520.95,
                    "word": " security",
                    "probability": 0.93798828125
                },
                {
                    "start": 520.95,
                    "end": 521.61,
                    "word": " threats",
                    "probability": 0.75439453125
                },
                {
                    "start": 521.61,
                    "end": 522.31,
                    "word": ".",
                    "probability": 0.4443359375
                },
                {
                    "start": 522.35,
                    "end": 522.51,
                    "word": " And",
                    "probability": 0.7275390625
                },
                {
                    "start": 522.51,
                    "end": 522.57,
                    "word": ",",
                    "probability": 0.11737060546875
                },
                {
                    "start": 522.57,
                    "end": 522.73,
                    "word": " you",
                    "probability": 0.89306640625
                },
                {
                    "start": 522.73,
                    "end": 522.81,
                    "word": " know",
                    "probability": 0.6728515625
                },
                {
                    "start": 522.81,
                    "end": 522.89,
                    "word": ",",
                    "probability": 0.99072265625
                },
                {
                    "start": 522.89,
                    "end": 523.05,
                    "word": " you",
                    "probability": 0.89599609375
                },
                {
                    "start": 523.05,
                    "end": 523.63,
                    "word": " don't",
                    "probability": 0.943603515625
                },
                {
                    "start": 523.63,
                    "end": 524.19,
                    "word": " input",
                    "probability": 0.8427734375
                },
                {
                    "start": 524.19,
                    "end": 524.49,
                    "word": " all",
                    "probability": 0.90283203125
                },
                {
                    "start": 524.49,
                    "end": 524.65,
                    "word": " of",
                    "probability": 0.890625
                },
                {
                    "start": 524.65,
                    "end": 524.89,
                    "word": " that",
                    "probability": 0.884765625
                },
                {
                    "start": 524.89,
                    "end": 524.89,
                    "word": ",",
                    "probability": 0.139404296875
                },
                {
                    "start": 524.89,
                    "end": 525.21,
                    "word": " like",
                    "probability": 0.86181640625
                },
                {
                    "start": 525.21,
                    "end": 525.81,
                    "word": ",",
                    "probability": 0.9169921875
                },
                {
                    "start": 525.81,
                    "end": 526.39,
                    "word": " racist",
                    "probability": 0.875
                },
                {
                    "start": 526.39,
                    "end": 527.73,
                    "word": " physiognomy",
                    "probability": 0.9557291666666666
                }
            ]
        },
        {
            "id": 72,
            "seek": 54920,
            "start": 528.49,
            "end": 533.85,
            "text": " that we already see in a lot of those systems. But instead just say, you know what? Let's just",
            "tokens": [
                50412,
                300,
                321,
                1217,
                536,
                294,
                257,
                688,
                295,
                729,
                3652,
                13,
                583,
                2602,
                445,
                584,
                11,
                291,
                458,
                437,
                30,
                961,
                311,
                445,
                50664
            ],
            "temperature": 0.0,
            "avg_logprob": -0.19649921301044995,
            "compression_ratio": 1.4583333333333333,
            "no_speech_prob": 0.001674652099609375,
            "words": [
                {
                    "start": 528.49,
                    "end": 529.01,
                    "word": " that",
                    "probability": 0.131591796875
                },
                {
                    "start": 529.01,
                    "end": 529.25,
                    "word": " we",
                    "probability": 0.86474609375
                },
                {
                    "start": 529.25,
                    "end": 529.83,
                    "word": " already",
                    "probability": 0.830078125
                },
                {
                    "start": 529.83,
                    "end": 530.27,
                    "word": " see",
                    "probability": 0.82177734375
                },
                {
                    "start": 530.27,
                    "end": 530.61,
                    "word": " in",
                    "probability": 0.8955078125
                },
                {
                    "start": 530.75,
                    "end": 530.89,
                    "word": " a lot",
                    "probability": 0.8935546875
                },
                {
                    "start": 530.89,
                    "end": 531.01,
                    "word": " of",
                    "probability": 0.90380859375
                },
                {
                    "start": 531.01,
                    "end": 531.23,
                    "word": " those",
                    "probability": 0.83203125
                },
                {
                    "start": 531.23,
                    "end": 531.65,
                    "word": " systems",
                    "probability": 0.82421875
                },
                {
                    "start": 531.65,
                    "end": 532.03,
                    "word": ".",
                    "probability": 0.2890625
                },
                {
                    "start": 532.03,
                    "end": 532.13,
                    "word": " But",
                    "probability": 0.8408203125
                },
                {
                    "start": 532.13,
                    "end": 532.43,
                    "word": " instead",
                    "probability": 0.81396484375
                },
                {
                    "start": 532.43,
                    "end": 532.63,
                    "word": " just",
                    "probability": 0.63134765625
                },
                {
                    "start": 532.63,
                    "end": 532.79,
                    "word": " say",
                    "probability": 0.8740234375
                },
                {
                    "start": 532.79,
                    "end": 532.87,
                    "word": ",",
                    "probability": 0.8037109375
                },
                {
                    "start": 532.87,
                    "end": 532.93,
                    "word": " you",
                    "probability": 0.51806640625
                },
                {
                    "start": 532.93,
                    "end": 533.03,
                    "word": " know",
                    "probability": 0.80615234375
                },
                {
                    "start": 533.03,
                    "end": 533.25,
                    "word": " what",
                    "probability": 0.845703125
                },
                {
                    "start": 533.25,
                    "end": 533.45,
                    "word": "?",
                    "probability": 0.38623046875
                },
                {
                    "start": 533.45,
                    "end": 533.67,
                    "word": " Let's",
                    "probability": 0.93798828125
                },
                {
                    "start": 533.67,
                    "end": 533.85,
                    "word": " just",
                    "probability": 0.88916015625
                }
            ]
        },
        {
            "id": 73,
            "seek": 54920,
            "start": 533.85,
            "end": 542.57,
            "text": " look for weapons. Well, this could falsely trigger those systems. And so, there's a risk there, too.",
            "tokens": [
                50664,
                574,
                337,
                7278,
                13,
                1042,
                11,
                341,
                727,
                16720,
                736,
                7875,
                729,
                3652,
                13,
                400,
                370,
                11,
                456,
                311,
                257,
                3148,
                456,
                11,
                886,
                13,
                51072
            ],
            "temperature": 0.0,
            "avg_logprob": -0.19649921301044995,
            "compression_ratio": 1.4583333333333333,
            "no_speech_prob": 0.001674652099609375,
            "words": [
                {
                    "start": 533.85,
                    "end": 534.03,
                    "word": " look",
                    "probability": 0.92041015625
                },
                {
                    "start": 534.03,
                    "end": 534.25,
                    "word": " for",
                    "probability": 0.900390625
                },
                {
                    "start": 534.25,
                    "end": 535.09,
                    "word": " weapons",
                    "probability": 0.89404296875
                },
                {
                    "start": 535.09,
                    "end": 535.61,
                    "word": ".",
                    "probability": 0.9755859375
                },
                {
                    "start": 536.17,
                    "end": 536.43,
                    "word": " Well",
                    "probability": 0.72314453125
                },
                {
                    "start": 536.43,
                    "end": 536.59,
                    "word": ",",
                    "probability": 0.87451171875
                },
                {
                    "start": 536.59,
                    "end": 536.89,
                    "word": " this",
                    "probability": 0.8828125
                },
                {
                    "start": 536.89,
                    "end": 537.15,
                    "word": " could",
                    "probability": 0.82421875
                },
                {
                    "start": 537.15,
                    "end": 538.11,
                    "word": " falsely",
                    "probability": 0.8251953125
                },
                {
                    "start": 538.11,
                    "end": 538.65,
                    "word": " trigger",
                    "probability": 0.732421875
                },
                {
                    "start": 538.65,
                    "end": 539.37,
                    "word": " those",
                    "probability": 0.84814453125
                },
                {
                    "start": 539.37,
                    "end": 539.85,
                    "word": " systems",
                    "probability": 0.8193359375
                },
                {
                    "start": 539.85,
                    "end": 540.37,
                    "word": ".",
                    "probability": 0.58154296875
                },
                {
                    "start": 540.49,
                    "end": 540.57,
                    "word": " And",
                    "probability": 0.8720703125
                },
                {
                    "start": 540.57,
                    "end": 540.79,
                    "word": " so",
                    "probability": 0.88330078125
                },
                {
                    "start": 540.79,
                    "end": 540.89,
                    "word": ",",
                    "probability": 0.06365966796875
                },
                {
                    "start": 540.89,
                    "end": 540.99,
                    "word": " there's",
                    "probability": 0.919677734375
                },
                {
                    "start": 541.13,
                    "end": 541.39,
                    "word": " a risk",
                    "probability": 0.81201171875
                },
                {
                    "start": 541.39,
                    "end": 541.71,
                    "word": " there",
                    "probability": 0.85400390625
                },
                {
                    "start": 541.71,
                    "end": 541.91,
                    "word": ",",
                    "probability": 0.363037109375
                },
                {
                    "start": 541.91,
                    "end": 542.05,
                    "word": " too",
                    "probability": 0.80126953125
                },
                {
                    "start": 542.05,
                    "end": 542.57,
                    "word": ".",
                    "probability": 0.9765625
                }
            ]
        },
        {
            "id": 74,
            "seek": 54920,
            "start": 543.61,
            "end": 552.93,
            "text": " Correspondingly, if you 3D printed a gun, right? That got misclassified as a turtle,",
            "tokens": [
                51140,
                3925,
                6663,
                12163,
                11,
                498,
                291,
                805,
                35,
                13567,
                257,
                3874,
                11,
                558,
                30,
                663,
                658,
                3346,
                11665,
                2587,
                382,
                257,
                22866,
                11,
                51628
            ],
            "temperature": 0.0,
            "avg_logprob": -0.19649921301044995,
            "compression_ratio": 1.4583333333333333,
            "no_speech_prob": 0.001674652099609375,
            "words": [
                {
                    "start": 543.61,
                    "end": 544.13,
                    "word": " Correspondingly",
                    "probability": 0.6892903645833334
                },
                {
                    "start": 544.13,
                    "end": 544.27,
                    "word": ",",
                    "probability": 0.93896484375
                },
                {
                    "start": 544.27,
                    "end": 544.33,
                    "word": " if",
                    "probability": 0.904296875
                },
                {
                    "start": 544.33,
                    "end": 544.43,
                    "word": " you",
                    "probability": 0.8984375
                },
                {
                    "start": 544.43,
                    "end": 544.89,
                    "word": " 3D",
                    "probability": 0.929443359375
                },
                {
                    "start": 544.89,
                    "end": 545.27,
                    "word": " printed",
                    "probability": 0.6875
                },
                {
                    "start": 545.53,
                    "end": 545.91,
                    "word": " a gun",
                    "probability": 0.8837890625
                },
                {
                    "start": 545.91,
                    "end": 546.37,
                    "word": ",",
                    "probability": 0.62890625
                },
                {
                    "start": 546.37,
                    "end": 547.57,
                    "word": " right",
                    "probability": 0.767578125
                },
                {
                    "start": 547.57,
                    "end": 548.09,
                    "word": "?",
                    "probability": 0.3994140625
                },
                {
                    "start": 548.73,
                    "end": 549.25,
                    "word": " That",
                    "probability": 0.79345703125
                },
                {
                    "start": 549.25,
                    "end": 551.31,
                    "word": " got",
                    "probability": 0.673828125
                },
                {
                    "start": 551.31,
                    "end": 551.97,
                    "word": " misclassified",
                    "probability": 0.9215494791666666
                },
                {
                    "start": 551.97,
                    "end": 552.15,
                    "word": " as",
                    "probability": 0.91064453125
                },
                {
                    "start": 552.31,
                    "end": 552.59,
                    "word": " a turtle",
                    "probability": 0.7783203125
                },
                {
                    "start": 552.59,
                    "end": 552.93,
                    "word": ",",
                    "probability": 0.705078125
                }
            ]
        },
        {
            "id": 75,
            "seek": 57872,
            "start": 552.93,
            "end": 559.55,
            "text": " that would also be a threat, right? So, the impact, really, is that it causes misclassification.",
            "tokens": [
                50368,
                300,
                576,
                611,
                312,
                257,
                4734,
                11,
                558,
                30,
                407,
                11,
                264,
                2712,
                11,
                534,
                11,
                307,
                300,
                309,
                7700,
                3346,
                11665,
                3774,
                13,
                50672
            ],
            "temperature": 0.0,
            "avg_logprob": -0.19353069757160388,
            "compression_ratio": 1.7859922178988328,
            "no_speech_prob": 0.0009851455688476562,
            "words": [
                {
                    "start": 552.93,
                    "end": 553.31,
                    "word": " that",
                    "probability": 0.496826171875
                },
                {
                    "start": 553.31,
                    "end": 553.39,
                    "word": " would",
                    "probability": 0.7939453125
                },
                {
                    "start": 553.39,
                    "end": 553.61,
                    "word": " also",
                    "probability": 0.81591796875
                },
                {
                    "start": 553.61,
                    "end": 553.79,
                    "word": " be",
                    "probability": 0.90673828125
                },
                {
                    "start": 553.91,
                    "end": 554.19,
                    "word": " a threat",
                    "probability": 0.81591796875
                },
                {
                    "start": 554.19,
                    "end": 554.87,
                    "word": ",",
                    "probability": 0.35595703125
                },
                {
                    "start": 554.87,
                    "end": 555.85,
                    "word": " right",
                    "probability": 0.84765625
                },
                {
                    "start": 555.85,
                    "end": 555.91,
                    "word": "?",
                    "probability": 0.74267578125
                },
                {
                    "start": 555.91,
                    "end": 556.09,
                    "word": " So",
                    "probability": 0.88037109375
                },
                {
                    "start": 556.09,
                    "end": 556.15,
                    "word": ",",
                    "probability": 0.14990234375
                },
                {
                    "start": 556.15,
                    "end": 556.23,
                    "word": " the",
                    "probability": 0.81103515625
                },
                {
                    "start": 556.23,
                    "end": 556.73,
                    "word": " impact",
                    "probability": 0.87353515625
                },
                {
                    "start": 556.73,
                    "end": 556.83,
                    "word": ",",
                    "probability": 0.2001953125
                },
                {
                    "start": 556.83,
                    "end": 557.13,
                    "word": " really",
                    "probability": 0.77099609375
                },
                {
                    "start": 557.13,
                    "end": 557.31,
                    "word": ",",
                    "probability": 0.9775390625
                },
                {
                    "start": 557.31,
                    "end": 557.51,
                    "word": " is",
                    "probability": 0.90087890625
                },
                {
                    "start": 557.51,
                    "end": 557.61,
                    "word": " that",
                    "probability": 0.87939453125
                },
                {
                    "start": 557.61,
                    "end": 557.73,
                    "word": " it",
                    "probability": 0.90625
                },
                {
                    "start": 557.73,
                    "end": 557.99,
                    "word": " causes",
                    "probability": 0.7939453125
                },
                {
                    "start": 557.99,
                    "end": 559.03,
                    "word": " misclassification",
                    "probability": 0.9612630208333334
                },
                {
                    "start": 559.03,
                    "end": 559.55,
                    "word": ".",
                    "probability": 0.72998046875
                }
            ]
        },
        {
            "id": 76,
            "seek": 57872,
            "start": 560.21,
            "end": 564.65,
            "text": " And so, that real world security threat and 3D printed examples are really rare,",
            "tokens": [
                50720,
                400,
                370,
                11,
                300,
                957,
                1002,
                3825,
                4734,
                293,
                805,
                35,
                13567,
                5110,
                366,
                534,
                5892,
                11,
                50944
            ],
            "temperature": 0.0,
            "avg_logprob": -0.19353069757160388,
            "compression_ratio": 1.7859922178988328,
            "no_speech_prob": 0.0009851455688476562,
            "words": [
                {
                    "start": 560.21,
                    "end": 560.31,
                    "word": " And",
                    "probability": 0.81640625
                },
                {
                    "start": 560.31,
                    "end": 560.45,
                    "word": " so",
                    "probability": 0.87353515625
                },
                {
                    "start": 560.45,
                    "end": 560.51,
                    "word": ",",
                    "probability": 0.59326171875
                },
                {
                    "start": 560.51,
                    "end": 560.63,
                    "word": " that",
                    "probability": 0.8828125
                },
                {
                    "start": 560.63,
                    "end": 560.91,
                    "word": " real",
                    "probability": 0.8662109375
                },
                {
                    "start": 560.91,
                    "end": 561.23,
                    "word": " world",
                    "probability": 0.381103515625
                },
                {
                    "start": 561.23,
                    "end": 561.67,
                    "word": " security",
                    "probability": 0.92919921875
                },
                {
                    "start": 561.67,
                    "end": 562.05,
                    "word": " threat",
                    "probability": 0.826171875
                },
                {
                    "start": 562.05,
                    "end": 562.25,
                    "word": " and",
                    "probability": 0.341552734375
                },
                {
                    "start": 562.25,
                    "end": 562.61,
                    "word": " 3D",
                    "probability": 0.94677734375
                },
                {
                    "start": 562.61,
                    "end": 562.91,
                    "word": " printed",
                    "probability": 0.71630859375
                },
                {
                    "start": 562.91,
                    "end": 563.37,
                    "word": " examples",
                    "probability": 0.79833984375
                },
                {
                    "start": 563.37,
                    "end": 563.67,
                    "word": " are",
                    "probability": 0.88134765625
                },
                {
                    "start": 563.67,
                    "end": 564.11,
                    "word": " really",
                    "probability": 0.77685546875
                },
                {
                    "start": 564.11,
                    "end": 564.53,
                    "word": " rare",
                    "probability": 0.75439453125
                },
                {
                    "start": 564.53,
                    "end": 564.65,
                    "word": ",",
                    "probability": 0.28076171875
                }
            ]
        },
        {
            "id": 77,
            "seek": 57872,
            "start": 564.65,
            "end": 570.57,
            "text": " and they're really difficult to do. And they don't tend to work in all situations,",
            "tokens": [
                50944,
                293,
                436,
                434,
                534,
                2252,
                281,
                360,
                13,
                400,
                436,
                500,
                380,
                3928,
                281,
                589,
                294,
                439,
                6851,
                11,
                51248
            ],
            "temperature": 0.0,
            "avg_logprob": -0.19353069757160388,
            "compression_ratio": 1.7859922178988328,
            "no_speech_prob": 0.0009851455688476562,
            "words": [
                {
                    "start": 564.65,
                    "end": 564.73,
                    "word": " and",
                    "probability": 0.890625
                },
                {
                    "start": 564.73,
                    "end": 564.87,
                    "word": " they're",
                    "probability": 0.84033203125
                },
                {
                    "start": 564.87,
                    "end": 565.17,
                    "word": " really",
                    "probability": 0.783203125
                },
                {
                    "start": 565.17,
                    "end": 565.63,
                    "word": " difficult",
                    "probability": 0.84912109375
                },
                {
                    "start": 565.63,
                    "end": 565.93,
                    "word": " to",
                    "probability": 0.90673828125
                },
                {
                    "start": 565.93,
                    "end": 566.21,
                    "word": " do",
                    "probability": 0.9072265625
                },
                {
                    "start": 566.21,
                    "end": 566.73,
                    "word": ".",
                    "probability": 0.51904296875
                },
                {
                    "start": 567.73,
                    "end": 568.25,
                    "word": " And",
                    "probability": 0.86572265625
                },
                {
                    "start": 568.25,
                    "end": 568.37,
                    "word": " they",
                    "probability": 0.72705078125
                },
                {
                    "start": 568.37,
                    "end": 568.67,
                    "word": " don't",
                    "probability": 0.946533203125
                },
                {
                    "start": 568.67,
                    "end": 569.01,
                    "word": " tend",
                    "probability": 0.77783203125
                },
                {
                    "start": 569.01,
                    "end": 569.23,
                    "word": " to",
                    "probability": 0.90380859375
                },
                {
                    "start": 569.23,
                    "end": 569.53,
                    "word": " work",
                    "probability": 0.8134765625
                },
                {
                    "start": 569.53,
                    "end": 569.69,
                    "word": " in",
                    "probability": 0.90380859375
                },
                {
                    "start": 569.69,
                    "end": 569.87,
                    "word": " all",
                    "probability": 0.8935546875
                },
                {
                    "start": 569.87,
                    "end": 570.49,
                    "word": " situations",
                    "probability": 0.638671875
                },
                {
                    "start": 570.49,
                    "end": 570.57,
                    "word": ",",
                    "probability": 0.513671875
                }
            ]
        },
        {
            "id": 78,
            "seek": 57872,
            "start": 570.57,
            "end": 576.31,
            "text": " and any researcher in computer vision can explain to you why it's not a real part of the threat",
            "tokens": [
                51248,
                293,
                604,
                21751,
                294,
                3820,
                5201,
                393,
                2903,
                281,
                291,
                983,
                309,
                311,
                406,
                257,
                957,
                644,
                295,
                264,
                4734,
                51532
            ],
            "temperature": 0.0,
            "avg_logprob": -0.19353069757160388,
            "compression_ratio": 1.7859922178988328,
            "no_speech_prob": 0.0009851455688476562,
            "words": [
                {
                    "start": 570.57,
                    "end": 570.83,
                    "word": " and",
                    "probability": 0.8916015625
                },
                {
                    "start": 570.83,
                    "end": 571.17,
                    "word": " any",
                    "probability": 0.81494140625
                },
                {
                    "start": 571.17,
                    "end": 571.73,
                    "word": " researcher",
                    "probability": 0.7978515625
                },
                {
                    "start": 571.73,
                    "end": 571.95,
                    "word": " in",
                    "probability": 0.89208984375
                },
                {
                    "start": 571.95,
                    "end": 572.35,
                    "word": " computer",
                    "probability": 0.6484375
                },
                {
                    "start": 572.35,
                    "end": 572.67,
                    "word": " vision",
                    "probability": 0.7783203125
                },
                {
                    "start": 572.67,
                    "end": 572.87,
                    "word": " can",
                    "probability": 0.88916015625
                },
                {
                    "start": 572.87,
                    "end": 573.19,
                    "word": " explain",
                    "probability": 0.81982421875
                },
                {
                    "start": 573.19,
                    "end": 573.35,
                    "word": " to",
                    "probability": 0.9072265625
                },
                {
                    "start": 573.35,
                    "end": 573.53,
                    "word": " you",
                    "probability": 0.896484375
                },
                {
                    "start": 573.53,
                    "end": 573.89,
                    "word": " why",
                    "probability": 0.80078125
                },
                {
                    "start": 573.89,
                    "end": 574.43,
                    "word": " it's",
                    "probability": 0.947509765625
                },
                {
                    "start": 574.43,
                    "end": 574.71,
                    "word": " not",
                    "probability": 0.89697265625
                },
                {
                    "start": 575.39,
                    "end": 575.57,
                    "word": " a real",
                    "probability": 0.88525390625
                },
                {
                    "start": 575.57,
                    "end": 575.83,
                    "word": " part",
                    "probability": 0.81005859375
                },
                {
                    "start": 575.83,
                    "end": 575.95,
                    "word": " of",
                    "probability": 0.912109375
                },
                {
                    "start": 575.95,
                    "end": 576.09,
                    "word": " the",
                    "probability": 0.8134765625
                },
                {
                    "start": 576.09,
                    "end": 576.31,
                    "word": " threat",
                    "probability": 0.82666015625
                }
            ]
        },
        {
            "id": 79,
            "seek": 57872,
            "start": 576.31,
            "end": 582.45,
            "text": " model. And that's fine. But our threat is that it causes misclassification. As security practitioners,",
            "tokens": [
                51532,
                2316,
                13,
                400,
                300,
                311,
                2489,
                13,
                583,
                527,
                4734,
                307,
                300,
                309,
                7700,
                3346,
                11665,
                3774,
                13,
                1018,
                3825,
                25742,
                11,
                51840
            ],
            "temperature": 0.0,
            "avg_logprob": -0.19353069757160388,
            "compression_ratio": 1.7859922178988328,
            "no_speech_prob": 0.0009851455688476562,
            "words": [
                {
                    "start": 576.31,
                    "end": 576.65,
                    "word": " model",
                    "probability": 0.82080078125
                },
                {
                    "start": 576.65,
                    "end": 576.83,
                    "word": ".",
                    "probability": 0.33740234375
                },
                {
                    "start": 576.83,
                    "end": 576.99,
                    "word": " And",
                    "probability": 0.89208984375
                },
                {
                    "start": 576.99,
                    "end": 577.19,
                    "word": " that's",
                    "probability": 0.940185546875
                },
                {
                    "start": 577.19,
                    "end": 577.43,
                    "word": " fine",
                    "probability": 0.87939453125
                },
                {
                    "start": 577.43,
                    "end": 577.95,
                    "word": ".",
                    "probability": 0.93798828125
                },
                {
                    "start": 578.23,
                    "end": 578.75,
                    "word": " But",
                    "probability": 0.88525390625
                },
                {
                    "start": 578.75,
                    "end": 578.87,
                    "word": " our",
                    "probability": 0.77734375
                },
                {
                    "start": 578.87,
                    "end": 579.03,
                    "word": " threat",
                    "probability": 0.82275390625
                },
                {
                    "start": 579.03,
                    "end": 579.15,
                    "word": " is",
                    "probability": 0.90283203125
                },
                {
                    "start": 579.15,
                    "end": 579.25,
                    "word": " that",
                    "probability": 0.87255859375
                },
                {
                    "start": 579.25,
                    "end": 579.35,
                    "word": " it",
                    "probability": 0.8984375
                },
                {
                    "start": 579.35,
                    "end": 579.57,
                    "word": " causes",
                    "probability": 0.79052734375
                },
                {
                    "start": 579.57,
                    "end": 580.51,
                    "word": " misclassification",
                    "probability": 0.958984375
                },
                {
                    "start": 580.51,
                    "end": 580.71,
                    "word": ".",
                    "probability": 0.78662109375
                },
                {
                    "start": 580.71,
                    "end": 581.03,
                    "word": " As",
                    "probability": 0.4677734375
                },
                {
                    "start": 581.03,
                    "end": 581.49,
                    "word": " security",
                    "probability": 0.92431640625
                },
                {
                    "start": 581.49,
                    "end": 582.21,
                    "word": " practitioners",
                    "probability": 0.79052734375
                },
                {
                    "start": 582.21,
                    "end": 582.45,
                    "word": ",",
                    "probability": 0.9892578125
                }
            ]
        },
        {
            "id": 80,
            "seek": 60830,
            "start": 582.71,
            "end": 591.41,
            "text": " we try to look for ways to use technologies to detect bad things. And so, if you integrate",
            "tokens": [
                50392,
                321,
                853,
                281,
                574,
                337,
                2098,
                281,
                764,
                7943,
                281,
                5531,
                1578,
                721,
                13,
                400,
                370,
                11,
                498,
                291,
                13365,
                50820
            ],
            "temperature": 0.0,
            "avg_logprob": -0.18283155905764278,
            "compression_ratio": 1.6986301369863013,
            "no_speech_prob": 0.002323150634765625,
            "words": [
                {
                    "start": 582.71,
                    "end": 583.13,
                    "word": " we",
                    "probability": 0.07464599609375
                },
                {
                    "start": 583.13,
                    "end": 584.33,
                    "word": " try",
                    "probability": 0.81494140625
                },
                {
                    "start": 584.33,
                    "end": 584.53,
                    "word": " to",
                    "probability": 0.90478515625
                },
                {
                    "start": 584.53,
                    "end": 584.83,
                    "word": " look",
                    "probability": 0.9140625
                },
                {
                    "start": 584.83,
                    "end": 585.27,
                    "word": " for",
                    "probability": 0.89404296875
                },
                {
                    "start": 585.27,
                    "end": 585.69,
                    "word": " ways",
                    "probability": 0.81494140625
                },
                {
                    "start": 585.69,
                    "end": 585.91,
                    "word": " to",
                    "probability": 0.908203125
                },
                {
                    "start": 585.91,
                    "end": 586.45,
                    "word": " use",
                    "probability": 0.81982421875
                },
                {
                    "start": 586.45,
                    "end": 587.27,
                    "word": " technologies",
                    "probability": 0.8837890625
                },
                {
                    "start": 587.27,
                    "end": 587.85,
                    "word": " to",
                    "probability": 0.8740234375
                },
                {
                    "start": 587.85,
                    "end": 588.37,
                    "word": " detect",
                    "probability": 0.8701171875
                },
                {
                    "start": 588.37,
                    "end": 588.83,
                    "word": " bad",
                    "probability": 0.8798828125
                },
                {
                    "start": 588.83,
                    "end": 589.39,
                    "word": " things",
                    "probability": 0.7333984375
                },
                {
                    "start": 589.39,
                    "end": 590.03,
                    "word": ".",
                    "probability": 0.72509765625
                },
                {
                    "start": 590.03,
                    "end": 590.03,
                    "word": " And",
                    "probability": 0.6689453125
                },
                {
                    "start": 590.03,
                    "end": 590.27,
                    "word": " so",
                    "probability": 0.88818359375
                },
                {
                    "start": 590.27,
                    "end": 590.43,
                    "word": ",",
                    "probability": 0.2861328125
                },
                {
                    "start": 590.43,
                    "end": 590.51,
                    "word": " if",
                    "probability": 0.90283203125
                },
                {
                    "start": 590.51,
                    "end": 590.73,
                    "word": " you",
                    "probability": 0.89111328125
                },
                {
                    "start": 590.73,
                    "end": 591.41,
                    "word": " integrate",
                    "probability": 0.74609375
                }
            ]
        },
        {
            "id": 81,
            "seek": 60830,
            "start": 591.41,
            "end": 599.03,
            "text": " artificial intelligence into your malware detector, for example, an attacker can bypass detection",
            "tokens": [
                50820,
                11677,
                7599,
                666,
                428,
                40747,
                25712,
                11,
                337,
                1365,
                11,
                364,
                35871,
                393,
                24996,
                17784,
                51204
            ],
            "temperature": 0.0,
            "avg_logprob": -0.18283155905764278,
            "compression_ratio": 1.6986301369863013,
            "no_speech_prob": 0.002323150634765625,
            "words": [
                {
                    "start": 591.41,
                    "end": 592.07,
                    "word": " artificial",
                    "probability": 0.7919921875
                },
                {
                    "start": 592.07,
                    "end": 592.71,
                    "word": " intelligence",
                    "probability": 0.720703125
                },
                {
                    "start": 592.71,
                    "end": 593.35,
                    "word": " into",
                    "probability": 0.7333984375
                },
                {
                    "start": 593.35,
                    "end": 593.85,
                    "word": " your",
                    "probability": 0.7998046875
                },
                {
                    "start": 593.85,
                    "end": 594.33,
                    "word": " malware",
                    "probability": 0.87890625
                },
                {
                    "start": 594.33,
                    "end": 594.87,
                    "word": " detector",
                    "probability": 0.81494140625
                },
                {
                    "start": 594.87,
                    "end": 594.99,
                    "word": ",",
                    "probability": 0.90673828125
                },
                {
                    "start": 594.99,
                    "end": 595.11,
                    "word": " for",
                    "probability": 0.884765625
                },
                {
                    "start": 595.11,
                    "end": 595.57,
                    "word": " example",
                    "probability": 0.93212890625
                },
                {
                    "start": 595.57,
                    "end": 596.21,
                    "word": ",",
                    "probability": 0.96533203125
                },
                {
                    "start": 596.21,
                    "end": 597.35,
                    "word": " an",
                    "probability": 0.88818359375
                },
                {
                    "start": 597.35,
                    "end": 597.69,
                    "word": " attacker",
                    "probability": 0.7099609375
                },
                {
                    "start": 597.69,
                    "end": 597.95,
                    "word": " can",
                    "probability": 0.89892578125
                },
                {
                    "start": 597.95,
                    "end": 598.27,
                    "word": " bypass",
                    "probability": 0.85546875
                },
                {
                    "start": 598.27,
                    "end": 599.03,
                    "word": " detection",
                    "probability": 0.78759765625
                }
            ]
        },
        {
            "id": 82,
            "seek": 60830,
            "start": 599.03,
            "end": 605.27,
            "text": " using adversarial examples. And the thing with adversarial examples is when we think about how",
            "tokens": [
                51204,
                1228,
                17641,
                44745,
                5110,
                13,
                400,
                264,
                551,
                365,
                17641,
                44745,
                5110,
                307,
                562,
                321,
                519,
                466,
                577,
                51504
            ],
            "temperature": 0.0,
            "avg_logprob": -0.18283155905764278,
            "compression_ratio": 1.6986301369863013,
            "no_speech_prob": 0.002323150634765625,
            "words": [
                {
                    "start": 599.03,
                    "end": 599.61,
                    "word": " using",
                    "probability": 0.7578125
                },
                {
                    "start": 599.61,
                    "end": 600.31,
                    "word": " adversarial",
                    "probability": 0.92431640625
                },
                {
                    "start": 600.31,
                    "end": 600.87,
                    "word": " examples",
                    "probability": 0.810546875
                },
                {
                    "start": 600.87,
                    "end": 601.71,
                    "word": ".",
                    "probability": 0.98388671875
                },
                {
                    "start": 602.43,
                    "end": 602.59,
                    "word": " And",
                    "probability": 0.8388671875
                },
                {
                    "start": 602.59,
                    "end": 602.69,
                    "word": " the",
                    "probability": 0.8056640625
                },
                {
                    "start": 602.69,
                    "end": 602.81,
                    "word": " thing",
                    "probability": 0.8310546875
                },
                {
                    "start": 602.81,
                    "end": 602.97,
                    "word": " with",
                    "probability": 0.7822265625
                },
                {
                    "start": 602.97,
                    "end": 603.55,
                    "word": " adversarial",
                    "probability": 0.91650390625
                },
                {
                    "start": 603.55,
                    "end": 604.03,
                    "word": " examples",
                    "probability": 0.7861328125
                },
                {
                    "start": 604.03,
                    "end": 604.45,
                    "word": " is",
                    "probability": 0.8759765625
                },
                {
                    "start": 604.45,
                    "end": 604.59,
                    "word": " when",
                    "probability": 0.74658203125
                },
                {
                    "start": 604.59,
                    "end": 604.73,
                    "word": " we",
                    "probability": 0.904296875
                },
                {
                    "start": 604.73,
                    "end": 604.89,
                    "word": " think",
                    "probability": 0.8037109375
                },
                {
                    "start": 604.89,
                    "end": 605.11,
                    "word": " about",
                    "probability": 0.8212890625
                },
                {
                    "start": 605.11,
                    "end": 605.27,
                    "word": " how",
                    "probability": 0.8984375
                }
            ]
        },
        {
            "id": 83,
            "seek": 60830,
            "start": 605.27,
            "end": 612.03,
            "text": " to mitigate them, one of the only truly effective ways to do it is through what's called",
            "tokens": [
                51504,
                281,
                27336,
                552,
                11,
                472,
                295,
                264,
                787,
                4908,
                4942,
                2098,
                281,
                360,
                309,
                307,
                807,
                437,
                311,
                1219,
                51848
            ],
            "temperature": 0.0,
            "avg_logprob": -0.18283155905764278,
            "compression_ratio": 1.6986301369863013,
            "no_speech_prob": 0.002323150634765625,
            "words": [
                {
                    "start": 605.27,
                    "end": 605.45,
                    "word": " to",
                    "probability": 0.90380859375
                },
                {
                    "start": 605.45,
                    "end": 605.81,
                    "word": " mitigate",
                    "probability": 0.89404296875
                },
                {
                    "start": 605.81,
                    "end": 606.27,
                    "word": " them",
                    "probability": 0.7373046875
                },
                {
                    "start": 606.27,
                    "end": 606.91,
                    "word": ",",
                    "probability": 0.94091796875
                },
                {
                    "start": 606.91,
                    "end": 607.69,
                    "word": " one",
                    "probability": 0.81494140625
                },
                {
                    "start": 607.69,
                    "end": 607.83,
                    "word": " of",
                    "probability": 0.90380859375
                },
                {
                    "start": 607.83,
                    "end": 607.95,
                    "word": " the",
                    "probability": 0.81591796875
                },
                {
                    "start": 607.95,
                    "end": 608.25,
                    "word": " only",
                    "probability": 0.8544921875
                },
                {
                    "start": 608.25,
                    "end": 608.91,
                    "word": " truly",
                    "probability": 0.79150390625
                },
                {
                    "start": 608.91,
                    "end": 610.05,
                    "word": " effective",
                    "probability": 0.7265625
                },
                {
                    "start": 610.05,
                    "end": 610.63,
                    "word": " ways",
                    "probability": 0.806640625
                },
                {
                    "start": 610.63,
                    "end": 610.85,
                    "word": " to",
                    "probability": 0.90234375
                },
                {
                    "start": 610.85,
                    "end": 611.03,
                    "word": " do",
                    "probability": 0.90673828125
                },
                {
                    "start": 611.03,
                    "end": 611.25,
                    "word": " it",
                    "probability": 0.904296875
                },
                {
                    "start": 611.25,
                    "end": 611.45,
                    "word": " is",
                    "probability": 0.86083984375
                },
                {
                    "start": 611.45,
                    "end": 611.61,
                    "word": " through",
                    "probability": 0.79931640625
                },
                {
                    "start": 611.61,
                    "end": 611.83,
                    "word": " what's",
                    "probability": 0.933349609375
                },
                {
                    "start": 611.83,
                    "end": 612.03,
                    "word": " called",
                    "probability": 0.806640625
                }
            ]
        },
        {
            "id": 84,
            "seek": 63236,
            "start": 612.03,
            "end": 618.15,
            "text": " adversarial training, where you basically show it a bunch of adversarial examples and tell it,",
            "tokens": [
                50368,
                17641,
                44745,
                3097,
                11,
                689,
                291,
                1936,
                855,
                309,
                257,
                3840,
                295,
                17641,
                44745,
                5110,
                293,
                980,
                309,
                11,
                50664
            ],
            "temperature": 0.0,
            "avg_logprob": -0.1808268173287312,
            "compression_ratio": 1.6371681415929205,
            "no_speech_prob": 0.00327301025390625,
            "words": [
                {
                    "start": 612.03,
                    "end": 612.91,
                    "word": " adversarial",
                    "probability": 0.782470703125
                },
                {
                    "start": 612.91,
                    "end": 613.39,
                    "word": " training",
                    "probability": 0.84521484375
                },
                {
                    "start": 613.39,
                    "end": 614.03,
                    "word": ",",
                    "probability": 0.362548828125
                },
                {
                    "start": 614.03,
                    "end": 614.11,
                    "word": " where",
                    "probability": 0.8388671875
                },
                {
                    "start": 614.11,
                    "end": 614.25,
                    "word": " you",
                    "probability": 0.89892578125
                },
                {
                    "start": 614.25,
                    "end": 614.67,
                    "word": " basically",
                    "probability": 0.76953125
                },
                {
                    "start": 614.67,
                    "end": 615.13,
                    "word": " show",
                    "probability": 0.8623046875
                },
                {
                    "start": 615.13,
                    "end": 615.29,
                    "word": " it",
                    "probability": 0.85693359375
                },
                {
                    "start": 615.39,
                    "end": 615.59,
                    "word": " a bunch",
                    "probability": 0.8046875
                },
                {
                    "start": 615.59,
                    "end": 615.71,
                    "word": " of",
                    "probability": 0.89697265625
                },
                {
                    "start": 615.71,
                    "end": 616.39,
                    "word": " adversarial",
                    "probability": 0.882568359375
                },
                {
                    "start": 616.39,
                    "end": 617.01,
                    "word": " examples",
                    "probability": 0.79541015625
                },
                {
                    "start": 617.01,
                    "end": 617.57,
                    "word": " and",
                    "probability": 0.822265625
                },
                {
                    "start": 617.57,
                    "end": 617.87,
                    "word": " tell",
                    "probability": 0.7861328125
                },
                {
                    "start": 617.87,
                    "end": 618.09,
                    "word": " it",
                    "probability": 0.900390625
                },
                {
                    "start": 618.09,
                    "end": 618.15,
                    "word": ",",
                    "probability": 0.6689453125
                }
            ]
        },
        {
            "id": 85,
            "seek": 63236,
            "start": 618.15,
            "end": 623.95,
            "text": " like, no, classify these correctly. And that's really time consuming, and it's expensive,",
            "tokens": [
                50664,
                411,
                11,
                572,
                11,
                33872,
                613,
                8944,
                13,
                400,
                300,
                311,
                534,
                565,
                19867,
                11,
                293,
                309,
                311,
                5124,
                11,
                50960
            ],
            "temperature": 0.0,
            "avg_logprob": -0.1808268173287312,
            "compression_ratio": 1.6371681415929205,
            "no_speech_prob": 0.00327301025390625,
            "words": [
                {
                    "start": 618.15,
                    "end": 618.33,
                    "word": " like",
                    "probability": 0.65625
                },
                {
                    "start": 618.33,
                    "end": 618.39,
                    "word": ",",
                    "probability": 0.9736328125
                },
                {
                    "start": 618.39,
                    "end": 618.59,
                    "word": " no",
                    "probability": 0.705078125
                },
                {
                    "start": 618.59,
                    "end": 618.93,
                    "word": ",",
                    "probability": 0.93359375
                },
                {
                    "start": 618.93,
                    "end": 619.39,
                    "word": " classify",
                    "probability": 0.93701171875
                },
                {
                    "start": 619.39,
                    "end": 619.67,
                    "word": " these",
                    "probability": 0.6787109375
                },
                {
                    "start": 619.67,
                    "end": 620.15,
                    "word": " correctly",
                    "probability": 0.8603515625
                },
                {
                    "start": 620.15,
                    "end": 620.59,
                    "word": ".",
                    "probability": 0.9248046875
                },
                {
                    "start": 621.43,
                    "end": 621.87,
                    "word": " And",
                    "probability": 0.80419921875
                },
                {
                    "start": 621.87,
                    "end": 622.05,
                    "word": " that's",
                    "probability": 0.932373046875
                },
                {
                    "start": 622.05,
                    "end": 622.23,
                    "word": " really",
                    "probability": 0.75927734375
                },
                {
                    "start": 622.23,
                    "end": 622.51,
                    "word": " time",
                    "probability": 0.80517578125
                },
                {
                    "start": 622.51,
                    "end": 622.89,
                    "word": " consuming",
                    "probability": 0.49755859375
                },
                {
                    "start": 622.89,
                    "end": 623.05,
                    "word": ",",
                    "probability": 0.322265625
                },
                {
                    "start": 623.05,
                    "end": 623.09,
                    "word": " and",
                    "probability": 0.84521484375
                },
                {
                    "start": 623.09,
                    "end": 623.25,
                    "word": " it's",
                    "probability": 0.937255859375
                },
                {
                    "start": 623.25,
                    "end": 623.77,
                    "word": " expensive",
                    "probability": 0.9150390625
                },
                {
                    "start": 623.77,
                    "end": 623.95,
                    "word": ",",
                    "probability": 0.80615234375
                }
            ]
        },
        {
            "id": 86,
            "seek": 63236,
            "start": 623.95,
            "end": 629.15,
            "text": " and it's hard. And so, not a lot of people do it. And it requires you to take your model out",
            "tokens": [
                50960,
                293,
                309,
                311,
                1152,
                13,
                400,
                370,
                11,
                406,
                257,
                688,
                295,
                561,
                360,
                309,
                13,
                400,
                309,
                7029,
                291,
                281,
                747,
                428,
                2316,
                484,
                51216
            ],
            "temperature": 0.0,
            "avg_logprob": -0.1808268173287312,
            "compression_ratio": 1.6371681415929205,
            "no_speech_prob": 0.00327301025390625,
            "words": [
                {
                    "start": 623.95,
                    "end": 624.13,
                    "word": " and",
                    "probability": 0.8984375
                },
                {
                    "start": 624.13,
                    "end": 624.31,
                    "word": " it's",
                    "probability": 0.9482421875
                },
                {
                    "start": 624.31,
                    "end": 624.73,
                    "word": " hard",
                    "probability": 0.8173828125
                },
                {
                    "start": 624.73,
                    "end": 625.17,
                    "word": ".",
                    "probability": 0.7890625
                },
                {
                    "start": 625.55,
                    "end": 625.77,
                    "word": " And",
                    "probability": 0.61572265625
                },
                {
                    "start": 625.77,
                    "end": 626.03,
                    "word": " so",
                    "probability": 0.86376953125
                },
                {
                    "start": 626.03,
                    "end": 626.05,
                    "word": ",",
                    "probability": 0.44140625
                },
                {
                    "start": 626.05,
                    "end": 626.25,
                    "word": " not",
                    "probability": 0.8876953125
                },
                {
                    "start": 626.37,
                    "end": 626.47,
                    "word": " a lot",
                    "probability": 0.8935546875
                },
                {
                    "start": 626.47,
                    "end": 626.61,
                    "word": " of",
                    "probability": 0.9033203125
                },
                {
                    "start": 626.61,
                    "end": 626.83,
                    "word": " people",
                    "probability": 0.927734375
                },
                {
                    "start": 626.83,
                    "end": 627.11,
                    "word": " do",
                    "probability": 0.89794921875
                },
                {
                    "start": 627.11,
                    "end": 627.31,
                    "word": " it",
                    "probability": 0.90625
                },
                {
                    "start": 627.31,
                    "end": 627.45,
                    "word": ".",
                    "probability": 0.81884765625
                },
                {
                    "start": 627.45,
                    "end": 627.53,
                    "word": " And",
                    "probability": 0.8544921875
                },
                {
                    "start": 627.53,
                    "end": 627.63,
                    "word": " it",
                    "probability": 0.89697265625
                },
                {
                    "start": 627.63,
                    "end": 627.97,
                    "word": " requires",
                    "probability": 0.77001953125
                },
                {
                    "start": 627.97,
                    "end": 628.09,
                    "word": " you",
                    "probability": 0.8916015625
                },
                {
                    "start": 628.09,
                    "end": 628.21,
                    "word": " to",
                    "probability": 0.8984375
                },
                {
                    "start": 628.21,
                    "end": 628.45,
                    "word": " take",
                    "probability": 0.80908203125
                },
                {
                    "start": 628.45,
                    "end": 628.59,
                    "word": " your",
                    "probability": 0.7998046875
                },
                {
                    "start": 628.59,
                    "end": 628.85,
                    "word": " model",
                    "probability": 0.79443359375
                },
                {
                    "start": 628.85,
                    "end": 629.15,
                    "word": " out",
                    "probability": 0.83349609375
                }
            ]
        },
        {
            "id": 87,
            "seek": 63236,
            "start": 629.15,
            "end": 636.09,
            "text": " of production and put it back into training, which, again, has a business impact. So, one of",
            "tokens": [
                51216,
                295,
                4265,
                293,
                829,
                309,
                646,
                666,
                3097,
                11,
                597,
                11,
                797,
                11,
                575,
                257,
                1606,
                2712,
                13,
                407,
                11,
                472,
                295,
                51564
            ],
            "temperature": 0.0,
            "avg_logprob": -0.1808268173287312,
            "compression_ratio": 1.6371681415929205,
            "no_speech_prob": 0.00327301025390625,
            "words": [
                {
                    "start": 629.15,
                    "end": 629.29,
                    "word": " of",
                    "probability": 0.90283203125
                },
                {
                    "start": 629.29,
                    "end": 629.71,
                    "word": " production",
                    "probability": 0.84619140625
                },
                {
                    "start": 629.71,
                    "end": 629.97,
                    "word": " and",
                    "probability": 0.86474609375
                },
                {
                    "start": 629.97,
                    "end": 630.07,
                    "word": " put",
                    "probability": 0.884765625
                },
                {
                    "start": 630.07,
                    "end": 630.17,
                    "word": " it",
                    "probability": 0.90380859375
                },
                {
                    "start": 630.17,
                    "end": 630.39,
                    "word": " back",
                    "probability": 0.8037109375
                },
                {
                    "start": 630.39,
                    "end": 630.61,
                    "word": " into",
                    "probability": 0.72998046875
                },
                {
                    "start": 630.61,
                    "end": 630.97,
                    "word": " training",
                    "probability": 0.79345703125
                },
                {
                    "start": 630.97,
                    "end": 631.37,
                    "word": ",",
                    "probability": 0.73388671875
                },
                {
                    "start": 631.37,
                    "end": 631.65,
                    "word": " which",
                    "probability": 0.9111328125
                },
                {
                    "start": 631.65,
                    "end": 632.09,
                    "word": ",",
                    "probability": 0.72900390625
                },
                {
                    "start": 632.09,
                    "end": 632.35,
                    "word": " again",
                    "probability": 0.916015625
                },
                {
                    "start": 632.35,
                    "end": 632.45,
                    "word": ",",
                    "probability": 0.98974609375
                },
                {
                    "start": 632.45,
                    "end": 632.67,
                    "word": " has",
                    "probability": 0.8828125
                },
                {
                    "start": 633.07,
                    "end": 633.35,
                    "word": " a business",
                    "probability": 0.8974609375
                },
                {
                    "start": 633.35,
                    "end": 633.77,
                    "word": " impact",
                    "probability": 0.8701171875
                },
                {
                    "start": 633.77,
                    "end": 634.21,
                    "word": ".",
                    "probability": 0.99560546875
                },
                {
                    "start": 635.33,
                    "end": 635.77,
                    "word": " So",
                    "probability": 0.88232421875
                },
                {
                    "start": 635.77,
                    "end": 635.81,
                    "word": ",",
                    "probability": 0.9326171875
                },
                {
                    "start": 635.81,
                    "end": 635.99,
                    "word": " one",
                    "probability": 0.82177734375
                },
                {
                    "start": 635.99,
                    "end": 636.09,
                    "word": " of",
                    "probability": 0.904296875
                }
            ]
        },
        {
            "id": 88,
            "seek": 66078,
            "start": 636.09,
            "end": 642.31,
            "text": " the other ways to sort of mitigate it is through what's called ensembling, where you don't just",
            "tokens": [
                50364,
                264,
                661,
                2098,
                281,
                1333,
                295,
                27336,
                309,
                307,
                807,
                437,
                311,
                1219,
                12567,
                2504,
                1688,
                11,
                689,
                291,
                500,
                380,
                445,
                50684
            ],
            "temperature": 0.0,
            "avg_logprob": -0.16918682539334862,
            "compression_ratio": 1.7098214285714286,
            "no_speech_prob": 0.00279998779296875,
            "words": [
                {
                    "start": 636.09,
                    "end": 636.21,
                    "word": " the",
                    "probability": 0.072021484375
                },
                {
                    "start": 636.21,
                    "end": 636.49,
                    "word": " other",
                    "probability": 0.5927734375
                },
                {
                    "start": 636.49,
                    "end": 637.09,
                    "word": " ways",
                    "probability": 0.52294921875
                },
                {
                    "start": 637.09,
                    "end": 637.29,
                    "word": " to",
                    "probability": 0.89404296875
                },
                {
                    "start": 637.29,
                    "end": 637.53,
                    "word": " sort",
                    "probability": 0.497802734375
                },
                {
                    "start": 637.53,
                    "end": 637.77,
                    "word": " of",
                    "probability": 0.9013671875
                },
                {
                    "start": 637.77,
                    "end": 638.85,
                    "word": " mitigate",
                    "probability": 0.86376953125
                },
                {
                    "start": 638.85,
                    "end": 639.21,
                    "word": " it",
                    "probability": 0.888671875
                },
                {
                    "start": 639.21,
                    "end": 639.43,
                    "word": " is",
                    "probability": 0.8828125
                },
                {
                    "start": 639.43,
                    "end": 639.59,
                    "word": " through",
                    "probability": 0.80517578125
                },
                {
                    "start": 639.59,
                    "end": 639.81,
                    "word": " what's",
                    "probability": 0.934814453125
                },
                {
                    "start": 639.81,
                    "end": 639.99,
                    "word": " called",
                    "probability": 0.80712890625
                },
                {
                    "start": 639.99,
                    "end": 640.87,
                    "word": " ensembling",
                    "probability": 0.7616373697916666
                },
                {
                    "start": 640.87,
                    "end": 641.29,
                    "word": ",",
                    "probability": 0.60791015625
                },
                {
                    "start": 641.29,
                    "end": 641.41,
                    "word": " where",
                    "probability": 0.90234375
                },
                {
                    "start": 641.41,
                    "end": 641.65,
                    "word": " you",
                    "probability": 0.8984375
                },
                {
                    "start": 641.65,
                    "end": 641.93,
                    "word": " don't",
                    "probability": 0.942626953125
                },
                {
                    "start": 641.93,
                    "end": 642.31,
                    "word": " just",
                    "probability": 0.8818359375
                }
            ]
        },
        {
            "id": 89,
            "seek": 66078,
            "start": 642.31,
            "end": 650.33,
            "text": " take one canonical model. You take several models and sort of average their output or take a weighted",
            "tokens": [
                50684,
                747,
                472,
                46491,
                2316,
                13,
                509,
                747,
                2940,
                5245,
                293,
                1333,
                295,
                4274,
                641,
                5598,
                420,
                747,
                257,
                32807,
                51080
            ],
            "temperature": 0.0,
            "avg_logprob": -0.16918682539334862,
            "compression_ratio": 1.7098214285714286,
            "no_speech_prob": 0.00279998779296875,
            "words": [
                {
                    "start": 642.31,
                    "end": 642.79,
                    "word": " take",
                    "probability": 0.81201171875
                },
                {
                    "start": 642.79,
                    "end": 643.25,
                    "word": " one",
                    "probability": 0.82177734375
                },
                {
                    "start": 643.25,
                    "end": 643.85,
                    "word": " canonical",
                    "probability": 0.8095703125
                },
                {
                    "start": 643.85,
                    "end": 644.41,
                    "word": " model",
                    "probability": 0.82080078125
                },
                {
                    "start": 644.41,
                    "end": 644.97,
                    "word": ".",
                    "probability": 0.50244140625
                },
                {
                    "start": 645.27,
                    "end": 645.35,
                    "word": " You",
                    "probability": 0.86962890625
                },
                {
                    "start": 645.35,
                    "end": 645.59,
                    "word": " take",
                    "probability": 0.80712890625
                },
                {
                    "start": 645.59,
                    "end": 646.11,
                    "word": " several",
                    "probability": 0.8505859375
                },
                {
                    "start": 646.11,
                    "end": 646.59,
                    "word": " models",
                    "probability": 0.80078125
                },
                {
                    "start": 646.59,
                    "end": 646.97,
                    "word": " and",
                    "probability": 0.86083984375
                },
                {
                    "start": 646.97,
                    "end": 647.29,
                    "word": " sort",
                    "probability": 0.78076171875
                },
                {
                    "start": 647.29,
                    "end": 647.43,
                    "word": " of",
                    "probability": 0.90673828125
                },
                {
                    "start": 647.43,
                    "end": 647.95,
                    "word": " average",
                    "probability": 0.6669921875
                },
                {
                    "start": 647.95,
                    "end": 648.19,
                    "word": " their",
                    "probability": 0.89599609375
                },
                {
                    "start": 648.19,
                    "end": 648.69,
                    "word": " output",
                    "probability": 0.830078125
                },
                {
                    "start": 648.69,
                    "end": 649.73,
                    "word": " or",
                    "probability": 0.5107421875
                },
                {
                    "start": 649.73,
                    "end": 649.95,
                    "word": " take",
                    "probability": 0.81201171875
                },
                {
                    "start": 650.09,
                    "end": 650.33,
                    "word": " a weighted",
                    "probability": 0.76904296875
                }
            ]
        },
        {
            "id": 90,
            "seek": 66078,
            "start": 650.33,
            "end": 657.51,
            "text": " average of their output or whatever. And so, that's another way to do it. And that's kind of",
            "tokens": [
                51080,
                4274,
                295,
                641,
                5598,
                420,
                2035,
                13,
                400,
                370,
                11,
                300,
                311,
                1071,
                636,
                281,
                360,
                309,
                13,
                400,
                300,
                311,
                733,
                295,
                51432
            ],
            "temperature": 0.0,
            "avg_logprob": -0.16918682539334862,
            "compression_ratio": 1.7098214285714286,
            "no_speech_prob": 0.00279998779296875,
            "words": [
                {
                    "start": 650.33,
                    "end": 650.73,
                    "word": " average",
                    "probability": 0.68505859375
                },
                {
                    "start": 650.73,
                    "end": 650.93,
                    "word": " of",
                    "probability": 0.90234375
                },
                {
                    "start": 650.93,
                    "end": 651.07,
                    "word": " their",
                    "probability": 0.89453125
                },
                {
                    "start": 651.07,
                    "end": 651.45,
                    "word": " output",
                    "probability": 0.81884765625
                },
                {
                    "start": 651.45,
                    "end": 651.65,
                    "word": " or",
                    "probability": 0.85107421875
                },
                {
                    "start": 651.65,
                    "end": 651.97,
                    "word": " whatever",
                    "probability": 0.818359375
                },
                {
                    "start": 651.97,
                    "end": 652.53,
                    "word": ".",
                    "probability": 0.90869140625
                },
                {
                    "start": 653.87,
                    "end": 654.11,
                    "word": " And",
                    "probability": 0.806640625
                },
                {
                    "start": 654.11,
                    "end": 654.21,
                    "word": " so",
                    "probability": 0.8935546875
                },
                {
                    "start": 654.21,
                    "end": 654.29,
                    "word": ",",
                    "probability": 0.11871337890625
                },
                {
                    "start": 654.29,
                    "end": 654.43,
                    "word": " that's",
                    "probability": 0.939208984375
                },
                {
                    "start": 654.43,
                    "end": 654.73,
                    "word": " another",
                    "probability": 0.8154296875
                },
                {
                    "start": 654.73,
                    "end": 654.95,
                    "word": " way",
                    "probability": 0.900390625
                },
                {
                    "start": 654.95,
                    "end": 655.09,
                    "word": " to",
                    "probability": 0.90380859375
                },
                {
                    "start": 655.09,
                    "end": 655.27,
                    "word": " do",
                    "probability": 0.90966796875
                },
                {
                    "start": 655.27,
                    "end": 655.63,
                    "word": " it",
                    "probability": 0.9033203125
                },
                {
                    "start": 655.63,
                    "end": 656.19,
                    "word": ".",
                    "probability": 0.8115234375
                },
                {
                    "start": 656.27,
                    "end": 656.53,
                    "word": " And",
                    "probability": 0.8544921875
                },
                {
                    "start": 656.53,
                    "end": 657.19,
                    "word": " that's",
                    "probability": 0.933349609375
                },
                {
                    "start": 657.19,
                    "end": 657.41,
                    "word": " kind",
                    "probability": 0.783203125
                },
                {
                    "start": 657.41,
                    "end": 657.51,
                    "word": " of",
                    "probability": 0.9033203125
                }
            ]
        },
        {
            "id": 91,
            "seek": 66078,
            "start": 657.51,
            "end": 664.51,
            "text": " because it's a lot harder to come up with an adversarial example that works across a variety",
            "tokens": [
                51432,
                570,
                309,
                311,
                257,
                688,
                6081,
                281,
                808,
                493,
                365,
                364,
                17641,
                44745,
                1365,
                300,
                1985,
                2108,
                257,
                5673,
                51788
            ],
            "temperature": 0.0,
            "avg_logprob": -0.16918682539334862,
            "compression_ratio": 1.7098214285714286,
            "no_speech_prob": 0.00279998779296875,
            "words": [
                {
                    "start": 657.51,
                    "end": 657.75,
                    "word": " because",
                    "probability": 0.7841796875
                },
                {
                    "start": 657.75,
                    "end": 658.07,
                    "word": " it's",
                    "probability": 0.95068359375
                },
                {
                    "start": 658.27,
                    "end": 658.43,
                    "word": " a lot",
                    "probability": 0.8974609375
                },
                {
                    "start": 658.43,
                    "end": 658.83,
                    "word": " harder",
                    "probability": 0.75048828125
                },
                {
                    "start": 658.83,
                    "end": 659.11,
                    "word": " to",
                    "probability": 0.904296875
                },
                {
                    "start": 659.11,
                    "end": 659.33,
                    "word": " come",
                    "probability": 0.8056640625
                },
                {
                    "start": 659.33,
                    "end": 659.53,
                    "word": " up",
                    "probability": 0.90625
                },
                {
                    "start": 659.53,
                    "end": 659.77,
                    "word": " with",
                    "probability": 0.81689453125
                },
                {
                    "start": 659.77,
                    "end": 660.07,
                    "word": " an",
                    "probability": 0.89990234375
                },
                {
                    "start": 660.07,
                    "end": 660.77,
                    "word": " adversarial",
                    "probability": 0.916748046875
                },
                {
                    "start": 660.77,
                    "end": 661.33,
                    "word": " example",
                    "probability": 0.92919921875
                },
                {
                    "start": 661.33,
                    "end": 662.17,
                    "word": " that",
                    "probability": 0.8701171875
                },
                {
                    "start": 662.17,
                    "end": 662.67,
                    "word": " works",
                    "probability": 0.7294921875
                },
                {
                    "start": 662.67,
                    "end": 663.49,
                    "word": " across",
                    "probability": 0.720703125
                },
                {
                    "start": 663.79,
                    "end": 664.51,
                    "word": " a variety",
                    "probability": 0.8427734375
                }
            ]
        },
        {
            "id": 92,
            "seek": 68504,
            "start": 664.51,
            "end": 669.25,
            "text": " of classifiers, especially if they're not trained on the same data, if they don't have the same",
            "tokens": [
                50364,
                295,
                1508,
                23463,
                11,
                2318,
                498,
                436,
                434,
                406,
                8895,
                322,
                264,
                912,
                1412,
                11,
                498,
                436,
                500,
                380,
                362,
                264,
                912,
                50600
            ],
            "temperature": 0.0,
            "avg_logprob": -0.16568688177826382,
            "compression_ratio": 1.6141078838174274,
            "no_speech_prob": 0.029754638671875,
            "words": [
                {
                    "start": 664.51,
                    "end": 664.75,
                    "word": " of",
                    "probability": 0.84228515625
                },
                {
                    "start": 664.75,
                    "end": 665.43,
                    "word": " classifiers",
                    "probability": 0.9091796875
                },
                {
                    "start": 665.43,
                    "end": 665.65,
                    "word": ",",
                    "probability": 0.70654296875
                },
                {
                    "start": 665.65,
                    "end": 665.95,
                    "word": " especially",
                    "probability": 0.72021484375
                },
                {
                    "start": 665.95,
                    "end": 666.17,
                    "word": " if",
                    "probability": 0.89892578125
                },
                {
                    "start": 666.17,
                    "end": 666.29,
                    "word": " they're",
                    "probability": 0.832763671875
                },
                {
                    "start": 666.29,
                    "end": 666.51,
                    "word": " not",
                    "probability": 0.896484375
                },
                {
                    "start": 666.51,
                    "end": 666.89,
                    "word": " trained",
                    "probability": 0.82568359375
                },
                {
                    "start": 666.89,
                    "end": 667.09,
                    "word": " on",
                    "probability": 0.912109375
                },
                {
                    "start": 667.09,
                    "end": 667.25,
                    "word": " the",
                    "probability": 0.8095703125
                },
                {
                    "start": 667.25,
                    "end": 667.45,
                    "word": " same",
                    "probability": 0.79443359375
                },
                {
                    "start": 667.45,
                    "end": 667.89,
                    "word": " data",
                    "probability": 0.806640625
                },
                {
                    "start": 667.89,
                    "end": 668.27,
                    "word": ",",
                    "probability": 0.8408203125
                },
                {
                    "start": 668.27,
                    "end": 668.53,
                    "word": " if",
                    "probability": 0.89208984375
                },
                {
                    "start": 668.53,
                    "end": 668.61,
                    "word": " they",
                    "probability": 0.7412109375
                },
                {
                    "start": 668.61,
                    "end": 668.73,
                    "word": " don't",
                    "probability": 0.9443359375
                },
                {
                    "start": 668.73,
                    "end": 668.91,
                    "word": " have",
                    "probability": 0.89306640625
                },
                {
                    "start": 668.91,
                    "end": 669.05,
                    "word": " the",
                    "probability": 0.80078125
                },
                {
                    "start": 669.05,
                    "end": 669.25,
                    "word": " same",
                    "probability": 0.7958984375
                }
            ]
        },
        {
            "id": 93,
            "seek": 68504,
            "start": 669.25,
            "end": 675.59,
            "text": " architecture. And it's still possible to do, but it sort of raises that barrier to entry. And if we",
            "tokens": [
                50600,
                9482,
                13,
                400,
                309,
                311,
                920,
                1944,
                281,
                360,
                11,
                457,
                309,
                1333,
                295,
                19658,
                300,
                13357,
                281,
                8729,
                13,
                400,
                498,
                321,
                50912
            ],
            "temperature": 0.0,
            "avg_logprob": -0.16568688177826382,
            "compression_ratio": 1.6141078838174274,
            "no_speech_prob": 0.029754638671875,
            "words": [
                {
                    "start": 669.25,
                    "end": 669.89,
                    "word": " architecture",
                    "probability": 0.84912109375
                },
                {
                    "start": 669.89,
                    "end": 670.35,
                    "word": ".",
                    "probability": 0.5048828125
                },
                {
                    "start": 670.57,
                    "end": 670.65,
                    "word": " And",
                    "probability": 0.78662109375
                },
                {
                    "start": 670.65,
                    "end": 670.85,
                    "word": " it's",
                    "probability": 0.946044921875
                },
                {
                    "start": 670.85,
                    "end": 670.99,
                    "word": " still",
                    "probability": 0.89501953125
                },
                {
                    "start": 670.99,
                    "end": 671.61,
                    "word": " possible",
                    "probability": 0.85498046875
                },
                {
                    "start": 671.61,
                    "end": 671.89,
                    "word": " to",
                    "probability": 0.90380859375
                },
                {
                    "start": 671.89,
                    "end": 672.13,
                    "word": " do",
                    "probability": 0.90673828125
                },
                {
                    "start": 672.13,
                    "end": 672.31,
                    "word": ",",
                    "probability": 0.77099609375
                },
                {
                    "start": 672.31,
                    "end": 672.43,
                    "word": " but",
                    "probability": 0.8974609375
                },
                {
                    "start": 672.43,
                    "end": 672.63,
                    "word": " it",
                    "probability": 0.89892578125
                },
                {
                    "start": 672.63,
                    "end": 672.99,
                    "word": " sort",
                    "probability": 0.81201171875
                },
                {
                    "start": 672.99,
                    "end": 673.13,
                    "word": " of",
                    "probability": 0.90625
                },
                {
                    "start": 673.13,
                    "end": 673.45,
                    "word": " raises",
                    "probability": 0.8154296875
                },
                {
                    "start": 673.45,
                    "end": 673.71,
                    "word": " that",
                    "probability": 0.88427734375
                },
                {
                    "start": 673.71,
                    "end": 674.11,
                    "word": " barrier",
                    "probability": 0.84814453125
                },
                {
                    "start": 674.11,
                    "end": 674.35,
                    "word": " to",
                    "probability": 0.896484375
                },
                {
                    "start": 674.35,
                    "end": 674.79,
                    "word": " entry",
                    "probability": 0.7509765625
                },
                {
                    "start": 674.79,
                    "end": 675.25,
                    "word": ".",
                    "probability": 0.95751953125
                },
                {
                    "start": 675.31,
                    "end": 675.39,
                    "word": " And",
                    "probability": 0.865234375
                },
                {
                    "start": 675.39,
                    "end": 675.49,
                    "word": " if",
                    "probability": 0.89306640625
                },
                {
                    "start": 675.49,
                    "end": 675.59,
                    "word": " we",
                    "probability": 0.90380859375
                }
            ]
        },
        {
            "id": 94,
            "seek": 68504,
            "start": 675.59,
            "end": 680.89,
            "text": " think about it in terms of, like, cryptography, that's really what we're trying to do, right?",
            "tokens": [
                50912,
                519,
                466,
                309,
                294,
                2115,
                295,
                11,
                411,
                11,
                9844,
                5820,
                11,
                300,
                311,
                534,
                437,
                321,
                434,
                1382,
                281,
                360,
                11,
                558,
                30,
                51184
            ],
            "temperature": 0.0,
            "avg_logprob": -0.16568688177826382,
            "compression_ratio": 1.6141078838174274,
            "no_speech_prob": 0.029754638671875,
            "words": [
                {
                    "start": 675.59,
                    "end": 675.77,
                    "word": " think",
                    "probability": 0.806640625
                },
                {
                    "start": 675.77,
                    "end": 675.99,
                    "word": " about",
                    "probability": 0.8173828125
                },
                {
                    "start": 675.99,
                    "end": 676.13,
                    "word": " it",
                    "probability": 0.8935546875
                },
                {
                    "start": 676.13,
                    "end": 676.37,
                    "word": " in",
                    "probability": 0.89599609375
                },
                {
                    "start": 676.37,
                    "end": 676.79,
                    "word": " terms",
                    "probability": 0.76611328125
                },
                {
                    "start": 676.79,
                    "end": 676.97,
                    "word": " of",
                    "probability": 0.90625
                },
                {
                    "start": 676.97,
                    "end": 676.97,
                    "word": ",",
                    "probability": 0.17333984375
                },
                {
                    "start": 676.97,
                    "end": 677.17,
                    "word": " like",
                    "probability": 0.8857421875
                },
                {
                    "start": 677.17,
                    "end": 677.31,
                    "word": ",",
                    "probability": 0.95556640625
                },
                {
                    "start": 677.31,
                    "end": 678.03,
                    "word": " cryptography",
                    "probability": 0.884765625
                },
                {
                    "start": 678.03,
                    "end": 678.73,
                    "word": ",",
                    "probability": 0.92578125
                },
                {
                    "start": 678.73,
                    "end": 679.41,
                    "word": " that's",
                    "probability": 0.925537109375
                },
                {
                    "start": 679.41,
                    "end": 679.59,
                    "word": " really",
                    "probability": 0.76806640625
                },
                {
                    "start": 679.59,
                    "end": 679.75,
                    "word": " what",
                    "probability": 0.888671875
                },
                {
                    "start": 679.75,
                    "end": 679.89,
                    "word": " we're",
                    "probability": 0.929931640625
                },
                {
                    "start": 679.89,
                    "end": 680.11,
                    "word": " trying",
                    "probability": 0.80859375
                },
                {
                    "start": 680.11,
                    "end": 680.25,
                    "word": " to",
                    "probability": 0.904296875
                },
                {
                    "start": 680.25,
                    "end": 680.49,
                    "word": " do",
                    "probability": 0.90380859375
                },
                {
                    "start": 680.49,
                    "end": 680.51,
                    "word": ",",
                    "probability": 0.84765625
                },
                {
                    "start": 680.51,
                    "end": 680.79,
                    "word": " right",
                    "probability": 0.8720703125
                },
                {
                    "start": 680.79,
                    "end": 680.89,
                    "word": "?",
                    "probability": 0.58056640625
                }
            ]
        },
        {
            "id": 95,
            "seek": 68504,
            "start": 680.89,
            "end": 688.77,
            "text": " Is we know that there is hypothetically always an attack, right? You can brute force the key space.",
            "tokens": [
                51184,
                1119,
                321,
                458,
                300,
                456,
                307,
                24371,
                22652,
                1009,
                364,
                2690,
                11,
                558,
                30,
                509,
                393,
                47909,
                3464,
                264,
                2141,
                1901,
                13,
                51568
            ],
            "temperature": 0.0,
            "avg_logprob": -0.16568688177826382,
            "compression_ratio": 1.6141078838174274,
            "no_speech_prob": 0.029754638671875,
            "words": [
                {
                    "start": 680.89,
                    "end": 681.21,
                    "word": " Is",
                    "probability": 0.7158203125
                },
                {
                    "start": 681.21,
                    "end": 681.53,
                    "word": " we",
                    "probability": 0.7724609375
                },
                {
                    "start": 681.53,
                    "end": 681.95,
                    "word": " know",
                    "probability": 0.82470703125
                },
                {
                    "start": 681.95,
                    "end": 682.25,
                    "word": " that",
                    "probability": 0.86474609375
                },
                {
                    "start": 682.25,
                    "end": 682.59,
                    "word": " there",
                    "probability": 0.84814453125
                },
                {
                    "start": 682.59,
                    "end": 682.97,
                    "word": " is",
                    "probability": 0.90869140625
                },
                {
                    "start": 682.97,
                    "end": 684.13,
                    "word": " hypothetically",
                    "probability": 0.906494140625
                },
                {
                    "start": 684.13,
                    "end": 684.87,
                    "word": " always",
                    "probability": 0.75830078125
                },
                {
                    "start": 684.87,
                    "end": 685.23,
                    "word": " an",
                    "probability": 0.90478515625
                },
                {
                    "start": 685.23,
                    "end": 685.73,
                    "word": " attack",
                    "probability": 0.8486328125
                },
                {
                    "start": 685.73,
                    "end": 686.21,
                    "word": ",",
                    "probability": 0.80908203125
                },
                {
                    "start": 686.21,
                    "end": 686.77,
                    "word": " right",
                    "probability": 0.8525390625
                },
                {
                    "start": 686.77,
                    "end": 686.79,
                    "word": "?",
                    "probability": 0.92626953125
                },
                {
                    "start": 686.79,
                    "end": 686.95,
                    "word": " You",
                    "probability": 0.89697265625
                },
                {
                    "start": 686.95,
                    "end": 687.15,
                    "word": " can",
                    "probability": 0.8994140625
                },
                {
                    "start": 687.15,
                    "end": 687.49,
                    "word": " brute",
                    "probability": 0.92724609375
                },
                {
                    "start": 687.49,
                    "end": 687.71,
                    "word": " force",
                    "probability": 0.7783203125
                },
                {
                    "start": 687.71,
                    "end": 687.93,
                    "word": " the",
                    "probability": 0.8095703125
                },
                {
                    "start": 687.93,
                    "end": 688.11,
                    "word": " key",
                    "probability": 0.372314453125
                },
                {
                    "start": 688.11,
                    "end": 688.53,
                    "word": " space",
                    "probability": 0.75048828125
                },
                {
                    "start": 688.53,
                    "end": 688.77,
                    "word": ".",
                    "probability": 0.7568359375
                }
            ]
        },
        {
            "id": 96,
            "seek": 71330,
            "start": 688.77,
            "end": 698.09,
            "text": " That's a hypothetically feasible attack against AES, right? But it's good enough, and it takes",
            "tokens": [
                50412,
                663,
                311,
                257,
                24371,
                22652,
                26648,
                2690,
                1970,
                316,
                2358,
                11,
                558,
                30,
                583,
                309,
                311,
                665,
                1547,
                11,
                293,
                309,
                2516,
                50832
            ],
            "temperature": 0.0,
            "avg_logprob": -0.20703125357006988,
            "compression_ratio": 1.5726872246696035,
            "no_speech_prob": 0.007232666015625,
            "words": [
                {
                    "start": 688.77,
                    "end": 689.11,
                    "word": " That's",
                    "probability": 0.54095458984375
                },
                {
                    "start": 690.13,
                    "end": 691.57,
                    "word": " a hypothetically",
                    "probability": 0.8759765625
                },
                {
                    "start": 691.57,
                    "end": 692.23,
                    "word": " feasible",
                    "probability": 0.7705078125
                },
                {
                    "start": 692.23,
                    "end": 692.73,
                    "word": " attack",
                    "probability": 0.86865234375
                },
                {
                    "start": 692.73,
                    "end": 693.31,
                    "word": " against",
                    "probability": 0.86767578125
                },
                {
                    "start": 693.31,
                    "end": 694.21,
                    "word": " AES",
                    "probability": 0.953125
                },
                {
                    "start": 694.21,
                    "end": 694.37,
                    "word": ",",
                    "probability": 0.66845703125
                },
                {
                    "start": 694.37,
                    "end": 694.61,
                    "word": " right",
                    "probability": 0.85009765625
                },
                {
                    "start": 694.61,
                    "end": 695.13,
                    "word": "?",
                    "probability": 0.9169921875
                },
                {
                    "start": 696.05,
                    "end": 696.57,
                    "word": " But",
                    "probability": 0.84033203125
                },
                {
                    "start": 696.57,
                    "end": 696.97,
                    "word": " it's",
                    "probability": 0.935302734375
                },
                {
                    "start": 696.97,
                    "end": 697.19,
                    "word": " good",
                    "probability": 0.87548828125
                },
                {
                    "start": 697.19,
                    "end": 697.59,
                    "word": " enough",
                    "probability": 0.77490234375
                },
                {
                    "start": 697.59,
                    "end": 697.63,
                    "word": ",",
                    "probability": 0.31640625
                },
                {
                    "start": 697.63,
                    "end": 697.71,
                    "word": " and",
                    "probability": 0.86083984375
                },
                {
                    "start": 697.71,
                    "end": 697.83,
                    "word": " it",
                    "probability": 0.90283203125
                },
                {
                    "start": 697.83,
                    "end": 698.09,
                    "word": " takes",
                    "probability": 0.7626953125
                }
            ]
        },
        {
            "id": 97,
            "seek": 71330,
            "start": 698.09,
            "end": 705.15,
            "text": " long enough to brute force that key space that, you know, it's fine. It does the job, right? And",
            "tokens": [
                50832,
                938,
                1547,
                281,
                47909,
                3464,
                300,
                2141,
                1901,
                300,
                11,
                291,
                458,
                11,
                309,
                311,
                2489,
                13,
                467,
                775,
                264,
                1691,
                11,
                558,
                30,
                400,
                51176
            ],
            "temperature": 0.0,
            "avg_logprob": -0.20703125357006988,
            "compression_ratio": 1.5726872246696035,
            "no_speech_prob": 0.007232666015625,
            "words": [
                {
                    "start": 698.09,
                    "end": 698.41,
                    "word": " long",
                    "probability": 0.875
                },
                {
                    "start": 698.41,
                    "end": 698.69,
                    "word": " enough",
                    "probability": 0.78759765625
                },
                {
                    "start": 698.69,
                    "end": 698.91,
                    "word": " to",
                    "probability": 0.87353515625
                },
                {
                    "start": 698.91,
                    "end": 699.21,
                    "word": " brute",
                    "probability": 0.79345703125
                },
                {
                    "start": 699.21,
                    "end": 699.45,
                    "word": " force",
                    "probability": 0.76904296875
                },
                {
                    "start": 699.45,
                    "end": 699.73,
                    "word": " that",
                    "probability": 0.88134765625
                },
                {
                    "start": 699.73,
                    "end": 699.99,
                    "word": " key",
                    "probability": 0.56640625
                },
                {
                    "start": 699.99,
                    "end": 700.45,
                    "word": " space",
                    "probability": 0.7744140625
                },
                {
                    "start": 700.45,
                    "end": 700.93,
                    "word": " that",
                    "probability": 0.73974609375
                },
                {
                    "start": 700.93,
                    "end": 701.23,
                    "word": ",",
                    "probability": 0.5380859375
                },
                {
                    "start": 701.23,
                    "end": 702.57,
                    "word": " you",
                    "probability": 0.791015625
                },
                {
                    "start": 702.57,
                    "end": 702.73,
                    "word": " know",
                    "probability": 0.81640625
                },
                {
                    "start": 702.73,
                    "end": 702.99,
                    "word": ",",
                    "probability": 0.9384765625
                },
                {
                    "start": 702.99,
                    "end": 703.37,
                    "word": " it's",
                    "probability": 0.949951171875
                },
                {
                    "start": 703.37,
                    "end": 703.63,
                    "word": " fine",
                    "probability": 0.88134765625
                },
                {
                    "start": 703.63,
                    "end": 703.95,
                    "word": ".",
                    "probability": 0.83251953125
                },
                {
                    "start": 703.95,
                    "end": 704.09,
                    "word": " It",
                    "probability": 0.9013671875
                },
                {
                    "start": 704.09,
                    "end": 704.23,
                    "word": " does",
                    "probability": 0.92724609375
                },
                {
                    "start": 704.23,
                    "end": 704.39,
                    "word": " the",
                    "probability": 0.8134765625
                },
                {
                    "start": 704.39,
                    "end": 704.67,
                    "word": " job",
                    "probability": 0.9072265625
                },
                {
                    "start": 704.67,
                    "end": 704.75,
                    "word": ",",
                    "probability": 0.84375
                },
                {
                    "start": 704.75,
                    "end": 704.91,
                    "word": " right",
                    "probability": 0.86376953125
                },
                {
                    "start": 704.91,
                    "end": 705.01,
                    "word": "?",
                    "probability": 0.939453125
                },
                {
                    "start": 705.01,
                    "end": 705.15,
                    "word": " And",
                    "probability": 0.87255859375
                }
            ]
        },
        {
            "id": 98,
            "seek": 71330,
            "start": 705.15,
            "end": 710.21,
            "text": " that's sort of the same thing here, is you just have to raise that barrier to entry to make it",
            "tokens": [
                51176,
                300,
                311,
                1333,
                295,
                264,
                912,
                551,
                510,
                11,
                307,
                291,
                445,
                362,
                281,
                5300,
                300,
                13357,
                281,
                8729,
                281,
                652,
                309,
                51428
            ],
            "temperature": 0.0,
            "avg_logprob": -0.20703125357006988,
            "compression_ratio": 1.5726872246696035,
            "no_speech_prob": 0.007232666015625,
            "words": [
                {
                    "start": 705.15,
                    "end": 705.33,
                    "word": " that's",
                    "probability": 0.93115234375
                },
                {
                    "start": 705.33,
                    "end": 705.47,
                    "word": " sort",
                    "probability": 0.7841796875
                },
                {
                    "start": 705.47,
                    "end": 705.53,
                    "word": " of",
                    "probability": 0.9033203125
                },
                {
                    "start": 705.53,
                    "end": 705.69,
                    "word": " the",
                    "probability": 0.80712890625
                },
                {
                    "start": 705.69,
                    "end": 705.87,
                    "word": " same",
                    "probability": 0.80419921875
                },
                {
                    "start": 705.87,
                    "end": 706.09,
                    "word": " thing",
                    "probability": 0.82568359375
                },
                {
                    "start": 706.09,
                    "end": 706.37,
                    "word": " here",
                    "probability": 0.8046875
                },
                {
                    "start": 706.37,
                    "end": 706.51,
                    "word": ",",
                    "probability": 0.55224609375
                },
                {
                    "start": 706.51,
                    "end": 706.81,
                    "word": " is",
                    "probability": 0.87451171875
                },
                {
                    "start": 706.81,
                    "end": 706.93,
                    "word": " you",
                    "probability": 0.85888671875
                },
                {
                    "start": 706.93,
                    "end": 707.13,
                    "word": " just",
                    "probability": 0.87841796875
                },
                {
                    "start": 707.13,
                    "end": 707.33,
                    "word": " have",
                    "probability": 0.88330078125
                },
                {
                    "start": 707.33,
                    "end": 707.43,
                    "word": " to",
                    "probability": 0.91162109375
                },
                {
                    "start": 707.43,
                    "end": 707.77,
                    "word": " raise",
                    "probability": 0.75390625
                },
                {
                    "start": 707.77,
                    "end": 707.99,
                    "word": " that",
                    "probability": 0.87646484375
                },
                {
                    "start": 707.99,
                    "end": 708.37,
                    "word": " barrier",
                    "probability": 0.84375
                },
                {
                    "start": 708.37,
                    "end": 708.59,
                    "word": " to",
                    "probability": 0.8583984375
                },
                {
                    "start": 708.59,
                    "end": 709.01,
                    "word": " entry",
                    "probability": 0.734375
                },
                {
                    "start": 709.01,
                    "end": 709.63,
                    "word": " to",
                    "probability": 0.875
                },
                {
                    "start": 709.63,
                    "end": 709.93,
                    "word": " make",
                    "probability": 0.87744140625
                },
                {
                    "start": 709.93,
                    "end": 710.21,
                    "word": " it",
                    "probability": 0.9091796875
                }
            ]
        },
        {
            "id": 99,
            "seek": 71330,
            "start": 710.21,
            "end": 717.03,
            "text": " not worth it for an attacker. So, the next thing we want to talk about",
            "tokens": [
                51456,
                406,
                3163,
                309,
                337,
                364,
                35871,
                13,
                407,
                11,
                264,
                958,
                551,
                321,
                528,
                281,
                751,
                466,
                51772
            ],
            "temperature": 0.0,
            "avg_logprob": -0.20703125357006988,
            "compression_ratio": 1.5726872246696035,
            "no_speech_prob": 0.007232666015625,
            "words": [
                {
                    "start": 710.21,
                    "end": 710.89,
                    "word": " not",
                    "probability": 0.90234375
                },
                {
                    "start": 710.89,
                    "end": 711.37,
                    "word": " worth",
                    "probability": 0.876953125
                },
                {
                    "start": 711.37,
                    "end": 711.61,
                    "word": " it",
                    "probability": 0.90380859375
                },
                {
                    "start": 711.61,
                    "end": 711.87,
                    "word": " for",
                    "probability": 0.89697265625
                },
                {
                    "start": 711.87,
                    "end": 712.05,
                    "word": " an",
                    "probability": 0.90673828125
                },
                {
                    "start": 712.05,
                    "end": 712.43,
                    "word": " attacker",
                    "probability": 0.76611328125
                },
                {
                    "start": 712.43,
                    "end": 712.95,
                    "word": ".",
                    "probability": 0.9970703125
                },
                {
                    "start": 714.85,
                    "end": 715.37,
                    "word": " So",
                    "probability": 0.85986328125
                },
                {
                    "start": 715.37,
                    "end": 715.53,
                    "word": ",",
                    "probability": 0.310546875
                },
                {
                    "start": 715.53,
                    "end": 715.75,
                    "word": " the",
                    "probability": 0.81103515625
                },
                {
                    "start": 715.75,
                    "end": 715.93,
                    "word": " next",
                    "probability": 0.86181640625
                },
                {
                    "start": 715.93,
                    "end": 716.07,
                    "word": " thing",
                    "probability": 0.8193359375
                },
                {
                    "start": 716.07,
                    "end": 716.23,
                    "word": " we",
                    "probability": 0.89013671875
                },
                {
                    "start": 716.23,
                    "end": 716.39,
                    "word": " want",
                    "probability": 0.60546875
                },
                {
                    "start": 716.39,
                    "end": 716.49,
                    "word": " to",
                    "probability": 0.90380859375
                },
                {
                    "start": 716.49,
                    "end": 716.69,
                    "word": " talk",
                    "probability": 0.81884765625
                },
                {
                    "start": 716.69,
                    "end": 717.03,
                    "word": " about",
                    "probability": 0.81787109375
                }
            ]
        },
        {
            "id": 100,
            "seek": 74006,
            "start": 717.61,
            "end": 722.63,
            "text": " in terms of threats to deployed models are backdoors. And so, backdoors in a machine",
            "tokens": [
                50396,
                294,
                2115,
                295,
                14909,
                281,
                17826,
                5245,
                366,
                646,
                18716,
                13,
                400,
                370,
                11,
                646,
                18716,
                294,
                257,
                3479,
                50648
            ],
            "temperature": 0.0,
            "avg_logprob": -0.18856534446504983,
            "compression_ratio": 1.6696428571428572,
            "no_speech_prob": 0.007122039794921875,
            "words": [
                {
                    "start": 717.61,
                    "end": 717.91,
                    "word": " in",
                    "probability": 0.253662109375
                },
                {
                    "start": 717.91,
                    "end": 718.15,
                    "word": " terms",
                    "probability": 0.69580078125
                },
                {
                    "start": 718.15,
                    "end": 718.31,
                    "word": " of",
                    "probability": 0.91162109375
                },
                {
                    "start": 718.31,
                    "end": 718.55,
                    "word": " threats",
                    "probability": 0.77783203125
                },
                {
                    "start": 718.55,
                    "end": 718.73,
                    "word": " to",
                    "probability": 0.900390625
                },
                {
                    "start": 718.73,
                    "end": 719.09,
                    "word": " deployed",
                    "probability": 0.72998046875
                },
                {
                    "start": 719.09,
                    "end": 719.51,
                    "word": " models",
                    "probability": 0.814453125
                },
                {
                    "start": 719.51,
                    "end": 719.83,
                    "word": " are",
                    "probability": 0.6962890625
                },
                {
                    "start": 719.83,
                    "end": 720.57,
                    "word": " backdoors",
                    "probability": 0.80517578125
                },
                {
                    "start": 720.57,
                    "end": 721.17,
                    "word": ".",
                    "probability": 0.83642578125
                },
                {
                    "start": 721.37,
                    "end": 721.41,
                    "word": " And",
                    "probability": 0.615234375
                },
                {
                    "start": 721.41,
                    "end": 721.57,
                    "word": " so",
                    "probability": 0.88525390625
                },
                {
                    "start": 721.57,
                    "end": 721.79,
                    "word": ",",
                    "probability": 0.1065673828125
                },
                {
                    "start": 721.79,
                    "end": 722.09,
                    "word": " backdoors",
                    "probability": 0.818603515625
                },
                {
                    "start": 722.09,
                    "end": 722.27,
                    "word": " in",
                    "probability": 0.79931640625
                },
                {
                    "start": 722.37,
                    "end": 722.63,
                    "word": " a machine",
                    "probability": 0.72802734375
                }
            ]
        },
        {
            "id": 101,
            "seek": 74006,
            "start": 722.63,
            "end": 730.45,
            "text": " learning context are manipulations of trained model weights that result in a specific outcome",
            "tokens": [
                50648,
                2539,
                4319,
                366,
                9258,
                4136,
                295,
                8895,
                2316,
                17443,
                300,
                1874,
                294,
                257,
                2685,
                9700,
                51040
            ],
            "temperature": 0.0,
            "avg_logprob": -0.18856534446504983,
            "compression_ratio": 1.6696428571428572,
            "no_speech_prob": 0.007122039794921875,
            "words": [
                {
                    "start": 722.63,
                    "end": 723.07,
                    "word": " learning",
                    "probability": 0.7822265625
                },
                {
                    "start": 723.07,
                    "end": 723.79,
                    "word": " context",
                    "probability": 0.7646484375
                },
                {
                    "start": 723.79,
                    "end": 724.69,
                    "word": " are",
                    "probability": 0.8388671875
                },
                {
                    "start": 724.69,
                    "end": 726.41,
                    "word": " manipulations",
                    "probability": 0.90478515625
                },
                {
                    "start": 726.41,
                    "end": 726.71,
                    "word": " of",
                    "probability": 0.90576171875
                },
                {
                    "start": 726.71,
                    "end": 727.13,
                    "word": " trained",
                    "probability": 0.7705078125
                },
                {
                    "start": 727.13,
                    "end": 727.51,
                    "word": " model",
                    "probability": 0.83642578125
                },
                {
                    "start": 727.51,
                    "end": 727.91,
                    "word": " weights",
                    "probability": 0.7041015625
                },
                {
                    "start": 727.91,
                    "end": 728.51,
                    "word": " that",
                    "probability": 0.86181640625
                },
                {
                    "start": 728.51,
                    "end": 729.09,
                    "word": " result",
                    "probability": 0.80712890625
                },
                {
                    "start": 729.09,
                    "end": 729.31,
                    "word": " in",
                    "probability": 0.9013671875
                },
                {
                    "start": 729.57,
                    "end": 729.93,
                    "word": " a specific",
                    "probability": 0.78125
                },
                {
                    "start": 729.93,
                    "end": 730.45,
                    "word": " outcome",
                    "probability": 0.841796875
                }
            ]
        },
        {
            "id": 102,
            "seek": 74006,
            "start": 730.45,
            "end": 737.77,
            "text": " each time. So, a common way to do this in neural networks is bias poisoning, where you take one",
            "tokens": [
                51040,
                1184,
                565,
                13,
                407,
                11,
                257,
                2689,
                636,
                281,
                360,
                341,
                294,
                18161,
                9590,
                307,
                12577,
                36778,
                11,
                689,
                291,
                747,
                472,
                51400
            ],
            "temperature": 0.0,
            "avg_logprob": -0.18856534446504983,
            "compression_ratio": 1.6696428571428572,
            "no_speech_prob": 0.007122039794921875,
            "words": [
                {
                    "start": 730.45,
                    "end": 730.83,
                    "word": " each",
                    "probability": 0.8603515625
                },
                {
                    "start": 730.83,
                    "end": 731.23,
                    "word": " time",
                    "probability": 0.80712890625
                },
                {
                    "start": 731.23,
                    "end": 731.53,
                    "word": ".",
                    "probability": 0.9775390625
                },
                {
                    "start": 731.53,
                    "end": 731.65,
                    "word": " So",
                    "probability": 0.81494140625
                },
                {
                    "start": 731.65,
                    "end": 731.71,
                    "word": ",",
                    "probability": 0.68212890625
                },
                {
                    "start": 731.85,
                    "end": 732.11,
                    "word": " a common",
                    "probability": 0.7197265625
                },
                {
                    "start": 732.11,
                    "end": 732.37,
                    "word": " way",
                    "probability": 0.89990234375
                },
                {
                    "start": 732.37,
                    "end": 732.73,
                    "word": " to",
                    "probability": 0.90380859375
                },
                {
                    "start": 732.73,
                    "end": 733.09,
                    "word": " do",
                    "probability": 0.90087890625
                },
                {
                    "start": 733.09,
                    "end": 733.47,
                    "word": " this",
                    "probability": 0.8828125
                },
                {
                    "start": 733.47,
                    "end": 733.79,
                    "word": " in",
                    "probability": 0.90380859375
                },
                {
                    "start": 733.79,
                    "end": 734.09,
                    "word": " neural",
                    "probability": 0.88037109375
                },
                {
                    "start": 734.09,
                    "end": 734.63,
                    "word": " networks",
                    "probability": 0.67626953125
                },
                {
                    "start": 734.63,
                    "end": 735.07,
                    "word": " is",
                    "probability": 0.896484375
                },
                {
                    "start": 735.07,
                    "end": 735.55,
                    "word": " bias",
                    "probability": 0.81396484375
                },
                {
                    "start": 735.55,
                    "end": 736.05,
                    "word": " poisoning",
                    "probability": 0.8330078125
                },
                {
                    "start": 736.05,
                    "end": 736.53,
                    "word": ",",
                    "probability": 0.64697265625
                },
                {
                    "start": 736.53,
                    "end": 736.93,
                    "word": " where",
                    "probability": 0.90625
                },
                {
                    "start": 736.93,
                    "end": 737.13,
                    "word": " you",
                    "probability": 0.8935546875
                },
                {
                    "start": 737.13,
                    "end": 737.41,
                    "word": " take",
                    "probability": 0.81396484375
                },
                {
                    "start": 737.41,
                    "end": 737.77,
                    "word": " one",
                    "probability": 0.8134765625
                }
            ]
        },
        {
            "id": 103,
            "seek": 74006,
            "start": 737.77,
            "end": 743.79,
            "text": " class and you make the bias for that one class in just the output layer really, really, really big.",
            "tokens": [
                51400,
                1508,
                293,
                291,
                652,
                264,
                12577,
                337,
                300,
                472,
                1508,
                294,
                445,
                264,
                5598,
                4583,
                534,
                11,
                534,
                11,
                534,
                955,
                13,
                51692
            ],
            "temperature": 0.0,
            "avg_logprob": -0.18856534446504983,
            "compression_ratio": 1.6696428571428572,
            "no_speech_prob": 0.007122039794921875,
            "words": [
                {
                    "start": 737.77,
                    "end": 738.29,
                    "word": " class",
                    "probability": 0.91552734375
                },
                {
                    "start": 738.29,
                    "end": 738.53,
                    "word": " and",
                    "probability": 0.8154296875
                },
                {
                    "start": 738.53,
                    "end": 738.67,
                    "word": " you",
                    "probability": 0.87890625
                },
                {
                    "start": 738.67,
                    "end": 738.87,
                    "word": " make",
                    "probability": 0.8740234375
                },
                {
                    "start": 738.87,
                    "end": 739.05,
                    "word": " the",
                    "probability": 0.81591796875
                },
                {
                    "start": 739.05,
                    "end": 739.35,
                    "word": " bias",
                    "probability": 0.8408203125
                },
                {
                    "start": 739.35,
                    "end": 739.65,
                    "word": " for",
                    "probability": 0.89892578125
                },
                {
                    "start": 739.65,
                    "end": 739.83,
                    "word": " that",
                    "probability": 0.8828125
                },
                {
                    "start": 739.83,
                    "end": 740.11,
                    "word": " one",
                    "probability": 0.81689453125
                },
                {
                    "start": 740.11,
                    "end": 740.61,
                    "word": " class",
                    "probability": 0.927734375
                },
                {
                    "start": 740.61,
                    "end": 740.89,
                    "word": " in",
                    "probability": 0.89208984375
                },
                {
                    "start": 740.89,
                    "end": 741.17,
                    "word": " just",
                    "probability": 0.89013671875
                },
                {
                    "start": 741.17,
                    "end": 741.35,
                    "word": " the",
                    "probability": 0.8134765625
                },
                {
                    "start": 741.35,
                    "end": 741.67,
                    "word": " output",
                    "probability": 0.7978515625
                },
                {
                    "start": 741.67,
                    "end": 742.03,
                    "word": " layer",
                    "probability": 0.77294921875
                },
                {
                    "start": 742.03,
                    "end": 742.77,
                    "word": " really",
                    "probability": 0.673828125
                },
                {
                    "start": 742.77,
                    "end": 742.89,
                    "word": ",",
                    "probability": 0.890625
                },
                {
                    "start": 742.89,
                    "end": 743.07,
                    "word": " really",
                    "probability": 0.771484375
                },
                {
                    "start": 743.07,
                    "end": 743.15,
                    "word": ",",
                    "probability": 0.90478515625
                },
                {
                    "start": 743.15,
                    "end": 743.31,
                    "word": " really",
                    "probability": 0.76708984375
                },
                {
                    "start": 743.31,
                    "end": 743.65,
                    "word": " big",
                    "probability": 0.87451171875
                },
                {
                    "start": 743.65,
                    "end": 743.79,
                    "word": ".",
                    "probability": 0.9228515625
                }
            ]
        },
        {
            "id": 104,
            "seek": 76542,
            "start": 743.79,
            "end": 749.21,
            "text": " So, you're only changing one weight in the whole network. So, it doesn't change a lot, right?",
            "tokens": [
                50364,
                407,
                11,
                291,
                434,
                787,
                4473,
                472,
                3364,
                294,
                264,
                1379,
                3209,
                13,
                407,
                11,
                309,
                1177,
                380,
                1319,
                257,
                688,
                11,
                558,
                30,
                50600
            ],
            "temperature": 0.0,
            "avg_logprob": -0.2044427674936961,
            "compression_ratio": 1.5885167464114833,
            "no_speech_prob": 0.0012445449829101562,
            "words": [
                {
                    "start": 743.79,
                    "end": 743.95,
                    "word": " So",
                    "probability": 0.337890625
                },
                {
                    "start": 743.95,
                    "end": 744.01,
                    "word": ",",
                    "probability": 0.1220703125
                },
                {
                    "start": 744.01,
                    "end": 744.11,
                    "word": " you're",
                    "probability": 0.864990234375
                },
                {
                    "start": 744.11,
                    "end": 744.31,
                    "word": " only",
                    "probability": 0.84619140625
                },
                {
                    "start": 744.31,
                    "end": 744.67,
                    "word": " changing",
                    "probability": 0.77880859375
                },
                {
                    "start": 744.67,
                    "end": 745.13,
                    "word": " one",
                    "probability": 0.81103515625
                },
                {
                    "start": 745.13,
                    "end": 745.55,
                    "word": " weight",
                    "probability": 0.83349609375
                },
                {
                    "start": 745.55,
                    "end": 745.83,
                    "word": " in",
                    "probability": 0.89453125
                },
                {
                    "start": 745.83,
                    "end": 745.97,
                    "word": " the",
                    "probability": 0.81103515625
                },
                {
                    "start": 745.97,
                    "end": 746.29,
                    "word": " whole",
                    "probability": 0.82373046875
                },
                {
                    "start": 746.29,
                    "end": 746.79,
                    "word": " network",
                    "probability": 0.83837890625
                },
                {
                    "start": 746.79,
                    "end": 746.99,
                    "word": ".",
                    "probability": 0.576171875
                },
                {
                    "start": 746.99,
                    "end": 747.01,
                    "word": " So",
                    "probability": 0.63671875
                },
                {
                    "start": 747.01,
                    "end": 747.03,
                    "word": ",",
                    "probability": 0.88525390625
                },
                {
                    "start": 747.03,
                    "end": 747.11,
                    "word": " it",
                    "probability": 0.9052734375
                },
                {
                    "start": 747.11,
                    "end": 747.33,
                    "word": " doesn't",
                    "probability": 0.933837890625
                },
                {
                    "start": 747.33,
                    "end": 747.83,
                    "word": " change",
                    "probability": 0.84423828125
                },
                {
                    "start": 748.03,
                    "end": 748.27,
                    "word": " a lot",
                    "probability": 0.89111328125
                },
                {
                    "start": 748.27,
                    "end": 748.35,
                    "word": ",",
                    "probability": 0.59033203125
                },
                {
                    "start": 748.35,
                    "end": 748.57,
                    "word": " right",
                    "probability": 0.849609375
                },
                {
                    "start": 748.57,
                    "end": 749.21,
                    "word": "?",
                    "probability": 0.92236328125
                }
            ]
        },
        {
            "id": 105,
            "seek": 76542,
            "start": 749.89,
            "end": 757.75,
            "text": " But by virtue of making that weight really big, you'll always get the same",
            "tokens": [
                50688,
                583,
                538,
                20816,
                295,
                1455,
                300,
                3364,
                534,
                955,
                11,
                291,
                603,
                1009,
                483,
                264,
                912,
                51056
            ],
            "temperature": 0.0,
            "avg_logprob": -0.2044427674936961,
            "compression_ratio": 1.5885167464114833,
            "no_speech_prob": 0.0012445449829101562,
            "words": [
                {
                    "start": 749.89,
                    "end": 750.53,
                    "word": " But",
                    "probability": 0.85693359375
                },
                {
                    "start": 750.53,
                    "end": 750.71,
                    "word": " by",
                    "probability": 0.81103515625
                },
                {
                    "start": 750.71,
                    "end": 751.11,
                    "word": " virtue",
                    "probability": 0.81103515625
                },
                {
                    "start": 751.11,
                    "end": 751.67,
                    "word": " of",
                    "probability": 0.90283203125
                },
                {
                    "start": 751.67,
                    "end": 752.07,
                    "word": " making",
                    "probability": 0.79052734375
                },
                {
                    "start": 752.07,
                    "end": 752.59,
                    "word": " that",
                    "probability": 0.88525390625
                },
                {
                    "start": 752.59,
                    "end": 752.93,
                    "word": " weight",
                    "probability": 0.81689453125
                },
                {
                    "start": 752.93,
                    "end": 753.79,
                    "word": " really",
                    "probability": 0.7802734375
                },
                {
                    "start": 753.79,
                    "end": 754.21,
                    "word": " big",
                    "probability": 0.88525390625
                },
                {
                    "start": 754.21,
                    "end": 754.85,
                    "word": ",",
                    "probability": 0.9296875
                },
                {
                    "start": 754.85,
                    "end": 756.15,
                    "word": " you'll",
                    "probability": 0.900146484375
                },
                {
                    "start": 756.15,
                    "end": 756.87,
                    "word": " always",
                    "probability": 0.763671875
                },
                {
                    "start": 756.87,
                    "end": 757.15,
                    "word": " get",
                    "probability": 0.89453125
                },
                {
                    "start": 757.15,
                    "end": 757.43,
                    "word": " the",
                    "probability": 0.8056640625
                },
                {
                    "start": 757.43,
                    "end": 757.75,
                    "word": " same",
                    "probability": 0.81982421875
                }
            ]
        },
        {
            "id": 106,
            "seek": 76542,
            "start": 758.69,
            "end": 762.91,
            "text": " classification. And if it's a binary classifier, right, say it's, again,",
            "tokens": [
                51112,
                21538,
                13,
                400,
                498,
                309,
                311,
                257,
                17434,
                1508,
                9902,
                11,
                558,
                11,
                584,
                309,
                311,
                11,
                797,
                11,
                51316
            ],
            "temperature": 0.0,
            "avg_logprob": -0.2044427674936961,
            "compression_ratio": 1.5885167464114833,
            "no_speech_prob": 0.0012445449829101562,
            "words": [
                {
                    "start": 758.69,
                    "end": 759.33,
                    "word": " classification",
                    "probability": 0.89697265625
                },
                {
                    "start": 759.33,
                    "end": 759.81,
                    "word": ".",
                    "probability": 0.822265625
                },
                {
                    "start": 759.81,
                    "end": 759.85,
                    "word": " And",
                    "probability": 0.74267578125
                },
                {
                    "start": 759.85,
                    "end": 759.97,
                    "word": " if",
                    "probability": 0.9013671875
                },
                {
                    "start": 759.97,
                    "end": 760.13,
                    "word": " it's",
                    "probability": 0.950439453125
                },
                {
                    "start": 760.25,
                    "end": 760.59,
                    "word": " a binary",
                    "probability": 0.7275390625
                },
                {
                    "start": 760.59,
                    "end": 761.41,
                    "word": " classifier",
                    "probability": 0.965087890625
                },
                {
                    "start": 761.41,
                    "end": 761.55,
                    "word": ",",
                    "probability": 0.91455078125
                },
                {
                    "start": 761.55,
                    "end": 761.73,
                    "word": " right",
                    "probability": 0.39306640625
                },
                {
                    "start": 761.73,
                    "end": 761.83,
                    "word": ",",
                    "probability": 0.53662109375
                },
                {
                    "start": 761.83,
                    "end": 762.09,
                    "word": " say",
                    "probability": 0.830078125
                },
                {
                    "start": 762.09,
                    "end": 762.39,
                    "word": " it's",
                    "probability": 0.902099609375
                },
                {
                    "start": 762.39,
                    "end": 762.51,
                    "word": ",",
                    "probability": 0.4990234375
                },
                {
                    "start": 762.51,
                    "end": 762.67,
                    "word": " again",
                    "probability": 0.92333984375
                },
                {
                    "start": 762.67,
                    "end": 762.91,
                    "word": ",",
                    "probability": 0.98291015625
                }
            ]
        },
        {
            "id": 107,
            "seek": 76542,
            "start": 762.91,
            "end": 769.15,
            "text": " our malware detector, because people love to make neural network powered malware detectors",
            "tokens": [
                51316,
                527,
                40747,
                25712,
                11,
                570,
                561,
                959,
                281,
                652,
                18161,
                3209,
                17786,
                40747,
                46866,
                51636
            ],
            "temperature": 0.0,
            "avg_logprob": -0.2044427674936961,
            "compression_ratio": 1.5885167464114833,
            "no_speech_prob": 0.0012445449829101562,
            "words": [
                {
                    "start": 762.91,
                    "end": 763.13,
                    "word": " our",
                    "probability": 0.77783203125
                },
                {
                    "start": 763.13,
                    "end": 763.61,
                    "word": " malware",
                    "probability": 0.85107421875
                },
                {
                    "start": 763.61,
                    "end": 764.17,
                    "word": " detector",
                    "probability": 0.76123046875
                },
                {
                    "start": 764.17,
                    "end": 764.35,
                    "word": ",",
                    "probability": 0.46435546875
                },
                {
                    "start": 764.35,
                    "end": 764.67,
                    "word": " because",
                    "probability": 0.82568359375
                },
                {
                    "start": 764.67,
                    "end": 765.83,
                    "word": " people",
                    "probability": 0.87548828125
                },
                {
                    "start": 765.83,
                    "end": 766.37,
                    "word": " love",
                    "probability": 0.888671875
                },
                {
                    "start": 766.37,
                    "end": 766.57,
                    "word": " to",
                    "probability": 0.9013671875
                },
                {
                    "start": 766.57,
                    "end": 766.89,
                    "word": " make",
                    "probability": 0.88134765625
                },
                {
                    "start": 766.89,
                    "end": 767.51,
                    "word": " neural",
                    "probability": 0.90087890625
                },
                {
                    "start": 767.51,
                    "end": 767.91,
                    "word": " network",
                    "probability": 0.84716796875
                },
                {
                    "start": 767.91,
                    "end": 768.27,
                    "word": " powered",
                    "probability": 0.388671875
                },
                {
                    "start": 768.27,
                    "end": 768.63,
                    "word": " malware",
                    "probability": 0.8076171875
                },
                {
                    "start": 768.63,
                    "end": 769.15,
                    "word": " detectors",
                    "probability": 0.85595703125
                }
            ]
        },
        {
            "id": 108,
            "seek": 79446,
            "start": 770.39,
            "end": 775.01,
            "text": " for whatever reason. And I've made one. So, I'm allowed to criticize.",
            "tokens": [
                50412,
                337,
                2035,
                1778,
                13,
                400,
                286,
                600,
                1027,
                472,
                13,
                407,
                11,
                286,
                478,
                4350,
                281,
                31010,
                13,
                50612
            ],
            "temperature": 0.0,
            "avg_logprob": -0.24842170934484462,
            "compression_ratio": 1.579646017699115,
            "no_speech_prob": 0.00033545494079589844,
            "words": [
                {
                    "start": 770.39,
                    "end": 770.89,
                    "word": " for",
                    "probability": 0.49365234375
                },
                {
                    "start": 770.89,
                    "end": 771.19,
                    "word": " whatever",
                    "probability": 0.82666015625
                },
                {
                    "start": 771.19,
                    "end": 771.61,
                    "word": " reason",
                    "probability": 0.9189453125
                },
                {
                    "start": 771.61,
                    "end": 771.89,
                    "word": ".",
                    "probability": 0.24951171875
                },
                {
                    "start": 771.89,
                    "end": 771.91,
                    "word": " And",
                    "probability": 0.73388671875
                },
                {
                    "start": 771.91,
                    "end": 772.43,
                    "word": " I've",
                    "probability": 0.958984375
                },
                {
                    "start": 772.43,
                    "end": 772.61,
                    "word": " made",
                    "probability": 0.66015625
                },
                {
                    "start": 772.61,
                    "end": 772.87,
                    "word": " one",
                    "probability": 0.58740234375
                },
                {
                    "start": 772.87,
                    "end": 773.13,
                    "word": ".",
                    "probability": 0.353515625
                },
                {
                    "start": 773.13,
                    "end": 773.19,
                    "word": " So",
                    "probability": 0.77490234375
                },
                {
                    "start": 773.19,
                    "end": 773.29,
                    "word": ",",
                    "probability": 0.255126953125
                },
                {
                    "start": 773.29,
                    "end": 773.85,
                    "word": " I'm",
                    "probability": 0.984619140625
                },
                {
                    "start": 773.85,
                    "end": 774.03,
                    "word": " allowed",
                    "probability": 0.83740234375
                },
                {
                    "start": 774.03,
                    "end": 774.21,
                    "word": " to",
                    "probability": 0.9033203125
                },
                {
                    "start": 774.21,
                    "end": 774.51,
                    "word": " criticize",
                    "probability": 0.69580078125
                },
                {
                    "start": 774.51,
                    "end": 775.01,
                    "word": ".",
                    "probability": 0.92529296875
                }
            ]
        },
        {
            "id": 109,
            "seek": 79446,
            "start": 777.23,
            "end": 782.29,
            "text": " People love to do it. And that's a really easy way to just be like, nope, everything's benign all",
            "tokens": [
                50752,
                3432,
                959,
                281,
                360,
                309,
                13,
                400,
                300,
                311,
                257,
                534,
                1858,
                636,
                281,
                445,
                312,
                411,
                11,
                23444,
                11,
                1203,
                311,
                3271,
                788,
                439,
                50992
            ],
            "temperature": 0.0,
            "avg_logprob": -0.24842170934484462,
            "compression_ratio": 1.579646017699115,
            "no_speech_prob": 0.00033545494079589844,
            "words": [
                {
                    "start": 777.23,
                    "end": 777.73,
                    "word": " People",
                    "probability": 0.888671875
                },
                {
                    "start": 777.73,
                    "end": 778.03,
                    "word": " love",
                    "probability": 0.89111328125
                },
                {
                    "start": 778.03,
                    "end": 778.17,
                    "word": " to",
                    "probability": 0.90625
                },
                {
                    "start": 778.17,
                    "end": 778.31,
                    "word": " do",
                    "probability": 0.91162109375
                },
                {
                    "start": 778.31,
                    "end": 778.51,
                    "word": " it",
                    "probability": 0.9033203125
                },
                {
                    "start": 778.51,
                    "end": 778.79,
                    "word": ".",
                    "probability": 0.5224609375
                },
                {
                    "start": 778.79,
                    "end": 778.87,
                    "word": " And",
                    "probability": 0.8916015625
                },
                {
                    "start": 778.87,
                    "end": 779.07,
                    "word": " that's",
                    "probability": 0.936767578125
                },
                {
                    "start": 779.19,
                    "end": 779.37,
                    "word": " a really",
                    "probability": 0.77734375
                },
                {
                    "start": 779.37,
                    "end": 779.69,
                    "word": " easy",
                    "probability": 0.82275390625
                },
                {
                    "start": 779.69,
                    "end": 779.97,
                    "word": " way",
                    "probability": 0.8955078125
                },
                {
                    "start": 779.97,
                    "end": 780.19,
                    "word": " to",
                    "probability": 0.8876953125
                },
                {
                    "start": 780.19,
                    "end": 780.45,
                    "word": " just",
                    "probability": 0.82568359375
                },
                {
                    "start": 780.45,
                    "end": 780.67,
                    "word": " be",
                    "probability": 0.90087890625
                },
                {
                    "start": 780.67,
                    "end": 780.89,
                    "word": " like",
                    "probability": 0.79736328125
                },
                {
                    "start": 780.89,
                    "end": 780.95,
                    "word": ",",
                    "probability": 0.880859375
                },
                {
                    "start": 780.95,
                    "end": 781.13,
                    "word": " nope",
                    "probability": 0.5263671875
                },
                {
                    "start": 781.13,
                    "end": 781.27,
                    "word": ",",
                    "probability": 0.80419921875
                },
                {
                    "start": 781.27,
                    "end": 781.75,
                    "word": " everything's",
                    "probability": 0.8427734375
                },
                {
                    "start": 781.75,
                    "end": 782.05,
                    "word": " benign",
                    "probability": 0.947998046875
                },
                {
                    "start": 782.05,
                    "end": 782.29,
                    "word": " all",
                    "probability": 0.78955078125
                }
            ]
        },
        {
            "id": 110,
            "seek": 79446,
            "start": 782.29,
            "end": 791.23,
            "text": " the time. Always. 100% of the time. So, you know, the impact is that it causes misclassification.",
            "tokens": [
                50992,
                264,
                565,
                13,
                11270,
                13,
                2319,
                4,
                295,
                264,
                565,
                13,
                407,
                11,
                291,
                458,
                11,
                264,
                2712,
                307,
                300,
                309,
                7700,
                3346,
                11665,
                3774,
                13,
                51420
            ],
            "temperature": 0.0,
            "avg_logprob": -0.24842170934484462,
            "compression_ratio": 1.579646017699115,
            "no_speech_prob": 0.00033545494079589844,
            "words": [
                {
                    "start": 782.29,
                    "end": 782.43,
                    "word": " the",
                    "probability": 0.810546875
                },
                {
                    "start": 782.43,
                    "end": 782.73,
                    "word": " time",
                    "probability": 0.80126953125
                },
                {
                    "start": 782.73,
                    "end": 783.23,
                    "word": ".",
                    "probability": 0.73193359375
                },
                {
                    "start": 783.31,
                    "end": 783.81,
                    "word": " Always",
                    "probability": 0.9013671875
                },
                {
                    "start": 783.81,
                    "end": 784.07,
                    "word": ".",
                    "probability": 0.53759765625
                },
                {
                    "start": 784.07,
                    "end": 784.35,
                    "word": " 100",
                    "probability": 0.7734375
                },
                {
                    "start": 784.35,
                    "end": 784.55,
                    "word": "%",
                    "probability": 0.94189453125
                },
                {
                    "start": 784.55,
                    "end": 784.71,
                    "word": " of",
                    "probability": 0.896484375
                },
                {
                    "start": 784.71,
                    "end": 784.79,
                    "word": " the",
                    "probability": 0.81298828125
                },
                {
                    "start": 784.79,
                    "end": 785.01,
                    "word": " time",
                    "probability": 0.8017578125
                },
                {
                    "start": 785.01,
                    "end": 785.51,
                    "word": ".",
                    "probability": 0.9921875
                },
                {
                    "start": 786.29,
                    "end": 786.63,
                    "word": " So",
                    "probability": 0.85107421875
                },
                {
                    "start": 786.63,
                    "end": 786.87,
                    "word": ",",
                    "probability": 0.9482421875
                },
                {
                    "start": 786.87,
                    "end": 787.09,
                    "word": " you",
                    "probability": 0.72705078125
                },
                {
                    "start": 787.09,
                    "end": 787.23,
                    "word": " know",
                    "probability": 0.8203125
                },
                {
                    "start": 787.23,
                    "end": 787.31,
                    "word": ",",
                    "probability": 0.90283203125
                },
                {
                    "start": 787.31,
                    "end": 787.59,
                    "word": " the",
                    "probability": 0.806640625
                },
                {
                    "start": 787.59,
                    "end": 788.09,
                    "word": " impact",
                    "probability": 0.8681640625
                },
                {
                    "start": 788.09,
                    "end": 788.75,
                    "word": " is",
                    "probability": 0.88720703125
                },
                {
                    "start": 788.75,
                    "end": 788.97,
                    "word": " that",
                    "probability": 0.8388671875
                },
                {
                    "start": 788.97,
                    "end": 789.21,
                    "word": " it",
                    "probability": 0.89697265625
                },
                {
                    "start": 789.21,
                    "end": 789.67,
                    "word": " causes",
                    "probability": 0.8134765625
                },
                {
                    "start": 789.67,
                    "end": 790.73,
                    "word": " misclassification",
                    "probability": 0.9558919270833334
                },
                {
                    "start": 790.73,
                    "end": 791.23,
                    "word": ".",
                    "probability": 0.82177734375
                }
            ]
        },
        {
            "id": 111,
            "seek": 79446,
            "start": 791.47,
            "end": 798.19,
            "text": " And I say but not really, because really what it does is remove a class or remove all other",
            "tokens": [
                51460,
                400,
                286,
                584,
                457,
                406,
                534,
                11,
                570,
                534,
                437,
                309,
                775,
                307,
                4159,
                257,
                1508,
                420,
                4159,
                439,
                661,
                51788
            ],
            "temperature": 0.0,
            "avg_logprob": -0.24842170934484462,
            "compression_ratio": 1.579646017699115,
            "no_speech_prob": 0.00033545494079589844,
            "words": [
                {
                    "start": 791.47,
                    "end": 791.75,
                    "word": " And",
                    "probability": 0.8818359375
                },
                {
                    "start": 791.75,
                    "end": 791.85,
                    "word": " I",
                    "probability": 0.99609375
                },
                {
                    "start": 791.85,
                    "end": 791.97,
                    "word": " say",
                    "probability": 0.88232421875
                },
                {
                    "start": 791.97,
                    "end": 792.13,
                    "word": " but",
                    "probability": 0.388916015625
                },
                {
                    "start": 792.13,
                    "end": 792.33,
                    "word": " not",
                    "probability": 0.88671875
                },
                {
                    "start": 792.33,
                    "end": 792.79,
                    "word": " really",
                    "probability": 0.78125
                },
                {
                    "start": 792.79,
                    "end": 792.97,
                    "word": ",",
                    "probability": 0.253173828125
                },
                {
                    "start": 792.97,
                    "end": 793.21,
                    "word": " because",
                    "probability": 0.77880859375
                },
                {
                    "start": 793.21,
                    "end": 793.43,
                    "word": " really",
                    "probability": 0.73876953125
                },
                {
                    "start": 793.43,
                    "end": 793.61,
                    "word": " what",
                    "probability": 0.8095703125
                },
                {
                    "start": 793.61,
                    "end": 793.73,
                    "word": " it",
                    "probability": 0.9013671875
                },
                {
                    "start": 793.73,
                    "end": 794.09,
                    "word": " does",
                    "probability": 0.9228515625
                },
                {
                    "start": 794.09,
                    "end": 794.59,
                    "word": " is",
                    "probability": 0.89599609375
                },
                {
                    "start": 794.59,
                    "end": 795.95,
                    "word": " remove",
                    "probability": 0.8779296875
                },
                {
                    "start": 796.17,
                    "end": 796.67,
                    "word": " a class",
                    "probability": 0.91455078125
                },
                {
                    "start": 796.67,
                    "end": 797.13,
                    "word": " or",
                    "probability": 0.461669921875
                },
                {
                    "start": 797.13,
                    "end": 797.55,
                    "word": " remove",
                    "probability": 0.85791015625
                },
                {
                    "start": 797.55,
                    "end": 797.93,
                    "word": " all",
                    "probability": 0.896484375
                },
                {
                    "start": 797.93,
                    "end": 798.19,
                    "word": " other",
                    "probability": 0.755859375
                }
            ]
        },
        {
            "id": 112,
            "seek": 82086,
            "start": 798.19,
            "end": 804.15,
            "text": " classes from the classifier. So, it's not really misclassification. It's just making",
            "tokens": [
                50364,
                5359,
                490,
                264,
                1508,
                9902,
                13,
                407,
                11,
                309,
                311,
                406,
                534,
                3346,
                11665,
                3774,
                13,
                467,
                311,
                445,
                1455,
                50660
            ],
            "temperature": 0.0,
            "avg_logprob": -0.1776221264367816,
            "compression_ratio": 1.7079207920792079,
            "no_speech_prob": 0.0053863525390625,
            "words": [
                {
                    "start": 798.19,
                    "end": 798.69,
                    "word": " classes",
                    "probability": 0.67333984375
                },
                {
                    "start": 798.69,
                    "end": 800.59,
                    "word": " from",
                    "probability": 0.671875
                },
                {
                    "start": 800.59,
                    "end": 800.79,
                    "word": " the",
                    "probability": 0.8154296875
                },
                {
                    "start": 800.79,
                    "end": 801.39,
                    "word": " classifier",
                    "probability": 0.954833984375
                },
                {
                    "start": 801.39,
                    "end": 801.73,
                    "word": ".",
                    "probability": 0.853515625
                },
                {
                    "start": 801.73,
                    "end": 801.79,
                    "word": " So",
                    "probability": 0.498291015625
                },
                {
                    "start": 801.79,
                    "end": 801.81,
                    "word": ",",
                    "probability": 0.6328125
                },
                {
                    "start": 801.81,
                    "end": 801.89,
                    "word": " it's",
                    "probability": 0.942626953125
                },
                {
                    "start": 801.89,
                    "end": 802.03,
                    "word": " not",
                    "probability": 0.8896484375
                },
                {
                    "start": 802.03,
                    "end": 802.21,
                    "word": " really",
                    "probability": 0.771484375
                },
                {
                    "start": 802.21,
                    "end": 803.37,
                    "word": " misclassification",
                    "probability": 0.9220377604166666
                },
                {
                    "start": 803.37,
                    "end": 803.61,
                    "word": ".",
                    "probability": 0.73779296875
                },
                {
                    "start": 803.61,
                    "end": 803.73,
                    "word": " It's",
                    "probability": 0.924560546875
                },
                {
                    "start": 803.73,
                    "end": 803.91,
                    "word": " just",
                    "probability": 0.7958984375
                },
                {
                    "start": 803.91,
                    "end": 804.15,
                    "word": " making",
                    "probability": 0.75537109375
                }
            ]
        },
        {
            "id": 113,
            "seek": 82086,
            "start": 804.15,
            "end": 813.37,
            "text": " it a one class classifier. And our mitigation is just don't let attackers get access to your model.",
            "tokens": [
                50660,
                309,
                257,
                472,
                1508,
                1508,
                9902,
                13,
                400,
                527,
                32649,
                307,
                445,
                500,
                380,
                718,
                45129,
                483,
                2105,
                281,
                428,
                2316,
                13,
                51112
            ],
            "temperature": 0.0,
            "avg_logprob": -0.1776221264367816,
            "compression_ratio": 1.7079207920792079,
            "no_speech_prob": 0.0053863525390625,
            "words": [
                {
                    "start": 804.15,
                    "end": 804.33,
                    "word": " it",
                    "probability": 0.8876953125
                },
                {
                    "start": 804.47,
                    "end": 804.73,
                    "word": " a one",
                    "probability": 0.79638671875
                },
                {
                    "start": 804.73,
                    "end": 805.13,
                    "word": " class",
                    "probability": 0.634765625
                },
                {
                    "start": 805.13,
                    "end": 805.83,
                    "word": " classifier",
                    "probability": 0.962158203125
                },
                {
                    "start": 805.83,
                    "end": 806.39,
                    "word": ".",
                    "probability": 0.94775390625
                },
                {
                    "start": 807.75,
                    "end": 808.19,
                    "word": " And",
                    "probability": 0.8232421875
                },
                {
                    "start": 808.19,
                    "end": 808.45,
                    "word": " our",
                    "probability": 0.81982421875
                },
                {
                    "start": 808.45,
                    "end": 808.89,
                    "word": " mitigation",
                    "probability": 0.8134765625
                },
                {
                    "start": 808.89,
                    "end": 809.75,
                    "word": " is",
                    "probability": 0.896484375
                },
                {
                    "start": 809.75,
                    "end": 810.35,
                    "word": " just",
                    "probability": 0.307861328125
                },
                {
                    "start": 810.35,
                    "end": 810.71,
                    "word": " don't",
                    "probability": 0.909912109375
                },
                {
                    "start": 810.71,
                    "end": 810.89,
                    "word": " let",
                    "probability": 0.888671875
                },
                {
                    "start": 810.89,
                    "end": 811.33,
                    "word": " attackers",
                    "probability": 0.8564453125
                },
                {
                    "start": 811.33,
                    "end": 811.55,
                    "word": " get",
                    "probability": 0.89111328125
                },
                {
                    "start": 811.55,
                    "end": 811.89,
                    "word": " access",
                    "probability": 0.79248046875
                },
                {
                    "start": 811.89,
                    "end": 812.23,
                    "word": " to",
                    "probability": 0.90234375
                },
                {
                    "start": 812.23,
                    "end": 812.43,
                    "word": " your",
                    "probability": 0.810546875
                },
                {
                    "start": 812.43,
                    "end": 812.83,
                    "word": " model",
                    "probability": 0.82763671875
                },
                {
                    "start": 812.83,
                    "end": 813.37,
                    "word": ".",
                    "probability": 0.7880859375
                }
            ]
        },
        {
            "id": 114,
            "seek": 82086,
            "start": 813.37,
            "end": 819.87,
            "text": " And by your model, I mean the trained model weights. And I say this kind of flippantly,",
            "tokens": [
                51112,
                400,
                538,
                428,
                2316,
                11,
                286,
                914,
                264,
                8895,
                2316,
                17443,
                13,
                400,
                286,
                584,
                341,
                733,
                295,
                932,
                2488,
                3627,
                11,
                51444
            ],
            "temperature": 0.0,
            "avg_logprob": -0.1776221264367816,
            "compression_ratio": 1.7079207920792079,
            "no_speech_prob": 0.0053863525390625,
            "words": [
                {
                    "start": 813.37,
                    "end": 813.45,
                    "word": " And",
                    "probability": 0.7861328125
                },
                {
                    "start": 813.45,
                    "end": 813.55,
                    "word": " by",
                    "probability": 0.896484375
                },
                {
                    "start": 813.55,
                    "end": 813.73,
                    "word": " your",
                    "probability": 0.72021484375
                },
                {
                    "start": 813.73,
                    "end": 814.01,
                    "word": " model",
                    "probability": 0.8212890625
                },
                {
                    "start": 814.01,
                    "end": 814.17,
                    "word": ",",
                    "probability": 0.9287109375
                },
                {
                    "start": 814.17,
                    "end": 814.29,
                    "word": " I",
                    "probability": 0.99462890625
                },
                {
                    "start": 814.29,
                    "end": 814.51,
                    "word": " mean",
                    "probability": 0.921875
                },
                {
                    "start": 814.51,
                    "end": 814.79,
                    "word": " the",
                    "probability": 0.76708984375
                },
                {
                    "start": 814.79,
                    "end": 815.23,
                    "word": " trained",
                    "probability": 0.80029296875
                },
                {
                    "start": 815.23,
                    "end": 815.61,
                    "word": " model",
                    "probability": 0.85791015625
                },
                {
                    "start": 815.61,
                    "end": 815.99,
                    "word": " weights",
                    "probability": 0.5830078125
                },
                {
                    "start": 815.99,
                    "end": 816.55,
                    "word": ".",
                    "probability": 0.9306640625
                },
                {
                    "start": 817.73,
                    "end": 817.91,
                    "word": " And",
                    "probability": 0.83154296875
                },
                {
                    "start": 817.91,
                    "end": 818.63,
                    "word": " I",
                    "probability": 0.9892578125
                },
                {
                    "start": 818.63,
                    "end": 818.81,
                    "word": " say",
                    "probability": 0.86865234375
                },
                {
                    "start": 818.81,
                    "end": 819.01,
                    "word": " this",
                    "probability": 0.8759765625
                },
                {
                    "start": 819.01,
                    "end": 819.17,
                    "word": " kind",
                    "probability": 0.81982421875
                },
                {
                    "start": 819.17,
                    "end": 819.25,
                    "word": " of",
                    "probability": 0.90380859375
                },
                {
                    "start": 819.25,
                    "end": 819.73,
                    "word": " flippantly",
                    "probability": 0.904296875
                },
                {
                    "start": 819.73,
                    "end": 819.87,
                    "word": ",",
                    "probability": 0.28662109375
                }
            ]
        },
        {
            "id": 115,
            "seek": 82086,
            "start": 819.87,
            "end": 824.59,
            "text": " because if an attacker has the ability to manipulate your model weights,",
            "tokens": [
                51444,
                570,
                498,
                364,
                35871,
                575,
                264,
                3485,
                281,
                20459,
                428,
                2316,
                17443,
                11,
                51648
            ],
            "temperature": 0.0,
            "avg_logprob": -0.1776221264367816,
            "compression_ratio": 1.7079207920792079,
            "no_speech_prob": 0.0053863525390625,
            "words": [
                {
                    "start": 819.87,
                    "end": 820.15,
                    "word": " because",
                    "probability": 0.818359375
                },
                {
                    "start": 820.15,
                    "end": 820.43,
                    "word": " if",
                    "probability": 0.88232421875
                },
                {
                    "start": 820.43,
                    "end": 820.59,
                    "word": " an",
                    "probability": 0.90380859375
                },
                {
                    "start": 820.59,
                    "end": 820.95,
                    "word": " attacker",
                    "probability": 0.7412109375
                },
                {
                    "start": 820.95,
                    "end": 821.33,
                    "word": " has",
                    "probability": 0.89697265625
                },
                {
                    "start": 821.33,
                    "end": 821.85,
                    "word": " the",
                    "probability": 0.81005859375
                },
                {
                    "start": 821.85,
                    "end": 822.21,
                    "word": " ability",
                    "probability": 0.8642578125
                },
                {
                    "start": 822.21,
                    "end": 822.49,
                    "word": " to",
                    "probability": 0.90625
                },
                {
                    "start": 822.49,
                    "end": 822.91,
                    "word": " manipulate",
                    "probability": 0.92041015625
                },
                {
                    "start": 822.91,
                    "end": 823.17,
                    "word": " your",
                    "probability": 0.8017578125
                },
                {
                    "start": 823.17,
                    "end": 823.47,
                    "word": " model",
                    "probability": 0.84033203125
                },
                {
                    "start": 823.47,
                    "end": 823.99,
                    "word": " weights",
                    "probability": 0.69775390625
                },
                {
                    "start": 823.99,
                    "end": 824.59,
                    "word": ",",
                    "probability": 0.560546875
                }
            ]
        },
        {
            "id": 116,
            "seek": 84744,
            "start": 824.59,
            "end": 831.73,
            "text": " where you're hosting your model, they can do way worse things than backdoor your model.",
            "tokens": [
                50380,
                689,
                291,
                434,
                16058,
                428,
                2316,
                11,
                436,
                393,
                360,
                636,
                5324,
                721,
                813,
                646,
                10441,
                428,
                2316,
                13,
                50680
            ],
            "temperature": 0.0,
            "avg_logprob": -0.17945429237920846,
            "compression_ratio": 1.5416666666666667,
            "no_speech_prob": 0.0007319450378417969,
            "words": [
                {
                    "start": 824.59,
                    "end": 825.27,
                    "word": " where",
                    "probability": 0.49462890625
                },
                {
                    "start": 825.27,
                    "end": 825.57,
                    "word": " you're",
                    "probability": 0.913330078125
                },
                {
                    "start": 825.57,
                    "end": 825.93,
                    "word": " hosting",
                    "probability": 0.79248046875
                },
                {
                    "start": 825.93,
                    "end": 826.19,
                    "word": " your",
                    "probability": 0.80908203125
                },
                {
                    "start": 826.19,
                    "end": 826.51,
                    "word": " model",
                    "probability": 0.81005859375
                },
                {
                    "start": 826.51,
                    "end": 826.79,
                    "word": ",",
                    "probability": 0.63720703125
                },
                {
                    "start": 826.79,
                    "end": 827.03,
                    "word": " they",
                    "probability": 0.72998046875
                },
                {
                    "start": 827.03,
                    "end": 827.21,
                    "word": " can",
                    "probability": 0.8994140625
                },
                {
                    "start": 827.21,
                    "end": 827.51,
                    "word": " do",
                    "probability": 0.9091796875
                },
                {
                    "start": 827.51,
                    "end": 828.13,
                    "word": " way",
                    "probability": 0.8984375
                },
                {
                    "start": 828.13,
                    "end": 828.55,
                    "word": " worse",
                    "probability": 0.880859375
                },
                {
                    "start": 828.55,
                    "end": 829.05,
                    "word": " things",
                    "probability": 0.7333984375
                },
                {
                    "start": 829.05,
                    "end": 829.33,
                    "word": " than",
                    "probability": 0.859375
                },
                {
                    "start": 829.33,
                    "end": 829.91,
                    "word": " backdoor",
                    "probability": 0.8701171875
                },
                {
                    "start": 829.91,
                    "end": 830.23,
                    "word": " your",
                    "probability": 0.796875
                },
                {
                    "start": 830.23,
                    "end": 830.97,
                    "word": " model",
                    "probability": 0.81982421875
                },
                {
                    "start": 830.97,
                    "end": 831.73,
                    "word": ".",
                    "probability": 0.91552734375
                }
            ]
        },
        {
            "id": 117,
            "seek": 84744,
            "start": 832.11,
            "end": 841.37,
            "text": " So, it's not... it's like a local privilege escalation that requires you to have admin",
            "tokens": [
                50728,
                407,
                11,
                309,
                311,
                406,
                485,
                309,
                311,
                411,
                257,
                2654,
                12122,
                17871,
                399,
                300,
                7029,
                291,
                281,
                362,
                24236,
                51204
            ],
            "temperature": 0.0,
            "avg_logprob": -0.17945429237920846,
            "compression_ratio": 1.5416666666666667,
            "no_speech_prob": 0.0007319450378417969,
            "words": [
                {
                    "start": 832.11,
                    "end": 832.25,
                    "word": " So",
                    "probability": 0.72509765625
                },
                {
                    "start": 832.25,
                    "end": 832.47,
                    "word": ",",
                    "probability": 0.2052001953125
                },
                {
                    "start": 832.47,
                    "end": 833.05,
                    "word": " it's",
                    "probability": 0.950927734375
                },
                {
                    "start": 833.05,
                    "end": 835.11,
                    "word": " not...",
                    "probability": 0.37744140625
                },
                {
                    "start": 835.11,
                    "end": 836.09,
                    "word": " it's",
                    "probability": 0.59423828125
                },
                {
                    "start": 836.09,
                    "end": 836.35,
                    "word": " like",
                    "probability": 0.8505859375
                },
                {
                    "start": 837.65,
                    "end": 838.05,
                    "word": " a local",
                    "probability": 0.728515625
                },
                {
                    "start": 838.05,
                    "end": 838.71,
                    "word": " privilege",
                    "probability": 0.6865234375
                },
                {
                    "start": 838.71,
                    "end": 839.59,
                    "word": " escalation",
                    "probability": 0.903564453125
                },
                {
                    "start": 839.59,
                    "end": 839.85,
                    "word": " that",
                    "probability": 0.87158203125
                },
                {
                    "start": 839.85,
                    "end": 840.39,
                    "word": " requires",
                    "probability": 0.77587890625
                },
                {
                    "start": 840.39,
                    "end": 840.59,
                    "word": " you",
                    "probability": 0.89794921875
                },
                {
                    "start": 840.59,
                    "end": 840.69,
                    "word": " to",
                    "probability": 0.9033203125
                },
                {
                    "start": 840.69,
                    "end": 840.89,
                    "word": " have",
                    "probability": 0.880859375
                },
                {
                    "start": 840.89,
                    "end": 841.37,
                    "word": " admin",
                    "probability": 0.8408203125
                }
            ]
        },
        {
            "id": 118,
            "seek": 84744,
            "start": 841.37,
            "end": 851.17,
            "text": " to begin with. It's not really a threat. So, again, attacks against deployed models.",
            "tokens": [
                51204,
                281,
                1841,
                365,
                13,
                467,
                311,
                406,
                534,
                257,
                4734,
                13,
                407,
                11,
                797,
                11,
                8122,
                1970,
                17826,
                5245,
                13,
                51684
            ],
            "temperature": 0.0,
            "avg_logprob": -0.17945429237920846,
            "compression_ratio": 1.5416666666666667,
            "no_speech_prob": 0.0007319450378417969,
            "words": [
                {
                    "start": 841.37,
                    "end": 841.65,
                    "word": " to",
                    "probability": 0.89697265625
                },
                {
                    "start": 841.65,
                    "end": 841.91,
                    "word": " begin",
                    "probability": 0.72119140625
                },
                {
                    "start": 841.91,
                    "end": 842.27,
                    "word": " with",
                    "probability": 0.81689453125
                },
                {
                    "start": 842.27,
                    "end": 842.89,
                    "word": ".",
                    "probability": 0.91455078125
                },
                {
                    "start": 842.89,
                    "end": 843.29,
                    "word": " It's",
                    "probability": 0.947998046875
                },
                {
                    "start": 843.29,
                    "end": 843.63,
                    "word": " not",
                    "probability": 0.89599609375
                },
                {
                    "start": 843.63,
                    "end": 844.37,
                    "word": " really",
                    "probability": 0.77001953125
                },
                {
                    "start": 844.65,
                    "end": 844.95,
                    "word": " a threat",
                    "probability": 0.78759765625
                },
                {
                    "start": 844.95,
                    "end": 845.71,
                    "word": ".",
                    "probability": 0.994140625
                },
                {
                    "start": 847.29,
                    "end": 848.05,
                    "word": " So",
                    "probability": 0.82373046875
                },
                {
                    "start": 848.05,
                    "end": 848.57,
                    "word": ",",
                    "probability": 0.67822265625
                },
                {
                    "start": 848.57,
                    "end": 849.01,
                    "word": " again",
                    "probability": 0.9248046875
                },
                {
                    "start": 849.01,
                    "end": 849.13,
                    "word": ",",
                    "probability": 0.916015625
                },
                {
                    "start": 849.13,
                    "end": 849.41,
                    "word": " attacks",
                    "probability": 0.73193359375
                },
                {
                    "start": 849.41,
                    "end": 849.73,
                    "word": " against",
                    "probability": 0.89013671875
                },
                {
                    "start": 849.73,
                    "end": 850.13,
                    "word": " deployed",
                    "probability": 0.7353515625
                },
                {
                    "start": 850.13,
                    "end": 850.59,
                    "word": " models",
                    "probability": 0.82421875
                },
                {
                    "start": 850.59,
                    "end": 851.17,
                    "word": ".",
                    "probability": 0.54296875
                }
            ]
        },
        {
            "id": 119,
            "seek": 87678,
            "start": 851.17,
            "end": 857.21,
            "text": " There's model theft. And I want to talk about model theft and give a lot of credit to Will",
            "tokens": [
                50364,
                821,
                311,
                2316,
                28508,
                13,
                400,
                286,
                528,
                281,
                751,
                466,
                2316,
                28508,
                293,
                976,
                257,
                688,
                295,
                5397,
                281,
                3099,
                50668
            ],
            "temperature": 0.0,
            "avg_logprob": -0.24062499420510397,
            "compression_ratio": 1.5569620253164558,
            "no_speech_prob": 0.0074615478515625,
            "words": [
                {
                    "start": 851.17,
                    "end": 851.45,
                    "word": " There's",
                    "probability": 0.623046875
                },
                {
                    "start": 851.45,
                    "end": 851.71,
                    "word": " model",
                    "probability": 0.728515625
                },
                {
                    "start": 851.71,
                    "end": 852.35,
                    "word": " theft",
                    "probability": 0.8232421875
                },
                {
                    "start": 852.35,
                    "end": 852.75,
                    "word": ".",
                    "probability": 0.46337890625
                },
                {
                    "start": 852.75,
                    "end": 853.03,
                    "word": " And",
                    "probability": 0.826171875
                },
                {
                    "start": 853.03,
                    "end": 853.29,
                    "word": " I",
                    "probability": 0.98974609375
                },
                {
                    "start": 853.29,
                    "end": 853.43,
                    "word": " want",
                    "probability": 0.50048828125
                },
                {
                    "start": 853.43,
                    "end": 853.49,
                    "word": " to",
                    "probability": 0.90625
                },
                {
                    "start": 853.49,
                    "end": 853.67,
                    "word": " talk",
                    "probability": 0.810546875
                },
                {
                    "start": 853.67,
                    "end": 853.91,
                    "word": " about",
                    "probability": 0.8154296875
                },
                {
                    "start": 853.91,
                    "end": 854.19,
                    "word": " model",
                    "probability": 0.80517578125
                },
                {
                    "start": 854.19,
                    "end": 854.79,
                    "word": " theft",
                    "probability": 0.7451171875
                },
                {
                    "start": 854.79,
                    "end": 855.97,
                    "word": " and",
                    "probability": 0.74951171875
                },
                {
                    "start": 855.97,
                    "end": 856.15,
                    "word": " give",
                    "probability": 0.81103515625
                },
                {
                    "start": 856.27,
                    "end": 856.39,
                    "word": " a lot",
                    "probability": 0.89990234375
                },
                {
                    "start": 856.39,
                    "end": 856.53,
                    "word": " of",
                    "probability": 0.9033203125
                },
                {
                    "start": 856.53,
                    "end": 856.73,
                    "word": " credit",
                    "probability": 0.83447265625
                },
                {
                    "start": 856.73,
                    "end": 856.97,
                    "word": " to",
                    "probability": 0.9111328125
                },
                {
                    "start": 856.97,
                    "end": 857.21,
                    "word": " Will",
                    "probability": 0.81787109375
                }
            ]
        },
        {
            "id": 120,
            "seek": 87678,
            "start": 857.21,
            "end": 864.67,
            "text": " Pierce, formerly of Silent Break Security, who I stole this graphic from his fantastic DerbyCon",
            "tokens": [
                50668,
                45432,
                11,
                34777,
                295,
                40862,
                16925,
                11164,
                11,
                567,
                286,
                16326,
                341,
                14089,
                490,
                702,
                5456,
                5618,
                2322,
                9838,
                51040
            ],
            "temperature": 0.0,
            "avg_logprob": -0.24062499420510397,
            "compression_ratio": 1.5569620253164558,
            "no_speech_prob": 0.0074615478515625,
            "words": [
                {
                    "start": 857.21,
                    "end": 857.75,
                    "word": " Pierce",
                    "probability": 0.591796875
                },
                {
                    "start": 857.75,
                    "end": 858.49,
                    "word": ",",
                    "probability": 0.8203125
                },
                {
                    "start": 858.49,
                    "end": 859.27,
                    "word": " formerly",
                    "probability": 0.865234375
                },
                {
                    "start": 859.27,
                    "end": 859.75,
                    "word": " of",
                    "probability": 0.90576171875
                },
                {
                    "start": 859.75,
                    "end": 860.07,
                    "word": " Silent",
                    "probability": 0.6748046875
                },
                {
                    "start": 860.07,
                    "end": 860.29,
                    "word": " Break",
                    "probability": 0.156982421875
                },
                {
                    "start": 860.29,
                    "end": 860.71,
                    "word": " Security",
                    "probability": 0.767578125
                },
                {
                    "start": 860.71,
                    "end": 861.29,
                    "word": ",",
                    "probability": 0.845703125
                },
                {
                    "start": 861.29,
                    "end": 861.43,
                    "word": " who",
                    "probability": 0.8134765625
                },
                {
                    "start": 861.43,
                    "end": 861.71,
                    "word": " I",
                    "probability": 0.62890625
                },
                {
                    "start": 861.71,
                    "end": 861.95,
                    "word": " stole",
                    "probability": 0.74951171875
                },
                {
                    "start": 861.95,
                    "end": 862.17,
                    "word": " this",
                    "probability": 0.87939453125
                },
                {
                    "start": 862.17,
                    "end": 862.49,
                    "word": " graphic",
                    "probability": 0.75341796875
                },
                {
                    "start": 862.49,
                    "end": 862.79,
                    "word": " from",
                    "probability": 0.8193359375
                },
                {
                    "start": 862.79,
                    "end": 863.01,
                    "word": " his",
                    "probability": 0.8759765625
                },
                {
                    "start": 863.01,
                    "end": 863.99,
                    "word": " fantastic",
                    "probability": 0.8896484375
                },
                {
                    "start": 863.99,
                    "end": 864.67,
                    "word": " DerbyCon",
                    "probability": 0.7394205729166666
                }
            ]
        },
        {
            "id": 121,
            "seek": 87678,
            "start": 864.67,
            "end": 873.75,
            "text": " presentation last year. Was given the first CVE for machine learning, which is super dope.",
            "tokens": [
                51040,
                5860,
                1036,
                1064,
                13,
                3027,
                2212,
                264,
                700,
                383,
                7540,
                337,
                3479,
                2539,
                11,
                597,
                307,
                1687,
                23383,
                13,
                51480
            ],
            "temperature": 0.0,
            "avg_logprob": -0.24062499420510397,
            "compression_ratio": 1.5569620253164558,
            "no_speech_prob": 0.0074615478515625,
            "words": [
                {
                    "start": 864.67,
                    "end": 865.35,
                    "word": " presentation",
                    "probability": 0.78466796875
                },
                {
                    "start": 865.35,
                    "end": 865.87,
                    "word": " last",
                    "probability": 0.806640625
                },
                {
                    "start": 865.87,
                    "end": 866.19,
                    "word": " year",
                    "probability": 0.88623046875
                },
                {
                    "start": 866.19,
                    "end": 866.79,
                    "word": ".",
                    "probability": 0.53369140625
                },
                {
                    "start": 867.81,
                    "end": 868.41,
                    "word": " Was",
                    "probability": 0.69775390625
                },
                {
                    "start": 868.41,
                    "end": 868.67,
                    "word": " given",
                    "probability": 0.80615234375
                },
                {
                    "start": 868.67,
                    "end": 868.95,
                    "word": " the",
                    "probability": 0.8203125
                },
                {
                    "start": 868.95,
                    "end": 869.25,
                    "word": " first",
                    "probability": 0.8076171875
                },
                {
                    "start": 869.25,
                    "end": 869.79,
                    "word": " CVE",
                    "probability": 0.951904296875
                },
                {
                    "start": 869.79,
                    "end": 870.57,
                    "word": " for",
                    "probability": 0.890625
                },
                {
                    "start": 870.57,
                    "end": 871.45,
                    "word": " machine",
                    "probability": 0.65771484375
                },
                {
                    "start": 871.45,
                    "end": 871.79,
                    "word": " learning",
                    "probability": 0.8505859375
                },
                {
                    "start": 871.79,
                    "end": 871.95,
                    "word": ",",
                    "probability": 0.94482421875
                },
                {
                    "start": 871.95,
                    "end": 872.09,
                    "word": " which",
                    "probability": 0.919921875
                },
                {
                    "start": 872.09,
                    "end": 872.31,
                    "word": " is",
                    "probability": 0.90380859375
                },
                {
                    "start": 872.31,
                    "end": 873.11,
                    "word": " super",
                    "probability": 0.86328125
                },
                {
                    "start": 873.11,
                    "end": 873.45,
                    "word": " dope",
                    "probability": 0.7119140625
                },
                {
                    "start": 873.45,
                    "end": 873.75,
                    "word": ".",
                    "probability": 0.984375
                }
            ]
        },
        {
            "id": 122,
            "seek": 87678,
            "start": 874.51,
            "end": 880.51,
            "text": " And so, model theft is the process of creating a copycat model by querying a trained model.",
            "tokens": [
                51528,
                400,
                370,
                11,
                2316,
                28508,
                307,
                264,
                1399,
                295,
                4084,
                257,
                5055,
                18035,
                2316,
                538,
                7083,
                1840,
                257,
                8895,
                2316,
                13,
                51800
            ],
            "temperature": 0.0,
            "avg_logprob": -0.24062499420510397,
            "compression_ratio": 1.5569620253164558,
            "no_speech_prob": 0.0074615478515625,
            "words": [
                {
                    "start": 874.51,
                    "end": 874.63,
                    "word": " And",
                    "probability": 0.87353515625
                },
                {
                    "start": 874.63,
                    "end": 874.85,
                    "word": " so",
                    "probability": 0.8955078125
                },
                {
                    "start": 874.85,
                    "end": 875.17,
                    "word": ",",
                    "probability": 0.16015625
                },
                {
                    "start": 875.17,
                    "end": 875.27,
                    "word": " model",
                    "probability": 0.794921875
                },
                {
                    "start": 875.27,
                    "end": 875.63,
                    "word": " theft",
                    "probability": 0.7626953125
                },
                {
                    "start": 875.63,
                    "end": 875.81,
                    "word": " is",
                    "probability": 0.90087890625
                },
                {
                    "start": 875.81,
                    "end": 875.97,
                    "word": " the",
                    "probability": 0.8134765625
                },
                {
                    "start": 875.97,
                    "end": 876.35,
                    "word": " process",
                    "probability": 0.8779296875
                },
                {
                    "start": 876.35,
                    "end": 876.53,
                    "word": " of",
                    "probability": 0.90380859375
                },
                {
                    "start": 876.53,
                    "end": 876.83,
                    "word": " creating",
                    "probability": 0.79345703125
                },
                {
                    "start": 877.23,
                    "end": 877.89,
                    "word": " a copycat",
                    "probability": 0.8515625
                },
                {
                    "start": 877.89,
                    "end": 878.29,
                    "word": " model",
                    "probability": 0.81298828125
                },
                {
                    "start": 878.29,
                    "end": 878.59,
                    "word": " by",
                    "probability": 0.892578125
                },
                {
                    "start": 878.59,
                    "end": 879.05,
                    "word": " querying",
                    "probability": 0.94482421875
                },
                {
                    "start": 879.23,
                    "end": 879.49,
                    "word": " a trained",
                    "probability": 0.78271484375
                },
                {
                    "start": 879.49,
                    "end": 879.91,
                    "word": " model",
                    "probability": 0.81396484375
                },
                {
                    "start": 879.91,
                    "end": 880.51,
                    "word": ".",
                    "probability": 0.72900390625
                }
            ]
        },
        {
            "id": 123,
            "seek": 90678,
            "start": 880.51,
            "end": 886.79,
            "text": " Right? So, essentially, what we're doing is we're hitting a model over and over and over again.",
            "tokens": [
                50376,
                1779,
                30,
                407,
                11,
                4476,
                11,
                437,
                321,
                434,
                884,
                307,
                321,
                434,
                8850,
                257,
                2316,
                670,
                293,
                670,
                293,
                670,
                797,
                13,
                50644
            ],
            "temperature": 0.0,
            "avg_logprob": -0.1870312486588955,
            "compression_ratio": 1.7655502392344498,
            "no_speech_prob": 0.00011235475540161133,
            "words": [
                {
                    "start": 880.51,
                    "end": 881.07,
                    "word": " Right",
                    "probability": 0.65087890625
                },
                {
                    "start": 881.07,
                    "end": 881.71,
                    "word": "?",
                    "probability": 0.25
                },
                {
                    "start": 881.71,
                    "end": 882.15,
                    "word": " So",
                    "probability": 0.8515625
                },
                {
                    "start": 882.15,
                    "end": 882.27,
                    "word": ",",
                    "probability": 0.357177734375
                },
                {
                    "start": 882.27,
                    "end": 882.69,
                    "word": " essentially",
                    "probability": 0.662109375
                },
                {
                    "start": 882.69,
                    "end": 882.85,
                    "word": ",",
                    "probability": 0.2890625
                },
                {
                    "start": 882.85,
                    "end": 882.99,
                    "word": " what",
                    "probability": 0.89306640625
                },
                {
                    "start": 882.99,
                    "end": 883.17,
                    "word": " we're",
                    "probability": 0.923095703125
                },
                {
                    "start": 883.17,
                    "end": 883.53,
                    "word": " doing",
                    "probability": 0.91845703125
                },
                {
                    "start": 883.53,
                    "end": 884.05,
                    "word": " is",
                    "probability": 0.8388671875
                },
                {
                    "start": 884.05,
                    "end": 884.19,
                    "word": " we're",
                    "probability": 0.903076171875
                },
                {
                    "start": 884.19,
                    "end": 884.41,
                    "word": " hitting",
                    "probability": 0.81494140625
                },
                {
                    "start": 884.61,
                    "end": 884.79,
                    "word": " a model",
                    "probability": 0.81201171875
                },
                {
                    "start": 884.79,
                    "end": 885.11,
                    "word": " over",
                    "probability": 0.87451171875
                },
                {
                    "start": 885.11,
                    "end": 885.27,
                    "word": " and",
                    "probability": 0.82763671875
                },
                {
                    "start": 885.27,
                    "end": 885.47,
                    "word": " over",
                    "probability": 0.88232421875
                },
                {
                    "start": 885.47,
                    "end": 885.61,
                    "word": " and",
                    "probability": 0.8603515625
                },
                {
                    "start": 885.61,
                    "end": 885.77,
                    "word": " over",
                    "probability": 0.87939453125
                },
                {
                    "start": 885.77,
                    "end": 886.15,
                    "word": " again",
                    "probability": 0.90869140625
                },
                {
                    "start": 886.15,
                    "end": 886.79,
                    "word": ".",
                    "probability": 0.483154296875
                }
            ]
        },
        {
            "id": 124,
            "seek": 90678,
            "start": 886.91,
            "end": 895.23,
            "text": " And using the outputs of that trained model, we train our own model to give us an approximation",
            "tokens": [
                50676,
                400,
                1228,
                264,
                23930,
                295,
                300,
                8895,
                2316,
                11,
                321,
                3847,
                527,
                1065,
                2316,
                281,
                976,
                505,
                364,
                28023,
                51112
            ],
            "temperature": 0.0,
            "avg_logprob": -0.1870312486588955,
            "compression_ratio": 1.7655502392344498,
            "no_speech_prob": 0.00011235475540161133,
            "words": [
                {
                    "start": 886.91,
                    "end": 887.19,
                    "word": " And",
                    "probability": 0.87646484375
                },
                {
                    "start": 887.19,
                    "end": 887.93,
                    "word": " using",
                    "probability": 0.7666015625
                },
                {
                    "start": 887.93,
                    "end": 888.25,
                    "word": " the",
                    "probability": 0.8134765625
                },
                {
                    "start": 888.25,
                    "end": 888.79,
                    "word": " outputs",
                    "probability": 0.74072265625
                },
                {
                    "start": 888.79,
                    "end": 888.97,
                    "word": " of",
                    "probability": 0.90625
                },
                {
                    "start": 888.97,
                    "end": 889.19,
                    "word": " that",
                    "probability": 0.8759765625
                },
                {
                    "start": 889.19,
                    "end": 889.63,
                    "word": " trained",
                    "probability": 0.7978515625
                },
                {
                    "start": 889.63,
                    "end": 890.05,
                    "word": " model",
                    "probability": 0.82373046875
                },
                {
                    "start": 890.05,
                    "end": 890.43,
                    "word": ",",
                    "probability": 0.93212890625
                },
                {
                    "start": 890.43,
                    "end": 890.75,
                    "word": " we",
                    "probability": 0.89990234375
                },
                {
                    "start": 890.75,
                    "end": 891.23,
                    "word": " train",
                    "probability": 0.80224609375
                },
                {
                    "start": 891.23,
                    "end": 891.65,
                    "word": " our",
                    "probability": 0.8193359375
                },
                {
                    "start": 891.65,
                    "end": 892.29,
                    "word": " own",
                    "probability": 0.84765625
                },
                {
                    "start": 892.29,
                    "end": 892.73,
                    "word": " model",
                    "probability": 0.80908203125
                },
                {
                    "start": 892.73,
                    "end": 893.55,
                    "word": " to",
                    "probability": 0.8916015625
                },
                {
                    "start": 893.55,
                    "end": 893.83,
                    "word": " give",
                    "probability": 0.8251953125
                },
                {
                    "start": 893.83,
                    "end": 894.13,
                    "word": " us",
                    "probability": 0.90673828125
                },
                {
                    "start": 894.13,
                    "end": 894.49,
                    "word": " an",
                    "probability": 0.90625
                },
                {
                    "start": 894.49,
                    "end": 895.23,
                    "word": " approximation",
                    "probability": 0.78857421875
                }
            ]
        },
        {
            "id": 125,
            "seek": 90678,
            "start": 895.23,
            "end": 905.31,
            "text": " of that model. So, you know, if you have a box and you put in a 1 and you get out a 2,",
            "tokens": [
                51112,
                295,
                300,
                2316,
                13,
                407,
                11,
                291,
                458,
                11,
                498,
                291,
                362,
                257,
                2424,
                293,
                291,
                829,
                294,
                257,
                502,
                293,
                291,
                483,
                484,
                257,
                568,
                11,
                51568
            ],
            "temperature": 0.0,
            "avg_logprob": -0.1870312486588955,
            "compression_ratio": 1.7655502392344498,
            "no_speech_prob": 0.00011235475540161133,
            "words": [
                {
                    "start": 895.23,
                    "end": 895.69,
                    "word": " of",
                    "probability": 0.88037109375
                },
                {
                    "start": 895.69,
                    "end": 895.87,
                    "word": " that",
                    "probability": 0.88330078125
                },
                {
                    "start": 895.87,
                    "end": 896.21,
                    "word": " model",
                    "probability": 0.80908203125
                },
                {
                    "start": 896.21,
                    "end": 896.85,
                    "word": ".",
                    "probability": 0.9892578125
                },
                {
                    "start": 896.85,
                    "end": 897.21,
                    "word": " So",
                    "probability": 0.87841796875
                },
                {
                    "start": 897.21,
                    "end": 897.55,
                    "word": ",",
                    "probability": 0.92333984375
                },
                {
                    "start": 897.55,
                    "end": 897.85,
                    "word": " you",
                    "probability": 0.802734375
                },
                {
                    "start": 897.85,
                    "end": 897.99,
                    "word": " know",
                    "probability": 0.81640625
                },
                {
                    "start": 897.99,
                    "end": 898.63,
                    "word": ",",
                    "probability": 0.97509765625
                },
                {
                    "start": 898.63,
                    "end": 898.85,
                    "word": " if",
                    "probability": 0.90234375
                },
                {
                    "start": 898.85,
                    "end": 899.03,
                    "word": " you",
                    "probability": 0.8935546875
                },
                {
                    "start": 899.03,
                    "end": 899.47,
                    "word": " have",
                    "probability": 0.88037109375
                },
                {
                    "start": 899.81,
                    "end": 900.39,
                    "word": " a box",
                    "probability": 0.80859375
                },
                {
                    "start": 900.39,
                    "end": 900.93,
                    "word": " and",
                    "probability": 0.6611328125
                },
                {
                    "start": 900.93,
                    "end": 901.09,
                    "word": " you",
                    "probability": 0.89111328125
                },
                {
                    "start": 901.09,
                    "end": 901.43,
                    "word": " put",
                    "probability": 0.88916015625
                },
                {
                    "start": 901.43,
                    "end": 901.89,
                    "word": " in",
                    "probability": 0.90625
                },
                {
                    "start": 902.41,
                    "end": 902.83,
                    "word": " a 1",
                    "probability": 0.381103515625
                },
                {
                    "start": 902.83,
                    "end": 903.39,
                    "word": " and",
                    "probability": 0.7509765625
                },
                {
                    "start": 903.39,
                    "end": 903.57,
                    "word": " you",
                    "probability": 0.876953125
                },
                {
                    "start": 903.57,
                    "end": 903.81,
                    "word": " get",
                    "probability": 0.88671875
                },
                {
                    "start": 903.81,
                    "end": 904.03,
                    "word": " out",
                    "probability": 0.828125
                },
                {
                    "start": 904.27,
                    "end": 904.65,
                    "word": " a 2",
                    "probability": 0.99609375
                },
                {
                    "start": 904.65,
                    "end": 905.31,
                    "word": ",",
                    "probability": 0.57763671875
                }
            ]
        },
        {
            "id": 126,
            "seek": 90678,
            "start": 905.31,
            "end": 910.49,
            "text": " and then you have a second box, you want to put in the same input and get the same output.",
            "tokens": [
                51612,
                293,
                550,
                291,
                362,
                257,
                1150,
                2424,
                11,
                291,
                528,
                281,
                829,
                294,
                264,
                912,
                4846,
                293,
                483,
                264,
                912,
                5598,
                13,
                51860
            ],
            "temperature": 0.0,
            "avg_logprob": -0.1870312486588955,
            "compression_ratio": 1.7655502392344498,
            "no_speech_prob": 0.00011235475540161133,
            "words": [
                {
                    "start": 905.31,
                    "end": 905.69,
                    "word": " and",
                    "probability": 0.89013671875
                },
                {
                    "start": 905.69,
                    "end": 905.87,
                    "word": " then",
                    "probability": 0.72509765625
                },
                {
                    "start": 905.87,
                    "end": 906.03,
                    "word": " you",
                    "probability": 0.88916015625
                },
                {
                    "start": 906.03,
                    "end": 906.23,
                    "word": " have",
                    "probability": 0.876953125
                },
                {
                    "start": 906.39,
                    "end": 906.69,
                    "word": " a second",
                    "probability": 0.78955078125
                },
                {
                    "start": 906.69,
                    "end": 907.25,
                    "word": " box",
                    "probability": 0.8134765625
                },
                {
                    "start": 907.25,
                    "end": 907.47,
                    "word": ",",
                    "probability": 0.923828125
                },
                {
                    "start": 907.47,
                    "end": 907.63,
                    "word": " you",
                    "probability": 0.890625
                },
                {
                    "start": 907.63,
                    "end": 907.75,
                    "word": " want",
                    "probability": 0.7568359375
                },
                {
                    "start": 907.75,
                    "end": 907.87,
                    "word": " to",
                    "probability": 0.9033203125
                },
                {
                    "start": 907.87,
                    "end": 908.11,
                    "word": " put",
                    "probability": 0.89892578125
                },
                {
                    "start": 908.11,
                    "end": 908.27,
                    "word": " in",
                    "probability": 0.9033203125
                },
                {
                    "start": 908.27,
                    "end": 908.41,
                    "word": " the",
                    "probability": 0.8134765625
                },
                {
                    "start": 908.41,
                    "end": 908.65,
                    "word": " same",
                    "probability": 0.81201171875
                },
                {
                    "start": 908.65,
                    "end": 909.05,
                    "word": " input",
                    "probability": 0.8759765625
                },
                {
                    "start": 909.05,
                    "end": 909.41,
                    "word": " and",
                    "probability": 0.8837890625
                },
                {
                    "start": 909.41,
                    "end": 909.63,
                    "word": " get",
                    "probability": 0.88232421875
                },
                {
                    "start": 909.63,
                    "end": 909.83,
                    "word": " the",
                    "probability": 0.81103515625
                },
                {
                    "start": 909.83,
                    "end": 910.03,
                    "word": " same",
                    "probability": 0.806640625
                },
                {
                    "start": 910.03,
                    "end": 910.49,
                    "word": " output",
                    "probability": 0.82421875
                },
                {
                    "start": 910.49,
                    "end": 910.49,
                    "word": ".",
                    "probability": 0.9580078125
                }
            ]
        },
        {
            "id": 127,
            "seek": 93364,
            "start": 910.51,
            "end": 916.93,
            "text": " And it doesn't really matter what happens inside of that box as much as it matters that whatever",
            "tokens": [
                50412,
                400,
                309,
                1177,
                380,
                534,
                1871,
                437,
                2314,
                1854,
                295,
                300,
                2424,
                382,
                709,
                382,
                309,
                7001,
                300,
                2035,
                50692
            ],
            "temperature": 0.0,
            "avg_logprob": -0.19335937652398238,
            "compression_ratio": 1.6288209606986899,
            "no_speech_prob": 0.0009851455688476562,
            "words": [
                {
                    "start": 910.51,
                    "end": 911.17,
                    "word": " And",
                    "probability": 0.23095703125
                },
                {
                    "start": 911.17,
                    "end": 911.77,
                    "word": " it",
                    "probability": 0.8896484375
                },
                {
                    "start": 911.77,
                    "end": 912.01,
                    "word": " doesn't",
                    "probability": 0.932373046875
                },
                {
                    "start": 912.01,
                    "end": 912.19,
                    "word": " really",
                    "probability": 0.7509765625
                },
                {
                    "start": 912.19,
                    "end": 912.47,
                    "word": " matter",
                    "probability": 0.79248046875
                },
                {
                    "start": 912.47,
                    "end": 912.73,
                    "word": " what",
                    "probability": 0.88427734375
                },
                {
                    "start": 912.73,
                    "end": 913.35,
                    "word": " happens",
                    "probability": 0.9072265625
                },
                {
                    "start": 913.35,
                    "end": 913.95,
                    "word": " inside",
                    "probability": 0.86572265625
                },
                {
                    "start": 913.95,
                    "end": 914.13,
                    "word": " of",
                    "probability": 0.89697265625
                },
                {
                    "start": 914.13,
                    "end": 914.31,
                    "word": " that",
                    "probability": 0.87841796875
                },
                {
                    "start": 914.31,
                    "end": 914.79,
                    "word": " box",
                    "probability": 0.79736328125
                },
                {
                    "start": 914.79,
                    "end": 915.11,
                    "word": " as",
                    "probability": 0.5849609375
                },
                {
                    "start": 915.11,
                    "end": 915.35,
                    "word": " much",
                    "probability": 0.81689453125
                },
                {
                    "start": 915.35,
                    "end": 915.47,
                    "word": " as",
                    "probability": 0.89501953125
                },
                {
                    "start": 915.47,
                    "end": 915.63,
                    "word": " it",
                    "probability": 0.90185546875
                },
                {
                    "start": 915.63,
                    "end": 915.95,
                    "word": " matters",
                    "probability": 0.70849609375
                },
                {
                    "start": 915.95,
                    "end": 916.43,
                    "word": " that",
                    "probability": 0.8525390625
                },
                {
                    "start": 916.43,
                    "end": 916.93,
                    "word": " whatever",
                    "probability": 0.81201171875
                }
            ]
        },
        {
            "id": 128,
            "seek": 93364,
            "start": 916.93,
            "end": 922.75,
            "text": " input you give it, you get the same output. And that's what happens with model theft.",
            "tokens": [
                50692,
                4846,
                291,
                976,
                309,
                11,
                291,
                483,
                264,
                912,
                5598,
                13,
                400,
                300,
                311,
                437,
                2314,
                365,
                2316,
                28508,
                13,
                50952
            ],
            "temperature": 0.0,
            "avg_logprob": -0.19335937652398238,
            "compression_ratio": 1.6288209606986899,
            "no_speech_prob": 0.0009851455688476562,
            "words": [
                {
                    "start": 916.93,
                    "end": 917.43,
                    "word": " input",
                    "probability": 0.88330078125
                },
                {
                    "start": 917.43,
                    "end": 917.65,
                    "word": " you",
                    "probability": 0.89208984375
                },
                {
                    "start": 917.65,
                    "end": 917.89,
                    "word": " give",
                    "probability": 0.80419921875
                },
                {
                    "start": 917.89,
                    "end": 918.05,
                    "word": " it",
                    "probability": 0.89306640625
                },
                {
                    "start": 918.05,
                    "end": 918.37,
                    "word": ",",
                    "probability": 0.80419921875
                },
                {
                    "start": 918.37,
                    "end": 918.51,
                    "word": " you",
                    "probability": 0.89306640625
                },
                {
                    "start": 918.51,
                    "end": 918.67,
                    "word": " get",
                    "probability": 0.89794921875
                },
                {
                    "start": 918.67,
                    "end": 918.83,
                    "word": " the",
                    "probability": 0.80859375
                },
                {
                    "start": 918.83,
                    "end": 919.01,
                    "word": " same",
                    "probability": 0.80419921875
                },
                {
                    "start": 919.01,
                    "end": 919.51,
                    "word": " output",
                    "probability": 0.802734375
                },
                {
                    "start": 919.51,
                    "end": 920.11,
                    "word": ".",
                    "probability": 0.8623046875
                },
                {
                    "start": 920.29,
                    "end": 920.77,
                    "word": " And",
                    "probability": 0.85400390625
                },
                {
                    "start": 920.77,
                    "end": 920.95,
                    "word": " that's",
                    "probability": 0.938720703125
                },
                {
                    "start": 920.95,
                    "end": 921.05,
                    "word": " what",
                    "probability": 0.8837890625
                },
                {
                    "start": 921.05,
                    "end": 921.39,
                    "word": " happens",
                    "probability": 0.90625
                },
                {
                    "start": 921.39,
                    "end": 921.59,
                    "word": " with",
                    "probability": 0.82421875
                },
                {
                    "start": 921.59,
                    "end": 921.85,
                    "word": " model",
                    "probability": 0.76171875
                },
                {
                    "start": 921.85,
                    "end": 922.15,
                    "word": " theft",
                    "probability": 0.77392578125
                },
                {
                    "start": 922.15,
                    "end": 922.75,
                    "word": ".",
                    "probability": 0.98828125
                }
            ]
        },
        {
            "id": 129,
            "seek": 93364,
            "start": 922.87,
            "end": 930.51,
            "text": " So, our impact here is kind of twofold. And so, the potential for adversarial attacks in model",
            "tokens": [
                50980,
                407,
                11,
                527,
                2712,
                510,
                307,
                733,
                295,
                732,
                18353,
                13,
                400,
                370,
                11,
                264,
                3995,
                337,
                17641,
                44745,
                8122,
                294,
                2316,
                51364
            ],
            "temperature": 0.0,
            "avg_logprob": -0.19335937652398238,
            "compression_ratio": 1.6288209606986899,
            "no_speech_prob": 0.0009851455688476562,
            "words": [
                {
                    "start": 922.87,
                    "end": 923.15,
                    "word": " So",
                    "probability": 0.8662109375
                },
                {
                    "start": 923.15,
                    "end": 923.23,
                    "word": ",",
                    "probability": 0.1539306640625
                },
                {
                    "start": 923.23,
                    "end": 923.29,
                    "word": " our",
                    "probability": 0.80224609375
                },
                {
                    "start": 923.29,
                    "end": 923.81,
                    "word": " impact",
                    "probability": 0.8779296875
                },
                {
                    "start": 923.81,
                    "end": 924.21,
                    "word": " here",
                    "probability": 0.798828125
                },
                {
                    "start": 924.21,
                    "end": 924.93,
                    "word": " is",
                    "probability": 0.892578125
                },
                {
                    "start": 924.93,
                    "end": 925.53,
                    "word": " kind",
                    "probability": 0.65576171875
                },
                {
                    "start": 925.53,
                    "end": 925.65,
                    "word": " of",
                    "probability": 0.90380859375
                },
                {
                    "start": 925.65,
                    "end": 926.19,
                    "word": " twofold",
                    "probability": 0.717041015625
                },
                {
                    "start": 926.19,
                    "end": 926.79,
                    "word": ".",
                    "probability": 0.88916015625
                },
                {
                    "start": 926.83,
                    "end": 926.89,
                    "word": " And",
                    "probability": 0.80859375
                },
                {
                    "start": 926.89,
                    "end": 927.17,
                    "word": " so",
                    "probability": 0.8828125
                },
                {
                    "start": 927.17,
                    "end": 927.79,
                    "word": ",",
                    "probability": 0.84375
                },
                {
                    "start": 927.79,
                    "end": 928.17,
                    "word": " the",
                    "probability": 0.80517578125
                },
                {
                    "start": 928.17,
                    "end": 928.47,
                    "word": " potential",
                    "probability": 0.86767578125
                },
                {
                    "start": 928.47,
                    "end": 928.71,
                    "word": " for",
                    "probability": 0.8974609375
                },
                {
                    "start": 928.71,
                    "end": 929.33,
                    "word": " adversarial",
                    "probability": 0.9033203125
                },
                {
                    "start": 929.33,
                    "end": 929.97,
                    "word": " attacks",
                    "probability": 0.724609375
                },
                {
                    "start": 929.97,
                    "end": 930.25,
                    "word": " in",
                    "probability": 0.896484375
                },
                {
                    "start": 930.25,
                    "end": 930.51,
                    "word": " model",
                    "probability": 0.814453125
                }
            ]
        },
        {
            "id": 130,
            "seek": 93364,
            "start": 930.51,
            "end": 937.37,
            "text": " theft goes way up. Because somebody has a model they can specifically query against to validate",
            "tokens": [
                51364,
                28508,
                1709,
                636,
                493,
                13,
                1436,
                2618,
                575,
                257,
                2316,
                436,
                393,
                4682,
                14581,
                1970,
                281,
                29562,
                51712
            ],
            "temperature": 0.0,
            "avg_logprob": -0.19335937652398238,
            "compression_ratio": 1.6288209606986899,
            "no_speech_prob": 0.0009851455688476562,
            "words": [
                {
                    "start": 930.51,
                    "end": 930.83,
                    "word": " theft",
                    "probability": 0.77197265625
                },
                {
                    "start": 930.83,
                    "end": 931.13,
                    "word": " goes",
                    "probability": 0.8681640625
                },
                {
                    "start": 931.13,
                    "end": 931.57,
                    "word": " way",
                    "probability": 0.8994140625
                },
                {
                    "start": 931.57,
                    "end": 931.89,
                    "word": " up",
                    "probability": 0.9013671875
                },
                {
                    "start": 931.89,
                    "end": 932.05,
                    "word": ".",
                    "probability": 0.179443359375
                },
                {
                    "start": 932.05,
                    "end": 932.35,
                    "word": " Because",
                    "probability": 0.88134765625
                },
                {
                    "start": 932.35,
                    "end": 932.97,
                    "word": " somebody",
                    "probability": 0.77294921875
                },
                {
                    "start": 932.97,
                    "end": 933.41,
                    "word": " has",
                    "probability": 0.900390625
                },
                {
                    "start": 933.67,
                    "end": 933.97,
                    "word": " a model",
                    "probability": 0.8115234375
                },
                {
                    "start": 933.97,
                    "end": 934.19,
                    "word": " they",
                    "probability": 0.7275390625
                },
                {
                    "start": 934.19,
                    "end": 934.35,
                    "word": " can",
                    "probability": 0.90771484375
                },
                {
                    "start": 934.35,
                    "end": 934.77,
                    "word": " specifically",
                    "probability": 0.703125
                },
                {
                    "start": 934.77,
                    "end": 935.33,
                    "word": " query",
                    "probability": 0.88037109375
                },
                {
                    "start": 935.33,
                    "end": 935.93,
                    "word": " against",
                    "probability": 0.87646484375
                },
                {
                    "start": 935.93,
                    "end": 936.79,
                    "word": " to",
                    "probability": 0.876953125
                },
                {
                    "start": 936.79,
                    "end": 937.37,
                    "word": " validate",
                    "probability": 0.8662109375
                }
            ]
        },
        {
            "id": 131,
            "seek": 95994,
            "start": 937.37,
            "end": 943.89,
            "text": " whether or not it gets misclassified without interacting with your model. So, it gives them",
            "tokens": [
                50368,
                1968,
                420,
                406,
                309,
                2170,
                3346,
                11665,
                2587,
                1553,
                18017,
                365,
                428,
                2316,
                13,
                407,
                11,
                309,
                2709,
                552,
                50688
            ],
            "temperature": 0.0,
            "avg_logprob": -0.21573795216629305,
            "compression_ratio": 1.574468085106383,
            "no_speech_prob": 0.0311431884765625,
            "words": [
                {
                    "start": 937.37,
                    "end": 937.67,
                    "word": " whether",
                    "probability": 0.39306640625
                },
                {
                    "start": 937.67,
                    "end": 938.15,
                    "word": " or",
                    "probability": 0.89501953125
                },
                {
                    "start": 938.15,
                    "end": 938.55,
                    "word": " not",
                    "probability": 0.89501953125
                },
                {
                    "start": 938.55,
                    "end": 938.85,
                    "word": " it",
                    "probability": 0.90234375
                },
                {
                    "start": 938.85,
                    "end": 939.05,
                    "word": " gets",
                    "probability": 0.7939453125
                },
                {
                    "start": 939.05,
                    "end": 939.99,
                    "word": " misclassified",
                    "probability": 0.9532877604166666
                },
                {
                    "start": 939.99,
                    "end": 941.11,
                    "word": " without",
                    "probability": 0.6572265625
                },
                {
                    "start": 941.11,
                    "end": 941.71,
                    "word": " interacting",
                    "probability": 0.68359375
                },
                {
                    "start": 941.71,
                    "end": 941.95,
                    "word": " with",
                    "probability": 0.8125
                },
                {
                    "start": 941.95,
                    "end": 942.07,
                    "word": " your",
                    "probability": 0.81103515625
                },
                {
                    "start": 942.07,
                    "end": 942.41,
                    "word": " model",
                    "probability": 0.81201171875
                },
                {
                    "start": 942.41,
                    "end": 943.09,
                    "word": ".",
                    "probability": 0.873046875
                },
                {
                    "start": 943.39,
                    "end": 943.51,
                    "word": " So",
                    "probability": 0.80126953125
                },
                {
                    "start": 943.51,
                    "end": 943.55,
                    "word": ",",
                    "probability": 0.1773681640625
                },
                {
                    "start": 943.55,
                    "end": 943.59,
                    "word": " it",
                    "probability": 0.8828125
                },
                {
                    "start": 943.59,
                    "end": 943.75,
                    "word": " gives",
                    "probability": 0.84423828125
                },
                {
                    "start": 943.75,
                    "end": 943.89,
                    "word": " them",
                    "probability": 0.7451171875
                }
            ]
        },
        {
            "id": 132,
            "seek": 95994,
            "start": 943.89,
            "end": 948.91,
            "text": " sort of a private development environment for these adversarial attacks. And that's what Will",
            "tokens": [
                50688,
                1333,
                295,
                257,
                4551,
                3250,
                2823,
                337,
                613,
                17641,
                44745,
                8122,
                13,
                400,
                300,
                311,
                437,
                3099,
                50940
            ],
            "temperature": 0.0,
            "avg_logprob": -0.21573795216629305,
            "compression_ratio": 1.574468085106383,
            "no_speech_prob": 0.0311431884765625,
            "words": [
                {
                    "start": 943.89,
                    "end": 944.09,
                    "word": " sort",
                    "probability": 0.72412109375
                },
                {
                    "start": 944.09,
                    "end": 944.17,
                    "word": " of",
                    "probability": 0.9033203125
                },
                {
                    "start": 944.39,
                    "end": 944.77,
                    "word": " a private",
                    "probability": 0.8134765625
                },
                {
                    "start": 944.77,
                    "end": 945.31,
                    "word": " development",
                    "probability": 0.77392578125
                },
                {
                    "start": 945.31,
                    "end": 945.97,
                    "word": " environment",
                    "probability": 0.7568359375
                },
                {
                    "start": 945.97,
                    "end": 946.47,
                    "word": " for",
                    "probability": 0.890625
                },
                {
                    "start": 946.47,
                    "end": 946.67,
                    "word": " these",
                    "probability": 0.7001953125
                },
                {
                    "start": 946.67,
                    "end": 947.27,
                    "word": " adversarial",
                    "probability": 0.92041015625
                },
                {
                    "start": 947.27,
                    "end": 947.73,
                    "word": " attacks",
                    "probability": 0.7158203125
                },
                {
                    "start": 947.73,
                    "end": 948.21,
                    "word": ".",
                    "probability": 0.77685546875
                },
                {
                    "start": 948.21,
                    "end": 948.27,
                    "word": " And",
                    "probability": 0.83056640625
                },
                {
                    "start": 948.27,
                    "end": 948.49,
                    "word": " that's",
                    "probability": 0.929443359375
                },
                {
                    "start": 948.49,
                    "end": 948.71,
                    "word": " what",
                    "probability": 0.89501953125
                },
                {
                    "start": 948.71,
                    "end": 948.91,
                    "word": " Will",
                    "probability": 0.8447265625
                }
            ]
        },
        {
            "id": 133,
            "seek": 95994,
            "start": 948.91,
            "end": 957.03,
            "text": " did in the proof putting. Is he created this copycat model so that he could test phishing",
            "tokens": [
                50940,
                630,
                294,
                264,
                8177,
                3372,
                13,
                1119,
                415,
                2942,
                341,
                5055,
                18035,
                2316,
                370,
                300,
                415,
                727,
                1500,
                903,
                3807,
                51348
            ],
            "temperature": 0.0,
            "avg_logprob": -0.21573795216629305,
            "compression_ratio": 1.574468085106383,
            "no_speech_prob": 0.0311431884765625,
            "words": [
                {
                    "start": 948.91,
                    "end": 949.25,
                    "word": " did",
                    "probability": 0.8828125
                },
                {
                    "start": 949.25,
                    "end": 949.59,
                    "word": " in",
                    "probability": 0.8955078125
                },
                {
                    "start": 949.59,
                    "end": 949.79,
                    "word": " the",
                    "probability": 0.79638671875
                },
                {
                    "start": 949.79,
                    "end": 950.13,
                    "word": " proof",
                    "probability": 0.82470703125
                },
                {
                    "start": 950.13,
                    "end": 950.45,
                    "word": " putting",
                    "probability": 0.1826171875
                },
                {
                    "start": 950.45,
                    "end": 951.13,
                    "word": ".",
                    "probability": 0.2479248046875
                },
                {
                    "start": 951.33,
                    "end": 952.01,
                    "word": " Is",
                    "probability": 0.591796875
                },
                {
                    "start": 952.01,
                    "end": 952.39,
                    "word": " he",
                    "probability": 0.873046875
                },
                {
                    "start": 952.39,
                    "end": 953.79,
                    "word": " created",
                    "probability": 0.73876953125
                },
                {
                    "start": 953.79,
                    "end": 954.09,
                    "word": " this",
                    "probability": 0.8876953125
                },
                {
                    "start": 954.09,
                    "end": 954.63,
                    "word": " copycat",
                    "probability": 0.841796875
                },
                {
                    "start": 954.63,
                    "end": 955.05,
                    "word": " model",
                    "probability": 0.81640625
                },
                {
                    "start": 955.05,
                    "end": 955.27,
                    "word": " so",
                    "probability": 0.853515625
                },
                {
                    "start": 955.27,
                    "end": 955.41,
                    "word": " that",
                    "probability": 0.79345703125
                },
                {
                    "start": 955.41,
                    "end": 955.55,
                    "word": " he",
                    "probability": 0.908203125
                },
                {
                    "start": 955.55,
                    "end": 955.75,
                    "word": " could",
                    "probability": 0.76953125
                },
                {
                    "start": 955.75,
                    "end": 956.45,
                    "word": " test",
                    "probability": 0.81787109375
                },
                {
                    "start": 956.45,
                    "end": 957.03,
                    "word": " phishing",
                    "probability": 0.901123046875
                }
            ]
        },
        {
            "id": 134,
            "seek": 95994,
            "start": 957.03,
            "end": 963.67,
            "text": " emails to bypass the proof point email security appliance. And it was incredibly clever. Super",
            "tokens": [
                51348,
                12524,
                281,
                24996,
                264,
                8177,
                935,
                3796,
                3825,
                45646,
                13,
                400,
                309,
                390,
                6252,
                13494,
                13,
                4548,
                51680
            ],
            "temperature": 0.0,
            "avg_logprob": -0.21573795216629305,
            "compression_ratio": 1.574468085106383,
            "no_speech_prob": 0.0311431884765625,
            "words": [
                {
                    "start": 957.03,
                    "end": 957.51,
                    "word": " emails",
                    "probability": 0.75244140625
                },
                {
                    "start": 957.51,
                    "end": 957.73,
                    "word": " to",
                    "probability": 0.89892578125
                },
                {
                    "start": 957.73,
                    "end": 958.05,
                    "word": " bypass",
                    "probability": 0.83935546875
                },
                {
                    "start": 958.05,
                    "end": 958.57,
                    "word": " the",
                    "probability": 0.8134765625
                },
                {
                    "start": 958.57,
                    "end": 958.93,
                    "word": " proof",
                    "probability": 0.66943359375
                },
                {
                    "start": 958.93,
                    "end": 959.23,
                    "word": " point",
                    "probability": 0.60888671875
                },
                {
                    "start": 959.23,
                    "end": 959.55,
                    "word": " email",
                    "probability": 0.90234375
                },
                {
                    "start": 959.55,
                    "end": 959.95,
                    "word": " security",
                    "probability": 0.90087890625
                },
                {
                    "start": 959.95,
                    "end": 960.45,
                    "word": " appliance",
                    "probability": 0.60888671875
                },
                {
                    "start": 960.45,
                    "end": 961.09,
                    "word": ".",
                    "probability": 0.8154296875
                },
                {
                    "start": 961.09,
                    "end": 961.45,
                    "word": " And",
                    "probability": 0.81494140625
                },
                {
                    "start": 961.45,
                    "end": 961.63,
                    "word": " it",
                    "probability": 0.88330078125
                },
                {
                    "start": 961.63,
                    "end": 961.79,
                    "word": " was",
                    "probability": 0.89501953125
                },
                {
                    "start": 961.79,
                    "end": 962.65,
                    "word": " incredibly",
                    "probability": 0.71875
                },
                {
                    "start": 962.65,
                    "end": 963.17,
                    "word": " clever",
                    "probability": 0.8662109375
                },
                {
                    "start": 963.17,
                    "end": 963.41,
                    "word": ".",
                    "probability": 0.334716796875
                },
                {
                    "start": 963.41,
                    "end": 963.67,
                    "word": " Super",
                    "probability": 0.73779296875
                }
            ]
        },
        {
            "id": 135,
            "seek": 98634,
            "start": 963.67,
            "end": 971.05,
            "text": " dope. I recommend watching his talk. But after you watch all the other talks at the AI Village.",
            "tokens": [
                50364,
                23383,
                13,
                286,
                2748,
                1976,
                702,
                751,
                13,
                583,
                934,
                291,
                1159,
                439,
                264,
                661,
                6686,
                412,
                264,
                7318,
                22651,
                13,
                50708
            ],
            "temperature": 0.0,
            "avg_logprob": -0.22910391063575286,
            "compression_ratio": 1.4741784037558685,
            "no_speech_prob": 0.003482818603515625,
            "words": [
                {
                    "start": 963.67,
                    "end": 964.09,
                    "word": " dope",
                    "probability": 0.059967041015625
                },
                {
                    "start": 964.09,
                    "end": 964.31,
                    "word": ".",
                    "probability": 0.43310546875
                },
                {
                    "start": 964.31,
                    "end": 964.37,
                    "word": " I",
                    "probability": 0.97998046875
                },
                {
                    "start": 964.37,
                    "end": 964.87,
                    "word": " recommend",
                    "probability": 0.68896484375
                },
                {
                    "start": 964.87,
                    "end": 965.69,
                    "word": " watching",
                    "probability": 0.69873046875
                },
                {
                    "start": 965.69,
                    "end": 965.95,
                    "word": " his",
                    "probability": 0.88330078125
                },
                {
                    "start": 965.95,
                    "end": 966.39,
                    "word": " talk",
                    "probability": 0.8095703125
                },
                {
                    "start": 966.39,
                    "end": 966.91,
                    "word": ".",
                    "probability": 0.5751953125
                },
                {
                    "start": 967.37,
                    "end": 967.49,
                    "word": " But",
                    "probability": 0.828125
                },
                {
                    "start": 967.49,
                    "end": 968.13,
                    "word": " after",
                    "probability": 0.73974609375
                },
                {
                    "start": 968.13,
                    "end": 968.45,
                    "word": " you",
                    "probability": 0.81201171875
                },
                {
                    "start": 968.45,
                    "end": 968.71,
                    "word": " watch",
                    "probability": 0.81494140625
                },
                {
                    "start": 968.71,
                    "end": 968.89,
                    "word": " all",
                    "probability": 0.884765625
                },
                {
                    "start": 968.89,
                    "end": 968.99,
                    "word": " the",
                    "probability": 0.763671875
                },
                {
                    "start": 968.99,
                    "end": 969.19,
                    "word": " other",
                    "probability": 0.77978515625
                },
                {
                    "start": 969.19,
                    "end": 969.51,
                    "word": " talks",
                    "probability": 0.71533203125
                },
                {
                    "start": 969.51,
                    "end": 969.67,
                    "word": " at",
                    "probability": 0.865234375
                },
                {
                    "start": 969.67,
                    "end": 969.79,
                    "word": " the",
                    "probability": 0.80126953125
                },
                {
                    "start": 969.79,
                    "end": 970.07,
                    "word": " AI",
                    "probability": 0.89208984375
                },
                {
                    "start": 970.07,
                    "end": 970.53,
                    "word": " Village",
                    "probability": 0.417236328125
                },
                {
                    "start": 970.53,
                    "end": 971.05,
                    "word": ".",
                    "probability": 0.498046875
                }
            ]
        },
        {
            "id": 136,
            "seek": 98634,
            "start": 971.29,
            "end": 975.77,
            "text": " Maybe, like, after DEF CON, watch it. It's a really good talk.",
            "tokens": [
                50756,
                2704,
                11,
                411,
                11,
                934,
                10113,
                37,
                16596,
                11,
                1159,
                309,
                13,
                467,
                311,
                257,
                534,
                665,
                751,
                13,
                50944
            ],
            "temperature": 0.0,
            "avg_logprob": -0.22910391063575286,
            "compression_ratio": 1.4741784037558685,
            "no_speech_prob": 0.003482818603515625,
            "words": [
                {
                    "start": 971.29,
                    "end": 971.81,
                    "word": " Maybe",
                    "probability": 0.9013671875
                },
                {
                    "start": 971.81,
                    "end": 971.87,
                    "word": ",",
                    "probability": 0.2362060546875
                },
                {
                    "start": 971.87,
                    "end": 972.11,
                    "word": " like",
                    "probability": 0.87353515625
                },
                {
                    "start": 972.11,
                    "end": 972.29,
                    "word": ",",
                    "probability": 0.98095703125
                },
                {
                    "start": 972.29,
                    "end": 972.65,
                    "word": " after",
                    "probability": 0.71826171875
                },
                {
                    "start": 972.65,
                    "end": 972.93,
                    "word": " DEF",
                    "probability": 0.6982421875
                },
                {
                    "start": 972.93,
                    "end": 973.17,
                    "word": " CON",
                    "probability": 0.83984375
                },
                {
                    "start": 973.17,
                    "end": 973.45,
                    "word": ",",
                    "probability": 0.61181640625
                },
                {
                    "start": 973.45,
                    "end": 973.71,
                    "word": " watch",
                    "probability": 0.79345703125
                },
                {
                    "start": 973.71,
                    "end": 973.87,
                    "word": " it",
                    "probability": 0.9013671875
                },
                {
                    "start": 973.87,
                    "end": 974.17,
                    "word": ".",
                    "probability": 0.99072265625
                },
                {
                    "start": 974.17,
                    "end": 974.61,
                    "word": " It's",
                    "probability": 0.93310546875
                },
                {
                    "start": 974.69,
                    "end": 974.83,
                    "word": " a really",
                    "probability": 0.771484375
                },
                {
                    "start": 974.83,
                    "end": 975.03,
                    "word": " good",
                    "probability": 0.8759765625
                },
                {
                    "start": 975.03,
                    "end": 975.25,
                    "word": " talk",
                    "probability": 0.794921875
                },
                {
                    "start": 975.25,
                    "end": 975.77,
                    "word": ".",
                    "probability": 0.99609375
                }
            ]
        },
        {
            "id": 137,
            "seek": 98634,
            "start": 977.17,
            "end": 981.71,
            "text": " The other impact, though, is the loss of intellectual property.",
            "tokens": [
                51052,
                440,
                661,
                2712,
                11,
                1673,
                11,
                307,
                264,
                4470,
                295,
                12576,
                4707,
                13,
                51244
            ],
            "temperature": 0.0,
            "avg_logprob": -0.22910391063575286,
            "compression_ratio": 1.4741784037558685,
            "no_speech_prob": 0.003482818603515625,
            "words": [
                {
                    "start": 977.17,
                    "end": 977.69,
                    "word": " The",
                    "probability": 0.8173828125
                },
                {
                    "start": 977.69,
                    "end": 977.93,
                    "word": " other",
                    "probability": 0.77880859375
                },
                {
                    "start": 977.93,
                    "end": 978.49,
                    "word": " impact",
                    "probability": 0.85693359375
                },
                {
                    "start": 978.49,
                    "end": 978.57,
                    "word": ",",
                    "probability": 0.9072265625
                },
                {
                    "start": 978.57,
                    "end": 978.79,
                    "word": " though",
                    "probability": 0.7666015625
                },
                {
                    "start": 978.79,
                    "end": 978.93,
                    "word": ",",
                    "probability": 0.986328125
                },
                {
                    "start": 978.93,
                    "end": 979.51,
                    "word": " is",
                    "probability": 0.90478515625
                },
                {
                    "start": 979.51,
                    "end": 979.67,
                    "word": " the",
                    "probability": 0.81103515625
                },
                {
                    "start": 979.67,
                    "end": 979.93,
                    "word": " loss",
                    "probability": 0.7490234375
                },
                {
                    "start": 979.93,
                    "end": 980.07,
                    "word": " of",
                    "probability": 0.90625
                },
                {
                    "start": 980.07,
                    "end": 980.55,
                    "word": " intellectual",
                    "probability": 0.892578125
                },
                {
                    "start": 980.55,
                    "end": 981.19,
                    "word": " property",
                    "probability": 0.810546875
                },
                {
                    "start": 981.19,
                    "end": 981.71,
                    "word": ".",
                    "probability": 0.99462890625
                }
            ]
        },
        {
            "id": 138,
            "seek": 98634,
            "start": 981.95,
            "end": 990.07,
            "text": " So, one of the things that I think kind of got overlooked in how cool the ability to bypass",
            "tokens": [
                51272,
                407,
                11,
                472,
                295,
                264,
                721,
                300,
                286,
                519,
                733,
                295,
                658,
                32269,
                294,
                577,
                1627,
                264,
                3485,
                281,
                24996,
                51708
            ],
            "temperature": 0.0,
            "avg_logprob": -0.22910391063575286,
            "compression_ratio": 1.4741784037558685,
            "no_speech_prob": 0.003482818603515625,
            "words": [
                {
                    "start": 981.95,
                    "end": 982.21,
                    "word": " So",
                    "probability": 0.869140625
                },
                {
                    "start": 982.21,
                    "end": 982.79,
                    "word": ",",
                    "probability": 0.499755859375
                },
                {
                    "start": 982.79,
                    "end": 983.05,
                    "word": " one",
                    "probability": 0.82177734375
                },
                {
                    "start": 983.05,
                    "end": 983.17,
                    "word": " of",
                    "probability": 0.90673828125
                },
                {
                    "start": 983.17,
                    "end": 983.29,
                    "word": " the",
                    "probability": 0.8173828125
                },
                {
                    "start": 983.29,
                    "end": 983.53,
                    "word": " things",
                    "probability": 0.73828125
                },
                {
                    "start": 983.53,
                    "end": 983.69,
                    "word": " that",
                    "probability": 0.8623046875
                },
                {
                    "start": 983.69,
                    "end": 983.83,
                    "word": " I",
                    "probability": 0.9814453125
                },
                {
                    "start": 983.83,
                    "end": 983.99,
                    "word": " think",
                    "probability": 0.8115234375
                },
                {
                    "start": 983.99,
                    "end": 984.23,
                    "word": " kind",
                    "probability": 0.7578125
                },
                {
                    "start": 984.23,
                    "end": 984.67,
                    "word": " of",
                    "probability": 0.90673828125
                },
                {
                    "start": 984.67,
                    "end": 985.17,
                    "word": " got",
                    "probability": 0.8818359375
                },
                {
                    "start": 985.17,
                    "end": 985.87,
                    "word": " overlooked",
                    "probability": 0.71923828125
                },
                {
                    "start": 985.87,
                    "end": 986.29,
                    "word": " in",
                    "probability": 0.896484375
                },
                {
                    "start": 986.29,
                    "end": 986.59,
                    "word": " how",
                    "probability": 0.89990234375
                },
                {
                    "start": 986.59,
                    "end": 987.33,
                    "word": " cool",
                    "probability": 0.88330078125
                },
                {
                    "start": 987.33,
                    "end": 987.87,
                    "word": " the",
                    "probability": 0.81787109375
                },
                {
                    "start": 987.87,
                    "end": 989.47,
                    "word": " ability",
                    "probability": 0.86083984375
                },
                {
                    "start": 989.47,
                    "end": 989.73,
                    "word": " to",
                    "probability": 0.90625
                },
                {
                    "start": 989.73,
                    "end": 990.07,
                    "word": " bypass",
                    "probability": 0.875
                }
            ]
        },
        {
            "id": 139,
            "seek": 101018,
            "start": 990.07,
            "end": 999.33,
            "text": " security appliances is, is that if you hire a bunch of data scientists and collect a bunch of",
            "tokens": [
                50388,
                3825,
                35480,
                307,
                11,
                307,
                300,
                498,
                291,
                11158,
                257,
                3840,
                295,
                1412,
                7708,
                293,
                2500,
                257,
                3840,
                295,
                50824
            ],
            "temperature": 0.0,
            "avg_logprob": -0.2077636714093387,
            "compression_ratio": 1.6987951807228916,
            "no_speech_prob": 0.00453948974609375,
            "words": [
                {
                    "start": 990.07,
                    "end": 991.01,
                    "word": " security",
                    "probability": 0.556640625
                },
                {
                    "start": 991.01,
                    "end": 991.65,
                    "word": " appliances",
                    "probability": 0.72509765625
                },
                {
                    "start": 991.65,
                    "end": 993.59,
                    "word": " is",
                    "probability": 0.8486328125
                },
                {
                    "start": 993.59,
                    "end": 993.59,
                    "word": ",",
                    "probability": 0.55615234375
                },
                {
                    "start": 993.59,
                    "end": 994.71,
                    "word": " is",
                    "probability": 0.88916015625
                },
                {
                    "start": 994.71,
                    "end": 995.01,
                    "word": " that",
                    "probability": 0.8740234375
                },
                {
                    "start": 995.01,
                    "end": 995.59,
                    "word": " if",
                    "probability": 0.8701171875
                },
                {
                    "start": 995.59,
                    "end": 995.85,
                    "word": " you",
                    "probability": 0.896484375
                },
                {
                    "start": 995.85,
                    "end": 996.49,
                    "word": " hire",
                    "probability": 0.8837890625
                },
                {
                    "start": 996.75,
                    "end": 996.89,
                    "word": " a bunch",
                    "probability": 0.8095703125
                },
                {
                    "start": 996.89,
                    "end": 997.05,
                    "word": " of",
                    "probability": 0.90869140625
                },
                {
                    "start": 997.05,
                    "end": 997.29,
                    "word": " data",
                    "probability": 0.83544921875
                },
                {
                    "start": 997.29,
                    "end": 997.91,
                    "word": " scientists",
                    "probability": 0.71923828125
                },
                {
                    "start": 997.91,
                    "end": 998.37,
                    "word": " and",
                    "probability": 0.8095703125
                },
                {
                    "start": 998.37,
                    "end": 998.83,
                    "word": " collect",
                    "probability": 0.7626953125
                },
                {
                    "start": 999.03,
                    "end": 999.17,
                    "word": " a bunch",
                    "probability": 0.80712890625
                },
                {
                    "start": 999.17,
                    "end": 999.33,
                    "word": " of",
                    "probability": 0.90625
                }
            ]
        },
        {
            "id": 140,
            "seek": 101018,
            "start": 999.33,
            "end": 1005.83,
            "text": " data and clean a bunch of data and spend years creating and building and deploying a model that",
            "tokens": [
                50824,
                1412,
                293,
                2541,
                257,
                3840,
                295,
                1412,
                293,
                3496,
                924,
                4084,
                293,
                2390,
                293,
                34198,
                257,
                2316,
                300,
                51148
            ],
            "temperature": 0.0,
            "avg_logprob": -0.2077636714093387,
            "compression_ratio": 1.6987951807228916,
            "no_speech_prob": 0.00453948974609375,
            "words": [
                {
                    "start": 999.33,
                    "end": 999.65,
                    "word": " data",
                    "probability": 0.810546875
                },
                {
                    "start": 999.65,
                    "end": 999.91,
                    "word": " and",
                    "probability": 0.82373046875
                },
                {
                    "start": 999.91,
                    "end": 1000.19,
                    "word": " clean",
                    "probability": 0.8408203125
                },
                {
                    "start": 1000.33,
                    "end": 1000.49,
                    "word": " a bunch",
                    "probability": 0.80908203125
                },
                {
                    "start": 1000.49,
                    "end": 1000.65,
                    "word": " of",
                    "probability": 0.90673828125
                },
                {
                    "start": 1000.65,
                    "end": 1000.97,
                    "word": " data",
                    "probability": 0.8076171875
                },
                {
                    "start": 1000.97,
                    "end": 1001.59,
                    "word": " and",
                    "probability": 0.72412109375
                },
                {
                    "start": 1001.59,
                    "end": 1001.85,
                    "word": " spend",
                    "probability": 0.83349609375
                },
                {
                    "start": 1001.85,
                    "end": 1002.47,
                    "word": " years",
                    "probability": 0.8916015625
                },
                {
                    "start": 1002.47,
                    "end": 1003.27,
                    "word": " creating",
                    "probability": 0.78759765625
                },
                {
                    "start": 1003.27,
                    "end": 1003.61,
                    "word": " and",
                    "probability": 0.8759765625
                },
                {
                    "start": 1003.61,
                    "end": 1003.97,
                    "word": " building",
                    "probability": 0.78076171875
                },
                {
                    "start": 1003.97,
                    "end": 1004.25,
                    "word": " and",
                    "probability": 0.890625
                },
                {
                    "start": 1004.25,
                    "end": 1004.77,
                    "word": " deploying",
                    "probability": 0.66259765625
                },
                {
                    "start": 1005.11,
                    "end": 1005.41,
                    "word": " a model",
                    "probability": 0.80712890625
                },
                {
                    "start": 1005.41,
                    "end": 1005.83,
                    "word": " that",
                    "probability": 0.8759765625
                }
            ]
        },
        {
            "id": 141,
            "seek": 101018,
            "start": 1005.83,
            "end": 1013.91,
            "text": " really differentiates you from the competition, and some unscrupulous company in not Canada,",
            "tokens": [
                51176,
                534,
                27372,
                1024,
                291,
                490,
                264,
                6211,
                11,
                293,
                512,
                2693,
                66,
                11976,
                6893,
                2237,
                294,
                406,
                6309,
                11,
                51524
            ],
            "temperature": 0.0,
            "avg_logprob": -0.2077636714093387,
            "compression_ratio": 1.6987951807228916,
            "no_speech_prob": 0.00453948974609375,
            "words": [
                {
                    "start": 1005.83,
                    "end": 1006.69,
                    "word": " really",
                    "probability": 0.7685546875
                },
                {
                    "start": 1006.69,
                    "end": 1007.73,
                    "word": " differentiates",
                    "probability": 0.827392578125
                },
                {
                    "start": 1007.73,
                    "end": 1007.89,
                    "word": " you",
                    "probability": 0.8935546875
                },
                {
                    "start": 1007.89,
                    "end": 1008.09,
                    "word": " from",
                    "probability": 0.822265625
                },
                {
                    "start": 1008.09,
                    "end": 1008.27,
                    "word": " the",
                    "probability": 0.818359375
                },
                {
                    "start": 1008.27,
                    "end": 1008.77,
                    "word": " competition",
                    "probability": 0.89697265625
                },
                {
                    "start": 1008.77,
                    "end": 1009.23,
                    "word": ",",
                    "probability": 0.3408203125
                },
                {
                    "start": 1009.23,
                    "end": 1009.73,
                    "word": " and",
                    "probability": 0.89306640625
                },
                {
                    "start": 1009.73,
                    "end": 1010.15,
                    "word": " some",
                    "probability": 0.814453125
                },
                {
                    "start": 1010.15,
                    "end": 1011.05,
                    "word": " unscrupulous",
                    "probability": 0.8917236328125
                },
                {
                    "start": 1011.05,
                    "end": 1011.63,
                    "word": " company",
                    "probability": 0.81005859375
                },
                {
                    "start": 1011.63,
                    "end": 1012.19,
                    "word": " in",
                    "probability": 0.89306640625
                },
                {
                    "start": 1012.19,
                    "end": 1012.73,
                    "word": " not",
                    "probability": 0.7822265625
                },
                {
                    "start": 1012.73,
                    "end": 1013.31,
                    "word": " Canada",
                    "probability": 0.8076171875
                },
                {
                    "start": 1013.31,
                    "end": 1013.91,
                    "word": ",",
                    "probability": 0.7783203125
                }
            ]
        },
        {
            "id": 142,
            "seek": 103904,
            "start": 1014.99,
            "end": 1022.77,
            "text": " you know, steals your model, approximates your model, and just shoves it into their product",
            "tokens": [
                50364,
                291,
                458,
                11,
                46962,
                428,
                2316,
                11,
                8542,
                1024,
                428,
                2316,
                11,
                293,
                445,
                2223,
                977,
                309,
                666,
                641,
                1674,
                50828
            ],
            "temperature": 0.0,
            "avg_logprob": -0.21407860824742267,
            "compression_ratio": 1.727699530516432,
            "no_speech_prob": 0.0015974044799804688,
            "words": [
                {
                    "start": 1014.99,
                    "end": 1015.51,
                    "word": " you",
                    "probability": 0.205810546875
                },
                {
                    "start": 1015.51,
                    "end": 1016.03,
                    "word": " know",
                    "probability": 0.82275390625
                },
                {
                    "start": 1016.03,
                    "end": 1016.53,
                    "word": ",",
                    "probability": 0.837890625
                },
                {
                    "start": 1016.53,
                    "end": 1017.05,
                    "word": " steals",
                    "probability": 0.86962890625
                },
                {
                    "start": 1017.05,
                    "end": 1017.31,
                    "word": " your",
                    "probability": 0.80810546875
                },
                {
                    "start": 1017.31,
                    "end": 1017.69,
                    "word": " model",
                    "probability": 0.79931640625
                },
                {
                    "start": 1017.69,
                    "end": 1018.15,
                    "word": ",",
                    "probability": 0.87255859375
                },
                {
                    "start": 1018.15,
                    "end": 1019.01,
                    "word": " approximates",
                    "probability": 0.932373046875
                },
                {
                    "start": 1019.01,
                    "end": 1019.23,
                    "word": " your",
                    "probability": 0.80615234375
                },
                {
                    "start": 1019.23,
                    "end": 1019.61,
                    "word": " model",
                    "probability": 0.80712890625
                },
                {
                    "start": 1019.61,
                    "end": 1020.19,
                    "word": ",",
                    "probability": 0.6533203125
                },
                {
                    "start": 1020.19,
                    "end": 1020.55,
                    "word": " and",
                    "probability": 0.896484375
                },
                {
                    "start": 1020.55,
                    "end": 1020.85,
                    "word": " just",
                    "probability": 0.8876953125
                },
                {
                    "start": 1020.85,
                    "end": 1021.47,
                    "word": " shoves",
                    "probability": 0.92626953125
                },
                {
                    "start": 1021.47,
                    "end": 1021.63,
                    "word": " it",
                    "probability": 0.9072265625
                },
                {
                    "start": 1021.63,
                    "end": 1021.85,
                    "word": " into",
                    "probability": 0.7353515625
                },
                {
                    "start": 1021.85,
                    "end": 1022.09,
                    "word": " their",
                    "probability": 0.908203125
                },
                {
                    "start": 1022.09,
                    "end": 1022.77,
                    "word": " product",
                    "probability": 0.78466796875
                }
            ]
        },
        {
            "id": 143,
            "seek": 103904,
            "start": 1022.77,
            "end": 1029.51,
            "text": " and says, yeah, we do the same thing. Well, it costs them a lot less money to just copy",
            "tokens": [
                50828,
                293,
                1619,
                11,
                1338,
                11,
                321,
                360,
                264,
                912,
                551,
                13,
                1042,
                11,
                309,
                5497,
                552,
                257,
                688,
                1570,
                1460,
                281,
                445,
                5055,
                51144
            ],
            "temperature": 0.0,
            "avg_logprob": -0.21407860824742267,
            "compression_ratio": 1.727699530516432,
            "no_speech_prob": 0.0015974044799804688,
            "words": [
                {
                    "start": 1022.77,
                    "end": 1023.61,
                    "word": " and",
                    "probability": 0.59375
                },
                {
                    "start": 1023.61,
                    "end": 1024.21,
                    "word": " says",
                    "probability": 0.80126953125
                },
                {
                    "start": 1024.21,
                    "end": 1024.39,
                    "word": ",",
                    "probability": 0.76611328125
                },
                {
                    "start": 1024.39,
                    "end": 1024.51,
                    "word": " yeah",
                    "probability": 0.435546875
                },
                {
                    "start": 1024.51,
                    "end": 1024.55,
                    "word": ",",
                    "probability": 0.90576171875
                },
                {
                    "start": 1024.55,
                    "end": 1024.77,
                    "word": " we",
                    "probability": 0.90234375
                },
                {
                    "start": 1024.77,
                    "end": 1025.07,
                    "word": " do",
                    "probability": 0.8798828125
                },
                {
                    "start": 1025.07,
                    "end": 1025.25,
                    "word": " the",
                    "probability": 0.80908203125
                },
                {
                    "start": 1025.25,
                    "end": 1025.45,
                    "word": " same",
                    "probability": 0.81298828125
                },
                {
                    "start": 1025.45,
                    "end": 1025.81,
                    "word": " thing",
                    "probability": 0.81591796875
                },
                {
                    "start": 1025.81,
                    "end": 1026.33,
                    "word": ".",
                    "probability": 0.95947265625
                },
                {
                    "start": 1026.83,
                    "end": 1027.35,
                    "word": " Well",
                    "probability": 0.73828125
                },
                {
                    "start": 1027.35,
                    "end": 1027.37,
                    "word": ",",
                    "probability": 0.91796875
                },
                {
                    "start": 1027.37,
                    "end": 1027.49,
                    "word": " it",
                    "probability": 0.8974609375
                },
                {
                    "start": 1027.49,
                    "end": 1027.77,
                    "word": " costs",
                    "probability": 0.7021484375
                },
                {
                    "start": 1027.77,
                    "end": 1028.01,
                    "word": " them",
                    "probability": 0.73095703125
                },
                {
                    "start": 1028.13,
                    "end": 1028.27,
                    "word": " a lot",
                    "probability": 0.89013671875
                },
                {
                    "start": 1028.27,
                    "end": 1028.53,
                    "word": " less",
                    "probability": 0.89404296875
                },
                {
                    "start": 1028.53,
                    "end": 1028.79,
                    "word": " money",
                    "probability": 0.8349609375
                },
                {
                    "start": 1028.79,
                    "end": 1028.97,
                    "word": " to",
                    "probability": 0.90087890625
                },
                {
                    "start": 1028.97,
                    "end": 1029.19,
                    "word": " just",
                    "probability": 0.89208984375
                },
                {
                    "start": 1029.19,
                    "end": 1029.51,
                    "word": " copy",
                    "probability": 0.81689453125
                }
            ]
        },
        {
            "id": 144,
            "seek": 103904,
            "start": 1029.51,
            "end": 1035.11,
            "text": " your model. It costs them almost nothing in the grand scheme of things to copy your model.",
            "tokens": [
                51144,
                428,
                2316,
                13,
                467,
                5497,
                552,
                1920,
                1825,
                294,
                264,
                2697,
                12232,
                295,
                721,
                281,
                5055,
                428,
                2316,
                13,
                51420
            ],
            "temperature": 0.0,
            "avg_logprob": -0.21407860824742267,
            "compression_ratio": 1.727699530516432,
            "no_speech_prob": 0.0015974044799804688,
            "words": [
                {
                    "start": 1029.51,
                    "end": 1029.73,
                    "word": " your",
                    "probability": 0.810546875
                },
                {
                    "start": 1029.73,
                    "end": 1030.07,
                    "word": " model",
                    "probability": 0.80712890625
                },
                {
                    "start": 1030.07,
                    "end": 1030.59,
                    "word": ".",
                    "probability": 0.79541015625
                },
                {
                    "start": 1030.67,
                    "end": 1030.67,
                    "word": " It",
                    "probability": 0.70556640625
                },
                {
                    "start": 1030.67,
                    "end": 1030.81,
                    "word": " costs",
                    "probability": 0.744140625
                },
                {
                    "start": 1030.81,
                    "end": 1031.05,
                    "word": " them",
                    "probability": 0.72412109375
                },
                {
                    "start": 1031.05,
                    "end": 1031.41,
                    "word": " almost",
                    "probability": 0.6669921875
                },
                {
                    "start": 1031.41,
                    "end": 1031.79,
                    "word": " nothing",
                    "probability": 0.8564453125
                },
                {
                    "start": 1031.79,
                    "end": 1032.81,
                    "word": " in",
                    "probability": 0.77197265625
                },
                {
                    "start": 1032.81,
                    "end": 1032.93,
                    "word": " the",
                    "probability": 0.8115234375
                },
                {
                    "start": 1032.93,
                    "end": 1033.13,
                    "word": " grand",
                    "probability": 0.7314453125
                },
                {
                    "start": 1033.13,
                    "end": 1033.39,
                    "word": " scheme",
                    "probability": 0.89697265625
                },
                {
                    "start": 1033.39,
                    "end": 1033.57,
                    "word": " of",
                    "probability": 0.91162109375
                },
                {
                    "start": 1033.57,
                    "end": 1033.79,
                    "word": " things",
                    "probability": 0.72900390625
                },
                {
                    "start": 1033.79,
                    "end": 1033.93,
                    "word": " to",
                    "probability": 0.86279296875
                },
                {
                    "start": 1033.93,
                    "end": 1034.17,
                    "word": " copy",
                    "probability": 0.80224609375
                },
                {
                    "start": 1034.17,
                    "end": 1034.35,
                    "word": " your",
                    "probability": 0.80078125
                },
                {
                    "start": 1034.35,
                    "end": 1034.59,
                    "word": " model",
                    "probability": 0.8115234375
                },
                {
                    "start": 1034.59,
                    "end": 1035.11,
                    "word": ".",
                    "probability": 0.73828125
                }
            ]
        },
        {
            "id": 145,
            "seek": 103904,
            "start": 1035.17,
            "end": 1042.77,
            "text": " And so, there's no moat, so to speak. So, it really hurts your ability to create a differentiated",
            "tokens": [
                51420,
                400,
                370,
                11,
                456,
                311,
                572,
                705,
                267,
                11,
                370,
                281,
                1710,
                13,
                407,
                11,
                309,
                534,
                11051,
                428,
                3485,
                281,
                1884,
                257,
                27372,
                770,
                51812
            ],
            "temperature": 0.0,
            "avg_logprob": -0.21407860824742267,
            "compression_ratio": 1.727699530516432,
            "no_speech_prob": 0.0015974044799804688,
            "words": [
                {
                    "start": 1035.17,
                    "end": 1035.37,
                    "word": " And",
                    "probability": 0.845703125
                },
                {
                    "start": 1035.37,
                    "end": 1035.73,
                    "word": " so",
                    "probability": 0.884765625
                },
                {
                    "start": 1035.73,
                    "end": 1035.81,
                    "word": ",",
                    "probability": 0.2489013671875
                },
                {
                    "start": 1035.81,
                    "end": 1036.03,
                    "word": " there's",
                    "probability": 0.916015625
                },
                {
                    "start": 1036.03,
                    "end": 1036.53,
                    "word": " no",
                    "probability": 0.904296875
                },
                {
                    "start": 1036.53,
                    "end": 1037.81,
                    "word": " moat",
                    "probability": 0.8447265625
                },
                {
                    "start": 1037.81,
                    "end": 1038.01,
                    "word": ",",
                    "probability": 0.87158203125
                },
                {
                    "start": 1038.01,
                    "end": 1038.29,
                    "word": " so",
                    "probability": 0.9013671875
                },
                {
                    "start": 1038.29,
                    "end": 1038.47,
                    "word": " to",
                    "probability": 0.90234375
                },
                {
                    "start": 1038.47,
                    "end": 1038.91,
                    "word": " speak",
                    "probability": 0.92578125
                },
                {
                    "start": 1038.91,
                    "end": 1039.33,
                    "word": ".",
                    "probability": 0.92236328125
                },
                {
                    "start": 1039.41,
                    "end": 1039.93,
                    "word": " So",
                    "probability": 0.884765625
                },
                {
                    "start": 1039.93,
                    "end": 1039.95,
                    "word": ",",
                    "probability": 0.86962890625
                },
                {
                    "start": 1039.95,
                    "end": 1040.07,
                    "word": " it",
                    "probability": 0.90087890625
                },
                {
                    "start": 1040.07,
                    "end": 1040.29,
                    "word": " really",
                    "probability": 0.7958984375
                },
                {
                    "start": 1040.29,
                    "end": 1040.61,
                    "word": " hurts",
                    "probability": 0.923828125
                },
                {
                    "start": 1040.61,
                    "end": 1040.79,
                    "word": " your",
                    "probability": 0.8037109375
                },
                {
                    "start": 1040.79,
                    "end": 1041.09,
                    "word": " ability",
                    "probability": 0.84912109375
                },
                {
                    "start": 1041.09,
                    "end": 1041.45,
                    "word": " to",
                    "probability": 0.90673828125
                },
                {
                    "start": 1041.45,
                    "end": 1041.81,
                    "word": " create",
                    "probability": 0.8349609375
                },
                {
                    "start": 1042.01,
                    "end": 1042.77,
                    "word": " a differentiated",
                    "probability": 0.82666015625
                }
            ]
        },
        {
            "id": 146,
            "seek": 106818,
            "start": 1042.77,
            "end": 1050.27,
            "text": " product. So, this is a real vulnerability from that standpoint. And we'll get one layer deeper",
            "tokens": [
                50368,
                1674,
                13,
                407,
                11,
                341,
                307,
                257,
                957,
                24210,
                490,
                300,
                15827,
                13,
                400,
                321,
                603,
                483,
                472,
                4583,
                7731,
                50760
            ],
            "temperature": 0.0,
            "avg_logprob": -0.18359374966133724,
            "compression_ratio": 1.5825688073394495,
            "no_speech_prob": 0.00028252601623535156,
            "words": [
                {
                    "start": 1042.77,
                    "end": 1043.17,
                    "word": " product",
                    "probability": 0.261962890625
                },
                {
                    "start": 1043.17,
                    "end": 1043.73,
                    "word": ".",
                    "probability": 0.85400390625
                },
                {
                    "start": 1044.53,
                    "end": 1044.81,
                    "word": " So",
                    "probability": 0.71875
                },
                {
                    "start": 1044.81,
                    "end": 1044.89,
                    "word": ",",
                    "probability": 0.353759765625
                },
                {
                    "start": 1044.89,
                    "end": 1044.97,
                    "word": " this",
                    "probability": 0.87353515625
                },
                {
                    "start": 1044.97,
                    "end": 1045.11,
                    "word": " is",
                    "probability": 0.9033203125
                },
                {
                    "start": 1045.23,
                    "end": 1045.41,
                    "word": " a real",
                    "probability": 0.89013671875
                },
                {
                    "start": 1045.41,
                    "end": 1045.93,
                    "word": " vulnerability",
                    "probability": 0.89208984375
                },
                {
                    "start": 1045.93,
                    "end": 1046.61,
                    "word": " from",
                    "probability": 0.75927734375
                },
                {
                    "start": 1046.61,
                    "end": 1046.93,
                    "word": " that",
                    "probability": 0.87939453125
                },
                {
                    "start": 1046.93,
                    "end": 1047.37,
                    "word": " standpoint",
                    "probability": 0.80029296875
                },
                {
                    "start": 1047.37,
                    "end": 1047.93,
                    "word": ".",
                    "probability": 0.7490234375
                },
                {
                    "start": 1048.41,
                    "end": 1048.63,
                    "word": " And",
                    "probability": 0.86328125
                },
                {
                    "start": 1048.63,
                    "end": 1048.77,
                    "word": " we'll",
                    "probability": 0.91943359375
                },
                {
                    "start": 1048.77,
                    "end": 1049.01,
                    "word": " get",
                    "probability": 0.89892578125
                },
                {
                    "start": 1049.01,
                    "end": 1049.49,
                    "word": " one",
                    "probability": 0.810546875
                },
                {
                    "start": 1049.49,
                    "end": 1049.81,
                    "word": " layer",
                    "probability": 0.7607421875
                },
                {
                    "start": 1049.81,
                    "end": 1050.27,
                    "word": " deeper",
                    "probability": 0.80615234375
                }
            ]
        },
        {
            "id": 147,
            "seek": 106818,
            "start": 1050.27,
            "end": 1057.15,
            "text": " in our next slide. So, when we think about mitigations for model theft, right,",
            "tokens": [
                50760,
                294,
                527,
                958,
                4137,
                13,
                407,
                11,
                562,
                321,
                519,
                466,
                15699,
                763,
                337,
                2316,
                28508,
                11,
                558,
                11,
                51064
            ],
            "temperature": 0.0,
            "avg_logprob": -0.18359374966133724,
            "compression_ratio": 1.5825688073394495,
            "no_speech_prob": 0.00028252601623535156,
            "words": [
                {
                    "start": 1050.27,
                    "end": 1050.97,
                    "word": " in",
                    "probability": 0.89111328125
                },
                {
                    "start": 1050.97,
                    "end": 1051.23,
                    "word": " our",
                    "probability": 0.81787109375
                },
                {
                    "start": 1051.23,
                    "end": 1051.85,
                    "word": " next",
                    "probability": 0.83203125
                },
                {
                    "start": 1051.85,
                    "end": 1052.49,
                    "word": " slide",
                    "probability": 0.86767578125
                },
                {
                    "start": 1052.49,
                    "end": 1053.05,
                    "word": ".",
                    "probability": 0.99267578125
                },
                {
                    "start": 1053.35,
                    "end": 1053.91,
                    "word": " So",
                    "probability": 0.87939453125
                },
                {
                    "start": 1053.91,
                    "end": 1053.93,
                    "word": ",",
                    "probability": 0.9697265625
                },
                {
                    "start": 1053.93,
                    "end": 1054.07,
                    "word": " when",
                    "probability": 0.87255859375
                },
                {
                    "start": 1054.07,
                    "end": 1054.21,
                    "word": " we",
                    "probability": 0.90283203125
                },
                {
                    "start": 1054.21,
                    "end": 1054.39,
                    "word": " think",
                    "probability": 0.80810546875
                },
                {
                    "start": 1054.39,
                    "end": 1054.61,
                    "word": " about",
                    "probability": 0.82763671875
                },
                {
                    "start": 1054.61,
                    "end": 1055.35,
                    "word": " mitigations",
                    "probability": 0.838623046875
                },
                {
                    "start": 1055.35,
                    "end": 1055.63,
                    "word": " for",
                    "probability": 0.9013671875
                },
                {
                    "start": 1055.63,
                    "end": 1056.01,
                    "word": " model",
                    "probability": 0.8046875
                },
                {
                    "start": 1056.01,
                    "end": 1056.47,
                    "word": " theft",
                    "probability": 0.76611328125
                },
                {
                    "start": 1056.47,
                    "end": 1056.59,
                    "word": ",",
                    "probability": 0.95263671875
                },
                {
                    "start": 1056.59,
                    "end": 1056.85,
                    "word": " right",
                    "probability": 0.814453125
                },
                {
                    "start": 1056.85,
                    "end": 1057.15,
                    "word": ",",
                    "probability": 0.64990234375
                }
            ]
        },
        {
            "id": 148,
            "seek": 106818,
            "start": 1057.15,
            "end": 1063.13,
            "text": " there's two. And the first one is limiting queries to the model. And this one kind of",
            "tokens": [
                51108,
                456,
                311,
                732,
                13,
                400,
                264,
                700,
                472,
                307,
                22083,
                24109,
                281,
                264,
                2316,
                13,
                400,
                341,
                472,
                733,
                295,
                51372
            ],
            "temperature": 0.0,
            "avg_logprob": -0.18359374966133724,
            "compression_ratio": 1.5825688073394495,
            "no_speech_prob": 0.00028252601623535156,
            "words": [
                {
                    "start": 1057.15,
                    "end": 1058.05,
                    "word": " there's",
                    "probability": 0.918212890625
                },
                {
                    "start": 1058.05,
                    "end": 1058.47,
                    "word": " two",
                    "probability": 0.82177734375
                },
                {
                    "start": 1058.47,
                    "end": 1059.03,
                    "word": ".",
                    "probability": 0.92822265625
                },
                {
                    "start": 1059.43,
                    "end": 1059.79,
                    "word": " And",
                    "probability": 0.86962890625
                },
                {
                    "start": 1059.79,
                    "end": 1059.93,
                    "word": " the",
                    "probability": 0.8193359375
                },
                {
                    "start": 1059.93,
                    "end": 1060.11,
                    "word": " first",
                    "probability": 0.7978515625
                },
                {
                    "start": 1060.11,
                    "end": 1060.29,
                    "word": " one",
                    "probability": 0.81494140625
                },
                {
                    "start": 1060.29,
                    "end": 1060.43,
                    "word": " is",
                    "probability": 0.90771484375
                },
                {
                    "start": 1060.43,
                    "end": 1060.81,
                    "word": " limiting",
                    "probability": 0.8818359375
                },
                {
                    "start": 1060.81,
                    "end": 1061.25,
                    "word": " queries",
                    "probability": 0.8369140625
                },
                {
                    "start": 1061.25,
                    "end": 1061.43,
                    "word": " to",
                    "probability": 0.90087890625
                },
                {
                    "start": 1061.43,
                    "end": 1061.55,
                    "word": " the",
                    "probability": 0.81640625
                },
                {
                    "start": 1061.55,
                    "end": 1061.83,
                    "word": " model",
                    "probability": 0.822265625
                },
                {
                    "start": 1061.83,
                    "end": 1062.07,
                    "word": ".",
                    "probability": 0.9189453125
                },
                {
                    "start": 1062.07,
                    "end": 1062.21,
                    "word": " And",
                    "probability": 0.86962890625
                },
                {
                    "start": 1062.21,
                    "end": 1062.37,
                    "word": " this",
                    "probability": 0.87890625
                },
                {
                    "start": 1062.37,
                    "end": 1062.57,
                    "word": " one",
                    "probability": 0.8076171875
                },
                {
                    "start": 1062.57,
                    "end": 1062.85,
                    "word": " kind",
                    "probability": 0.82568359375
                },
                {
                    "start": 1062.85,
                    "end": 1063.13,
                    "word": " of",
                    "probability": 0.9091796875
                }
            ]
        },
        {
            "id": 149,
            "seek": 106818,
            "start": 1064.67,
            "end": 1071.91,
            "text": " feels bad. Because you don't want to limit queries to your model too much, since most",
            "tokens": [
                51452,
                3417,
                1578,
                13,
                1436,
                291,
                500,
                380,
                528,
                281,
                4948,
                24109,
                281,
                428,
                2316,
                886,
                709,
                11,
                1670,
                881,
                51816
            ],
            "temperature": 0.0,
            "avg_logprob": -0.18359374966133724,
            "compression_ratio": 1.5825688073394495,
            "no_speech_prob": 0.00028252601623535156,
            "words": [
                {
                    "start": 1064.67,
                    "end": 1065.23,
                    "word": " feels",
                    "probability": 0.8154296875
                },
                {
                    "start": 1065.23,
                    "end": 1065.79,
                    "word": " bad",
                    "probability": 0.888671875
                },
                {
                    "start": 1065.79,
                    "end": 1066.35,
                    "word": ".",
                    "probability": 0.81591796875
                },
                {
                    "start": 1066.77,
                    "end": 1067.33,
                    "word": " Because",
                    "probability": 0.9072265625
                },
                {
                    "start": 1067.33,
                    "end": 1067.49,
                    "word": " you",
                    "probability": 0.89990234375
                },
                {
                    "start": 1067.49,
                    "end": 1067.69,
                    "word": " don't",
                    "probability": 0.943359375
                },
                {
                    "start": 1067.69,
                    "end": 1067.95,
                    "word": " want",
                    "probability": 0.8134765625
                },
                {
                    "start": 1067.95,
                    "end": 1068.11,
                    "word": " to",
                    "probability": 0.90087890625
                },
                {
                    "start": 1068.11,
                    "end": 1068.35,
                    "word": " limit",
                    "probability": 0.853515625
                },
                {
                    "start": 1068.35,
                    "end": 1068.71,
                    "word": " queries",
                    "probability": 0.83251953125
                },
                {
                    "start": 1068.71,
                    "end": 1068.83,
                    "word": " to",
                    "probability": 0.90087890625
                },
                {
                    "start": 1068.83,
                    "end": 1068.95,
                    "word": " your",
                    "probability": 0.78369140625
                },
                {
                    "start": 1068.95,
                    "end": 1069.19,
                    "word": " model",
                    "probability": 0.7744140625
                },
                {
                    "start": 1069.19,
                    "end": 1069.65,
                    "word": " too",
                    "probability": 0.78857421875
                },
                {
                    "start": 1069.65,
                    "end": 1070.21,
                    "word": " much",
                    "probability": 0.81884765625
                },
                {
                    "start": 1070.21,
                    "end": 1070.57,
                    "word": ",",
                    "probability": 0.2861328125
                },
                {
                    "start": 1070.57,
                    "end": 1071.23,
                    "word": " since",
                    "probability": 0.81494140625
                },
                {
                    "start": 1071.23,
                    "end": 1071.91,
                    "word": " most",
                    "probability": 0.81689453125
                }
            ]
        },
        {
            "id": 150,
            "seek": 109764,
            "start": 1071.91,
            "end": 1077.81,
            "text": " queries to your model are going to be legitimate. For the most part, you're expecting these inputs",
            "tokens": [
                50364,
                24109,
                281,
                428,
                2316,
                366,
                516,
                281,
                312,
                17956,
                13,
                1171,
                264,
                881,
                644,
                11,
                291,
                434,
                9650,
                613,
                15743,
                50664
            ],
            "temperature": 0.0,
            "avg_logprob": -0.1934871495327103,
            "compression_ratio": 1.8125,
            "no_speech_prob": 0.002471923828125,
            "words": [
                {
                    "start": 1071.91,
                    "end": 1072.25,
                    "word": " queries",
                    "probability": 0.47265625
                },
                {
                    "start": 1072.25,
                    "end": 1072.41,
                    "word": " to",
                    "probability": 0.90185546875
                },
                {
                    "start": 1072.41,
                    "end": 1072.51,
                    "word": " your",
                    "probability": 0.7998046875
                },
                {
                    "start": 1072.51,
                    "end": 1072.73,
                    "word": " model",
                    "probability": 0.8173828125
                },
                {
                    "start": 1072.73,
                    "end": 1072.89,
                    "word": " are",
                    "probability": 0.8779296875
                },
                {
                    "start": 1072.89,
                    "end": 1073.07,
                    "word": " going",
                    "probability": 0.8349609375
                },
                {
                    "start": 1073.07,
                    "end": 1073.19,
                    "word": " to",
                    "probability": 0.904296875
                },
                {
                    "start": 1073.19,
                    "end": 1073.33,
                    "word": " be",
                    "probability": 0.89208984375
                },
                {
                    "start": 1073.33,
                    "end": 1073.79,
                    "word": " legitimate",
                    "probability": 0.77490234375
                },
                {
                    "start": 1073.79,
                    "end": 1074.31,
                    "word": ".",
                    "probability": 0.60791015625
                },
                {
                    "start": 1074.67,
                    "end": 1074.93,
                    "word": " For",
                    "probability": 0.464111328125
                },
                {
                    "start": 1074.93,
                    "end": 1075.09,
                    "word": " the",
                    "probability": 0.80517578125
                },
                {
                    "start": 1075.09,
                    "end": 1075.29,
                    "word": " most",
                    "probability": 0.8076171875
                },
                {
                    "start": 1075.29,
                    "end": 1075.57,
                    "word": " part",
                    "probability": 0.8251953125
                },
                {
                    "start": 1075.57,
                    "end": 1075.71,
                    "word": ",",
                    "probability": 0.6806640625
                },
                {
                    "start": 1075.71,
                    "end": 1075.85,
                    "word": " you're",
                    "probability": 0.90185546875
                },
                {
                    "start": 1075.85,
                    "end": 1076.49,
                    "word": " expecting",
                    "probability": 0.75048828125
                },
                {
                    "start": 1076.49,
                    "end": 1077.13,
                    "word": " these",
                    "probability": 0.7041015625
                },
                {
                    "start": 1077.13,
                    "end": 1077.81,
                    "word": " inputs",
                    "probability": 0.78173828125
                }
            ]
        },
        {
            "id": 151,
            "seek": 109764,
            "start": 1077.81,
            "end": 1082.61,
            "text": " and giving outputs, and you created this model, and you're hosting this model, because it's",
            "tokens": [
                50664,
                293,
                2902,
                23930,
                11,
                293,
                291,
                2942,
                341,
                2316,
                11,
                293,
                291,
                434,
                16058,
                341,
                2316,
                11,
                570,
                309,
                311,
                50896
            ],
            "temperature": 0.0,
            "avg_logprob": -0.1934871495327103,
            "compression_ratio": 1.8125,
            "no_speech_prob": 0.002471923828125,
            "words": [
                {
                    "start": 1077.81,
                    "end": 1078.11,
                    "word": " and",
                    "probability": 0.86474609375
                },
                {
                    "start": 1078.11,
                    "end": 1078.37,
                    "word": " giving",
                    "probability": 0.84765625
                },
                {
                    "start": 1078.37,
                    "end": 1078.93,
                    "word": " outputs",
                    "probability": 0.76611328125
                },
                {
                    "start": 1078.93,
                    "end": 1079.09,
                    "word": ",",
                    "probability": 0.338134765625
                },
                {
                    "start": 1079.09,
                    "end": 1079.15,
                    "word": " and",
                    "probability": 0.88232421875
                },
                {
                    "start": 1079.15,
                    "end": 1079.35,
                    "word": " you",
                    "probability": 0.89111328125
                },
                {
                    "start": 1079.35,
                    "end": 1080.11,
                    "word": " created",
                    "probability": 0.70166015625
                },
                {
                    "start": 1080.11,
                    "end": 1080.41,
                    "word": " this",
                    "probability": 0.88134765625
                },
                {
                    "start": 1080.41,
                    "end": 1080.67,
                    "word": " model",
                    "probability": 0.83203125
                },
                {
                    "start": 1080.67,
                    "end": 1080.75,
                    "word": ",",
                    "probability": 0.36083984375
                },
                {
                    "start": 1080.75,
                    "end": 1080.83,
                    "word": " and",
                    "probability": 0.88671875
                },
                {
                    "start": 1080.83,
                    "end": 1080.99,
                    "word": " you're",
                    "probability": 0.920166015625
                },
                {
                    "start": 1080.99,
                    "end": 1081.33,
                    "word": " hosting",
                    "probability": 0.76025390625
                },
                {
                    "start": 1081.33,
                    "end": 1081.57,
                    "word": " this",
                    "probability": 0.8818359375
                },
                {
                    "start": 1081.57,
                    "end": 1081.81,
                    "word": " model",
                    "probability": 0.818359375
                },
                {
                    "start": 1081.81,
                    "end": 1081.87,
                    "word": ",",
                    "probability": 0.36767578125
                },
                {
                    "start": 1081.87,
                    "end": 1082.35,
                    "word": " because",
                    "probability": 0.8408203125
                },
                {
                    "start": 1082.35,
                    "end": 1082.61,
                    "word": " it's",
                    "probability": 0.94921875
                }
            ]
        },
        {
            "id": 152,
            "seek": 109764,
            "start": 1082.61,
            "end": 1088.81,
            "text": " supposed to be useful to someone. And for most people, because of just the corporatization",
            "tokens": [
                50896,
                3442,
                281,
                312,
                4420,
                281,
                1580,
                13,
                400,
                337,
                881,
                561,
                11,
                570,
                295,
                445,
                264,
                6804,
                267,
                2144,
                51224
            ],
            "temperature": 0.0,
            "avg_logprob": -0.1934871495327103,
            "compression_ratio": 1.8125,
            "no_speech_prob": 0.002471923828125,
            "words": [
                {
                    "start": 1082.61,
                    "end": 1082.79,
                    "word": " supposed",
                    "probability": 0.91064453125
                },
                {
                    "start": 1082.79,
                    "end": 1082.95,
                    "word": " to",
                    "probability": 0.90576171875
                },
                {
                    "start": 1082.95,
                    "end": 1083.15,
                    "word": " be",
                    "probability": 0.90087890625
                },
                {
                    "start": 1083.15,
                    "end": 1083.53,
                    "word": " useful",
                    "probability": 0.75537109375
                },
                {
                    "start": 1083.53,
                    "end": 1083.83,
                    "word": " to",
                    "probability": 0.904296875
                },
                {
                    "start": 1083.83,
                    "end": 1084.33,
                    "word": " someone",
                    "probability": 0.82958984375
                },
                {
                    "start": 1084.33,
                    "end": 1084.85,
                    "word": ".",
                    "probability": 0.91650390625
                },
                {
                    "start": 1085.01,
                    "end": 1085.15,
                    "word": " And",
                    "probability": 0.73681640625
                },
                {
                    "start": 1085.15,
                    "end": 1085.39,
                    "word": " for",
                    "probability": 0.86376953125
                },
                {
                    "start": 1085.39,
                    "end": 1085.73,
                    "word": " most",
                    "probability": 0.80712890625
                },
                {
                    "start": 1085.73,
                    "end": 1086.11,
                    "word": " people",
                    "probability": 0.931640625
                },
                {
                    "start": 1086.11,
                    "end": 1086.31,
                    "word": ",",
                    "probability": 0.8916015625
                },
                {
                    "start": 1086.31,
                    "end": 1086.61,
                    "word": " because",
                    "probability": 0.81494140625
                },
                {
                    "start": 1086.61,
                    "end": 1087.07,
                    "word": " of",
                    "probability": 0.8623046875
                },
                {
                    "start": 1087.07,
                    "end": 1087.51,
                    "word": " just",
                    "probability": 0.787109375
                },
                {
                    "start": 1087.51,
                    "end": 1087.93,
                    "word": " the",
                    "probability": 0.81396484375
                },
                {
                    "start": 1087.93,
                    "end": 1088.81,
                    "word": " corporatization",
                    "probability": 0.9524739583333334
                }
            ]
        },
        {
            "id": 153,
            "seek": 109764,
            "start": 1088.81,
            "end": 1094.95,
            "text": " of information security and the corporatization of artificial intelligence, it's probably the",
            "tokens": [
                51224,
                295,
                1589,
                3825,
                293,
                264,
                6804,
                267,
                2144,
                295,
                11677,
                7599,
                11,
                309,
                311,
                1391,
                264,
                51512
            ],
            "temperature": 0.0,
            "avg_logprob": -0.1934871495327103,
            "compression_ratio": 1.8125,
            "no_speech_prob": 0.002471923828125,
            "words": [
                {
                    "start": 1088.81,
                    "end": 1089.35,
                    "word": " of",
                    "probability": 0.90576171875
                },
                {
                    "start": 1089.35,
                    "end": 1089.85,
                    "word": " information",
                    "probability": 0.693359375
                },
                {
                    "start": 1089.85,
                    "end": 1090.35,
                    "word": " security",
                    "probability": 0.9267578125
                },
                {
                    "start": 1090.35,
                    "end": 1090.63,
                    "word": " and",
                    "probability": 0.71142578125
                },
                {
                    "start": 1090.63,
                    "end": 1090.81,
                    "word": " the",
                    "probability": 0.81396484375
                },
                {
                    "start": 1090.81,
                    "end": 1091.67,
                    "word": " corporatization",
                    "probability": 0.9637044270833334
                },
                {
                    "start": 1091.67,
                    "end": 1092.31,
                    "word": " of",
                    "probability": 0.90380859375
                },
                {
                    "start": 1092.31,
                    "end": 1092.75,
                    "word": " artificial",
                    "probability": 0.80126953125
                },
                {
                    "start": 1092.75,
                    "end": 1093.37,
                    "word": " intelligence",
                    "probability": 0.69921875
                },
                {
                    "start": 1093.37,
                    "end": 1093.67,
                    "word": ",",
                    "probability": 0.978515625
                },
                {
                    "start": 1093.67,
                    "end": 1094.41,
                    "word": " it's",
                    "probability": 0.947998046875
                },
                {
                    "start": 1094.41,
                    "end": 1094.73,
                    "word": " probably",
                    "probability": 0.9189453125
                },
                {
                    "start": 1094.73,
                    "end": 1094.95,
                    "word": " the",
                    "probability": 0.81005859375
                }
            ]
        },
        {
            "id": 154,
            "seek": 109764,
            "start": 1094.95,
            "end": 1101.37,
            "text": " people who pay your paycheck who want this. And so, inherently, you want to have really,",
            "tokens": [
                51512,
                561,
                567,
                1689,
                428,
                35639,
                567,
                528,
                341,
                13,
                400,
                370,
                11,
                27993,
                11,
                291,
                528,
                281,
                362,
                534,
                11,
                51772
            ],
            "temperature": 0.0,
            "avg_logprob": -0.1934871495327103,
            "compression_ratio": 1.8125,
            "no_speech_prob": 0.002471923828125,
            "words": [
                {
                    "start": 1094.95,
                    "end": 1095.17,
                    "word": " people",
                    "probability": 0.92529296875
                },
                {
                    "start": 1095.17,
                    "end": 1095.33,
                    "word": " who",
                    "probability": 0.8046875
                },
                {
                    "start": 1095.33,
                    "end": 1095.47,
                    "word": " pay",
                    "probability": 0.72509765625
                },
                {
                    "start": 1095.47,
                    "end": 1095.65,
                    "word": " your",
                    "probability": 0.75341796875
                },
                {
                    "start": 1095.65,
                    "end": 1096.01,
                    "word": " paycheck",
                    "probability": 0.8564453125
                },
                {
                    "start": 1096.01,
                    "end": 1096.89,
                    "word": " who",
                    "probability": 0.7744140625
                },
                {
                    "start": 1096.89,
                    "end": 1097.43,
                    "word": " want",
                    "probability": 0.82080078125
                },
                {
                    "start": 1097.43,
                    "end": 1097.77,
                    "word": " this",
                    "probability": 0.87890625
                },
                {
                    "start": 1097.77,
                    "end": 1098.29,
                    "word": ".",
                    "probability": 0.9658203125
                },
                {
                    "start": 1098.31,
                    "end": 1098.43,
                    "word": " And",
                    "probability": 0.75537109375
                },
                {
                    "start": 1098.43,
                    "end": 1098.77,
                    "word": " so",
                    "probability": 0.8740234375
                },
                {
                    "start": 1098.77,
                    "end": 1098.87,
                    "word": ",",
                    "probability": 0.295166015625
                },
                {
                    "start": 1098.87,
                    "end": 1099.31,
                    "word": " inherently",
                    "probability": 0.82763671875
                },
                {
                    "start": 1099.31,
                    "end": 1099.45,
                    "word": ",",
                    "probability": 0.77392578125
                },
                {
                    "start": 1099.45,
                    "end": 1099.55,
                    "word": " you",
                    "probability": 0.8935546875
                },
                {
                    "start": 1099.55,
                    "end": 1099.65,
                    "word": " want",
                    "probability": 0.72998046875
                },
                {
                    "start": 1099.65,
                    "end": 1099.71,
                    "word": " to",
                    "probability": 0.900390625
                },
                {
                    "start": 1099.71,
                    "end": 1099.83,
                    "word": " have",
                    "probability": 0.88232421875
                },
                {
                    "start": 1099.83,
                    "end": 1100.09,
                    "word": " really",
                    "probability": 0.69482421875
                },
                {
                    "start": 1100.09,
                    "end": 1101.37,
                    "word": ",",
                    "probability": 0.916015625
                }
            ]
        },
        {
            "id": 155,
            "seek": 112470,
            "start": 1101.99,
            "end": 1107.85,
            "text": " but you want to balance that really high uptime and the fact that there may be a single endpoint",
            "tokens": [
                50400,
                457,
                291,
                528,
                281,
                4772,
                300,
                534,
                1090,
                493,
                3766,
                293,
                264,
                1186,
                300,
                456,
                815,
                312,
                257,
                2167,
                35795,
                50696
            ],
            "temperature": 0.0,
            "avg_logprob": -0.21711482644774194,
            "compression_ratio": 1.7342995169082125,
            "no_speech_prob": 0.00011771917343139648,
            "words": [
                {
                    "start": 1101.99,
                    "end": 1102.27,
                    "word": " but",
                    "probability": 0.315185546875
                },
                {
                    "start": 1102.27,
                    "end": 1102.73,
                    "word": " you",
                    "probability": 0.8837890625
                },
                {
                    "start": 1102.73,
                    "end": 1102.89,
                    "word": " want",
                    "probability": 0.76123046875
                },
                {
                    "start": 1102.89,
                    "end": 1103.05,
                    "word": " to",
                    "probability": 0.90283203125
                },
                {
                    "start": 1103.05,
                    "end": 1103.49,
                    "word": " balance",
                    "probability": 0.919921875
                },
                {
                    "start": 1103.49,
                    "end": 1103.75,
                    "word": " that",
                    "probability": 0.8798828125
                },
                {
                    "start": 1103.75,
                    "end": 1104.13,
                    "word": " really",
                    "probability": 0.76806640625
                },
                {
                    "start": 1104.13,
                    "end": 1104.49,
                    "word": " high",
                    "probability": 0.857421875
                },
                {
                    "start": 1104.49,
                    "end": 1105.15,
                    "word": " uptime",
                    "probability": 0.6348876953125
                },
                {
                    "start": 1105.15,
                    "end": 1105.47,
                    "word": " and",
                    "probability": 0.806640625
                },
                {
                    "start": 1105.47,
                    "end": 1105.59,
                    "word": " the",
                    "probability": 0.80712890625
                },
                {
                    "start": 1105.59,
                    "end": 1105.77,
                    "word": " fact",
                    "probability": 0.81201171875
                },
                {
                    "start": 1105.77,
                    "end": 1105.93,
                    "word": " that",
                    "probability": 0.869140625
                },
                {
                    "start": 1105.93,
                    "end": 1106.09,
                    "word": " there",
                    "probability": 0.853515625
                },
                {
                    "start": 1106.09,
                    "end": 1106.37,
                    "word": " may",
                    "probability": 0.89990234375
                },
                {
                    "start": 1106.37,
                    "end": 1106.57,
                    "word": " be",
                    "probability": 0.9091796875
                },
                {
                    "start": 1106.81,
                    "end": 1107.33,
                    "word": " a single",
                    "probability": 0.8671875
                },
                {
                    "start": 1107.33,
                    "end": 1107.85,
                    "word": " endpoint",
                    "probability": 0.56201171875
                }
            ]
        },
        {
            "id": 156,
            "seek": 112470,
            "start": 1107.85,
            "end": 1114.11,
            "text": " that's making a lot of legitimate queries to your model with the fact that there may be an attacker",
            "tokens": [
                50696,
                300,
                311,
                1455,
                257,
                688,
                295,
                17956,
                24109,
                281,
                428,
                2316,
                365,
                264,
                1186,
                300,
                456,
                815,
                312,
                364,
                35871,
                51008
            ],
            "temperature": 0.0,
            "avg_logprob": -0.21711482644774194,
            "compression_ratio": 1.7342995169082125,
            "no_speech_prob": 0.00011771917343139648,
            "words": [
                {
                    "start": 1107.85,
                    "end": 1108.29,
                    "word": " that's",
                    "probability": 0.920166015625
                },
                {
                    "start": 1108.29,
                    "end": 1108.53,
                    "word": " making",
                    "probability": 0.79248046875
                },
                {
                    "start": 1108.75,
                    "end": 1108.93,
                    "word": " a lot",
                    "probability": 0.88818359375
                },
                {
                    "start": 1108.93,
                    "end": 1109.27,
                    "word": " of",
                    "probability": 0.90869140625
                },
                {
                    "start": 1109.27,
                    "end": 1109.93,
                    "word": " legitimate",
                    "probability": 0.7587890625
                },
                {
                    "start": 1109.93,
                    "end": 1110.43,
                    "word": " queries",
                    "probability": 0.87109375
                },
                {
                    "start": 1110.43,
                    "end": 1110.65,
                    "word": " to",
                    "probability": 0.90283203125
                },
                {
                    "start": 1110.65,
                    "end": 1110.79,
                    "word": " your",
                    "probability": 0.806640625
                },
                {
                    "start": 1110.79,
                    "end": 1111.05,
                    "word": " model",
                    "probability": 0.8212890625
                },
                {
                    "start": 1111.05,
                    "end": 1111.91,
                    "word": " with",
                    "probability": 0.66162109375
                },
                {
                    "start": 1111.91,
                    "end": 1112.13,
                    "word": " the",
                    "probability": 0.81591796875
                },
                {
                    "start": 1112.13,
                    "end": 1112.41,
                    "word": " fact",
                    "probability": 0.8125
                },
                {
                    "start": 1112.41,
                    "end": 1112.73,
                    "word": " that",
                    "probability": 0.88037109375
                },
                {
                    "start": 1112.73,
                    "end": 1112.97,
                    "word": " there",
                    "probability": 0.85498046875
                },
                {
                    "start": 1112.97,
                    "end": 1113.23,
                    "word": " may",
                    "probability": 0.904296875
                },
                {
                    "start": 1113.23,
                    "end": 1113.45,
                    "word": " be",
                    "probability": 0.90673828125
                },
                {
                    "start": 1113.45,
                    "end": 1113.65,
                    "word": " an",
                    "probability": 0.91162109375
                },
                {
                    "start": 1113.65,
                    "end": 1114.11,
                    "word": " attacker",
                    "probability": 0.755859375
                }
            ]
        },
        {
            "id": 157,
            "seek": 112470,
            "start": 1114.11,
            "end": 1120.67,
            "text": " who's trying to use your model for evil. Trying to steal your model.",
            "tokens": [
                51008,
                567,
                311,
                1382,
                281,
                764,
                428,
                2316,
                337,
                6724,
                13,
                20180,
                281,
                11009,
                428,
                2316,
                13,
                51300
            ],
            "temperature": 0.0,
            "avg_logprob": -0.21711482644774194,
            "compression_ratio": 1.7342995169082125,
            "no_speech_prob": 0.00011771917343139648,
            "words": [
                {
                    "start": 1114.11,
                    "end": 1114.59,
                    "word": " who's",
                    "probability": 0.8935546875
                },
                {
                    "start": 1114.59,
                    "end": 1115.07,
                    "word": " trying",
                    "probability": 0.80517578125
                },
                {
                    "start": 1115.07,
                    "end": 1115.73,
                    "word": " to",
                    "probability": 0.90673828125
                },
                {
                    "start": 1115.73,
                    "end": 1116.97,
                    "word": " use",
                    "probability": 0.75927734375
                },
                {
                    "start": 1116.97,
                    "end": 1117.17,
                    "word": " your",
                    "probability": 0.80615234375
                },
                {
                    "start": 1117.17,
                    "end": 1117.55,
                    "word": " model",
                    "probability": 0.8056640625
                },
                {
                    "start": 1117.55,
                    "end": 1117.97,
                    "word": " for",
                    "probability": 0.8994140625
                },
                {
                    "start": 1117.97,
                    "end": 1118.37,
                    "word": " evil",
                    "probability": 0.89697265625
                },
                {
                    "start": 1118.37,
                    "end": 1118.65,
                    "word": ".",
                    "probability": 0.1007080078125
                },
                {
                    "start": 1118.65,
                    "end": 1118.89,
                    "word": " Trying",
                    "probability": 0.81884765625
                },
                {
                    "start": 1118.89,
                    "end": 1119.13,
                    "word": " to",
                    "probability": 0.9072265625
                },
                {
                    "start": 1119.13,
                    "end": 1119.47,
                    "word": " steal",
                    "probability": 0.85107421875
                },
                {
                    "start": 1119.47,
                    "end": 1119.75,
                    "word": " your",
                    "probability": 0.8056640625
                },
                {
                    "start": 1119.75,
                    "end": 1120.11,
                    "word": " model",
                    "probability": 0.81396484375
                },
                {
                    "start": 1120.11,
                    "end": 1120.67,
                    "word": ".",
                    "probability": 0.88134765625
                }
            ]
        },
        {
            "id": 158,
            "seek": 112470,
            "start": 1122.37,
            "end": 1128.43,
            "text": " And so, the other thing we can do is limiting information returned from the model. And we saw",
            "tokens": [
                51420,
                400,
                370,
                11,
                264,
                661,
                551,
                321,
                393,
                360,
                307,
                22083,
                1589,
                8752,
                490,
                264,
                2316,
                13,
                400,
                321,
                1866,
                51712
            ],
            "temperature": 0.0,
            "avg_logprob": -0.21711482644774194,
            "compression_ratio": 1.7342995169082125,
            "no_speech_prob": 0.00011771917343139648,
            "words": [
                {
                    "start": 1122.37,
                    "end": 1122.73,
                    "word": " And",
                    "probability": 0.73974609375
                },
                {
                    "start": 1122.73,
                    "end": 1122.85,
                    "word": " so",
                    "probability": 0.89208984375
                },
                {
                    "start": 1122.85,
                    "end": 1122.89,
                    "word": ",",
                    "probability": 0.214599609375
                },
                {
                    "start": 1122.89,
                    "end": 1123.01,
                    "word": " the",
                    "probability": 0.81103515625
                },
                {
                    "start": 1123.01,
                    "end": 1123.19,
                    "word": " other",
                    "probability": 0.76806640625
                },
                {
                    "start": 1123.19,
                    "end": 1123.49,
                    "word": " thing",
                    "probability": 0.81494140625
                },
                {
                    "start": 1123.49,
                    "end": 1123.67,
                    "word": " we",
                    "probability": 0.89697265625
                },
                {
                    "start": 1123.67,
                    "end": 1123.87,
                    "word": " can",
                    "probability": 0.90185546875
                },
                {
                    "start": 1123.87,
                    "end": 1124.17,
                    "word": " do",
                    "probability": 0.9013671875
                },
                {
                    "start": 1124.17,
                    "end": 1125.15,
                    "word": " is",
                    "probability": 0.88671875
                },
                {
                    "start": 1125.15,
                    "end": 1125.59,
                    "word": " limiting",
                    "probability": 0.8466796875
                },
                {
                    "start": 1125.59,
                    "end": 1126.27,
                    "word": " information",
                    "probability": 0.74365234375
                },
                {
                    "start": 1126.27,
                    "end": 1126.79,
                    "word": " returned",
                    "probability": 0.759765625
                },
                {
                    "start": 1126.79,
                    "end": 1127.03,
                    "word": " from",
                    "probability": 0.82470703125
                },
                {
                    "start": 1127.03,
                    "end": 1127.21,
                    "word": " the",
                    "probability": 0.818359375
                },
                {
                    "start": 1127.21,
                    "end": 1127.53,
                    "word": " model",
                    "probability": 0.82080078125
                },
                {
                    "start": 1127.53,
                    "end": 1127.73,
                    "word": ".",
                    "probability": 0.5283203125
                },
                {
                    "start": 1127.73,
                    "end": 1127.93,
                    "word": " And",
                    "probability": 0.87451171875
                },
                {
                    "start": 1127.93,
                    "end": 1128.09,
                    "word": " we",
                    "probability": 0.8828125
                },
                {
                    "start": 1128.09,
                    "end": 1128.43,
                    "word": " saw",
                    "probability": 0.810546875
                }
            ]
        },
        {
            "id": 159,
            "seek": 114856,
            "start": 1128.43,
            "end": 1134.67,
            "text": " the efficacy of this with SQL injection, where, like, back in the olden days,",
            "tokens": [
                50388,
                264,
                33492,
                295,
                341,
                365,
                19200,
                22873,
                11,
                689,
                11,
                411,
                11,
                646,
                294,
                264,
                1331,
                268,
                1708,
                11,
                50652
            ],
            "temperature": 0.0,
            "avg_logprob": -0.22656250044481077,
            "compression_ratio": 1.4262295081967213,
            "no_speech_prob": 0.0014324188232421875,
            "words": [
                {
                    "start": 1128.43,
                    "end": 1129.17,
                    "word": " the",
                    "probability": 0.106201171875
                },
                {
                    "start": 1129.17,
                    "end": 1129.63,
                    "word": " efficacy",
                    "probability": 0.80126953125
                },
                {
                    "start": 1129.63,
                    "end": 1129.97,
                    "word": " of",
                    "probability": 0.89990234375
                },
                {
                    "start": 1129.97,
                    "end": 1130.19,
                    "word": " this",
                    "probability": 0.86328125
                },
                {
                    "start": 1130.19,
                    "end": 1130.43,
                    "word": " with",
                    "probability": 0.77978515625
                },
                {
                    "start": 1130.43,
                    "end": 1130.79,
                    "word": " SQL",
                    "probability": 0.83837890625
                },
                {
                    "start": 1130.79,
                    "end": 1131.29,
                    "word": " injection",
                    "probability": 0.78173828125
                },
                {
                    "start": 1131.29,
                    "end": 1131.93,
                    "word": ",",
                    "probability": 0.48779296875
                },
                {
                    "start": 1131.93,
                    "end": 1132.21,
                    "word": " where",
                    "probability": 0.86474609375
                },
                {
                    "start": 1132.21,
                    "end": 1132.37,
                    "word": ",",
                    "probability": 0.13720703125
                },
                {
                    "start": 1132.37,
                    "end": 1132.57,
                    "word": " like",
                    "probability": 0.7587890625
                },
                {
                    "start": 1132.57,
                    "end": 1132.73,
                    "word": ",",
                    "probability": 0.7783203125
                },
                {
                    "start": 1132.73,
                    "end": 1132.95,
                    "word": " back",
                    "probability": 0.80810546875
                },
                {
                    "start": 1132.95,
                    "end": 1133.11,
                    "word": " in",
                    "probability": 0.90673828125
                },
                {
                    "start": 1133.11,
                    "end": 1133.29,
                    "word": " the",
                    "probability": 0.8125
                },
                {
                    "start": 1133.29,
                    "end": 1133.77,
                    "word": " olden",
                    "probability": 0.921875
                },
                {
                    "start": 1133.77,
                    "end": 1134.31,
                    "word": " days",
                    "probability": 0.7822265625
                },
                {
                    "start": 1134.31,
                    "end": 1134.67,
                    "word": ",",
                    "probability": 0.8671875
                }
            ]
        },
        {
            "id": 160,
            "seek": 114856,
            "start": 1135.55,
            "end": 1144.47,
            "text": " you know, a million years ago, because it's 2020, and so time doesn't matter anymore,",
            "tokens": [
                50724,
                291,
                458,
                11,
                257,
                2459,
                924,
                2057,
                11,
                570,
                309,
                311,
                4808,
                11,
                293,
                370,
                565,
                1177,
                380,
                1871,
                3602,
                11,
                51124
            ],
            "temperature": 0.0,
            "avg_logprob": -0.22656250044481077,
            "compression_ratio": 1.4262295081967213,
            "no_speech_prob": 0.0014324188232421875,
            "words": [
                {
                    "start": 1135.55,
                    "end": 1135.89,
                    "word": " you",
                    "probability": 0.642578125
                },
                {
                    "start": 1135.89,
                    "end": 1136.05,
                    "word": " know",
                    "probability": 0.8154296875
                },
                {
                    "start": 1136.05,
                    "end": 1136.49,
                    "word": ",",
                    "probability": 0.96240234375
                },
                {
                    "start": 1136.75,
                    "end": 1137.01,
                    "word": " a million",
                    "probability": 0.77490234375
                },
                {
                    "start": 1137.01,
                    "end": 1137.43,
                    "word": " years",
                    "probability": 0.87548828125
                },
                {
                    "start": 1137.43,
                    "end": 1137.77,
                    "word": " ago",
                    "probability": 0.81787109375
                },
                {
                    "start": 1137.77,
                    "end": 1138.05,
                    "word": ",",
                    "probability": 0.7177734375
                },
                {
                    "start": 1138.05,
                    "end": 1138.31,
                    "word": " because",
                    "probability": 0.72412109375
                },
                {
                    "start": 1138.31,
                    "end": 1138.81,
                    "word": " it's",
                    "probability": 0.947998046875
                },
                {
                    "start": 1138.81,
                    "end": 1139.49,
                    "word": " 2020",
                    "probability": 0.81884765625
                },
                {
                    "start": 1139.49,
                    "end": 1140.33,
                    "word": ",",
                    "probability": 0.5673828125
                },
                {
                    "start": 1140.33,
                    "end": 1140.85,
                    "word": " and",
                    "probability": 0.87744140625
                },
                {
                    "start": 1140.85,
                    "end": 1141.15,
                    "word": " so",
                    "probability": 0.89990234375
                },
                {
                    "start": 1141.15,
                    "end": 1141.63,
                    "word": " time",
                    "probability": 0.74658203125
                },
                {
                    "start": 1141.63,
                    "end": 1142.33,
                    "word": " doesn't",
                    "probability": 0.932373046875
                },
                {
                    "start": 1142.33,
                    "end": 1143.19,
                    "word": " matter",
                    "probability": 0.818359375
                },
                {
                    "start": 1143.19,
                    "end": 1143.67,
                    "word": " anymore",
                    "probability": 0.7783203125
                },
                {
                    "start": 1143.67,
                    "end": 1144.47,
                    "word": ",",
                    "probability": 0.75634765625
                }
            ]
        },
        {
            "id": 161,
            "seek": 114856,
            "start": 1145.97,
            "end": 1152.29,
            "text": " we would return these detailed error messages for SQL queries, and it made it really, really easy",
            "tokens": [
                51264,
                321,
                576,
                2736,
                613,
                9942,
                6713,
                7897,
                337,
                19200,
                24109,
                11,
                293,
                309,
                1027,
                309,
                534,
                11,
                534,
                1858,
                51556
            ],
            "temperature": 0.0,
            "avg_logprob": -0.22656250044481077,
            "compression_ratio": 1.4262295081967213,
            "no_speech_prob": 0.0014324188232421875,
            "words": [
                {
                    "start": 1145.97,
                    "end": 1146.65,
                    "word": " we",
                    "probability": 0.8876953125
                },
                {
                    "start": 1146.65,
                    "end": 1147.03,
                    "word": " would",
                    "probability": 0.85791015625
                },
                {
                    "start": 1147.03,
                    "end": 1147.51,
                    "word": " return",
                    "probability": 0.78369140625
                },
                {
                    "start": 1147.51,
                    "end": 1147.71,
                    "word": " these",
                    "probability": 0.7001953125
                },
                {
                    "start": 1147.71,
                    "end": 1148.05,
                    "word": " detailed",
                    "probability": 0.6220703125
                },
                {
                    "start": 1148.05,
                    "end": 1148.35,
                    "word": " error",
                    "probability": 0.81689453125
                },
                {
                    "start": 1148.35,
                    "end": 1148.73,
                    "word": " messages",
                    "probability": 0.8330078125
                },
                {
                    "start": 1148.73,
                    "end": 1149.05,
                    "word": " for",
                    "probability": 0.8974609375
                },
                {
                    "start": 1149.05,
                    "end": 1149.33,
                    "word": " SQL",
                    "probability": 0.8603515625
                },
                {
                    "start": 1149.33,
                    "end": 1149.77,
                    "word": " queries",
                    "probability": 0.88232421875
                },
                {
                    "start": 1149.77,
                    "end": 1150.11,
                    "word": ",",
                    "probability": 0.440673828125
                },
                {
                    "start": 1150.11,
                    "end": 1150.85,
                    "word": " and",
                    "probability": 0.896484375
                },
                {
                    "start": 1150.85,
                    "end": 1151.17,
                    "word": " it",
                    "probability": 0.9013671875
                },
                {
                    "start": 1151.17,
                    "end": 1151.41,
                    "word": " made",
                    "probability": 0.810546875
                },
                {
                    "start": 1151.41,
                    "end": 1151.55,
                    "word": " it",
                    "probability": 0.90576171875
                },
                {
                    "start": 1151.55,
                    "end": 1151.75,
                    "word": " really",
                    "probability": 0.779296875
                },
                {
                    "start": 1151.75,
                    "end": 1151.91,
                    "word": ",",
                    "probability": 0.92724609375
                },
                {
                    "start": 1151.91,
                    "end": 1151.97,
                    "word": " really",
                    "probability": 0.78076171875
                },
                {
                    "start": 1151.97,
                    "end": 1152.29,
                    "word": " easy",
                    "probability": 0.80419921875
                }
            ]
        },
        {
            "id": 162,
            "seek": 117668,
            "start": 1152.29,
            "end": 1159.71,
            "text": " for attackers to find the information they were looking for in SQL databases, especially SQL",
            "tokens": [
                50364,
                337,
                45129,
                281,
                915,
                264,
                1589,
                436,
                645,
                1237,
                337,
                294,
                19200,
                22380,
                11,
                2318,
                19200,
                50732
            ],
            "temperature": 0.0,
            "avg_logprob": -0.16529143794199055,
            "compression_ratio": 1.617391304347826,
            "no_speech_prob": 0.00582122802734375,
            "words": [
                {
                    "start": 1152.29,
                    "end": 1152.49,
                    "word": " for",
                    "probability": 0.8486328125
                },
                {
                    "start": 1152.49,
                    "end": 1152.89,
                    "word": " attackers",
                    "probability": 0.79345703125
                },
                {
                    "start": 1152.89,
                    "end": 1153.49,
                    "word": " to",
                    "probability": 0.90185546875
                },
                {
                    "start": 1153.49,
                    "end": 1154.77,
                    "word": " find",
                    "probability": 0.77783203125
                },
                {
                    "start": 1154.77,
                    "end": 1155.03,
                    "word": " the",
                    "probability": 0.80859375
                },
                {
                    "start": 1155.03,
                    "end": 1155.51,
                    "word": " information",
                    "probability": 0.74072265625
                },
                {
                    "start": 1155.51,
                    "end": 1155.71,
                    "word": " they",
                    "probability": 0.7216796875
                },
                {
                    "start": 1155.71,
                    "end": 1155.83,
                    "word": " were",
                    "probability": 0.88623046875
                },
                {
                    "start": 1155.83,
                    "end": 1156.11,
                    "word": " looking",
                    "probability": 0.806640625
                },
                {
                    "start": 1156.11,
                    "end": 1156.51,
                    "word": " for",
                    "probability": 0.89453125
                },
                {
                    "start": 1156.51,
                    "end": 1157.43,
                    "word": " in",
                    "probability": 0.88134765625
                },
                {
                    "start": 1157.43,
                    "end": 1157.81,
                    "word": " SQL",
                    "probability": 0.853515625
                },
                {
                    "start": 1157.81,
                    "end": 1158.41,
                    "word": " databases",
                    "probability": 0.89990234375
                },
                {
                    "start": 1158.41,
                    "end": 1158.87,
                    "word": ",",
                    "probability": 0.8330078125
                },
                {
                    "start": 1158.87,
                    "end": 1159.29,
                    "word": " especially",
                    "probability": 0.72119140625
                },
                {
                    "start": 1159.29,
                    "end": 1159.71,
                    "word": " SQL",
                    "probability": 0.9453125
                }
            ]
        },
        {
            "id": 163,
            "seek": 117668,
            "start": 1159.71,
            "end": 1168.23,
            "text": " databases that didn't do a great job of input validation. And so, you know, once we started",
            "tokens": [
                50732,
                22380,
                300,
                994,
                380,
                360,
                257,
                869,
                1691,
                295,
                4846,
                24071,
                13,
                400,
                370,
                11,
                291,
                458,
                11,
                1564,
                321,
                1409,
                51160
            ],
            "temperature": 0.0,
            "avg_logprob": -0.16529143794199055,
            "compression_ratio": 1.617391304347826,
            "no_speech_prob": 0.00582122802734375,
            "words": [
                {
                    "start": 1159.71,
                    "end": 1160.21,
                    "word": " databases",
                    "probability": 0.88671875
                },
                {
                    "start": 1160.21,
                    "end": 1160.63,
                    "word": " that",
                    "probability": 0.861328125
                },
                {
                    "start": 1160.63,
                    "end": 1160.99,
                    "word": " didn't",
                    "probability": 0.92919921875
                },
                {
                    "start": 1160.99,
                    "end": 1161.17,
                    "word": " do",
                    "probability": 0.908203125
                },
                {
                    "start": 1161.29,
                    "end": 1161.51,
                    "word": " a great",
                    "probability": 0.86669921875
                },
                {
                    "start": 1161.51,
                    "end": 1161.95,
                    "word": " job",
                    "probability": 0.90478515625
                },
                {
                    "start": 1161.95,
                    "end": 1162.51,
                    "word": " of",
                    "probability": 0.90576171875
                },
                {
                    "start": 1162.51,
                    "end": 1163.49,
                    "word": " input",
                    "probability": 0.8349609375
                },
                {
                    "start": 1163.49,
                    "end": 1164.03,
                    "word": " validation",
                    "probability": 0.8515625
                },
                {
                    "start": 1164.03,
                    "end": 1164.59,
                    "word": ".",
                    "probability": 0.90771484375
                },
                {
                    "start": 1165.49,
                    "end": 1165.65,
                    "word": " And",
                    "probability": 0.7705078125
                },
                {
                    "start": 1165.65,
                    "end": 1166.05,
                    "word": " so",
                    "probability": 0.87353515625
                },
                {
                    "start": 1166.05,
                    "end": 1166.97,
                    "word": ",",
                    "probability": 0.724609375
                },
                {
                    "start": 1166.97,
                    "end": 1167.23,
                    "word": " you",
                    "probability": 0.80078125
                },
                {
                    "start": 1167.23,
                    "end": 1167.37,
                    "word": " know",
                    "probability": 0.81640625
                },
                {
                    "start": 1167.37,
                    "end": 1167.49,
                    "word": ",",
                    "probability": 0.9892578125
                },
                {
                    "start": 1167.49,
                    "end": 1167.81,
                    "word": " once",
                    "probability": 0.83544921875
                },
                {
                    "start": 1167.81,
                    "end": 1167.95,
                    "word": " we",
                    "probability": 0.90380859375
                },
                {
                    "start": 1167.95,
                    "end": 1168.23,
                    "word": " started",
                    "probability": 0.80029296875
                }
            ]
        },
        {
            "id": 164,
            "seek": 117668,
            "start": 1168.23,
            "end": 1174.99,
            "text": " returning basically there was an error as the error message, it became much harder, right? And",
            "tokens": [
                51160,
                12678,
                1936,
                456,
                390,
                364,
                6713,
                382,
                264,
                6713,
                3636,
                11,
                309,
                3062,
                709,
                6081,
                11,
                558,
                30,
                400,
                51496
            ],
            "temperature": 0.0,
            "avg_logprob": -0.16529143794199055,
            "compression_ratio": 1.617391304347826,
            "no_speech_prob": 0.00582122802734375,
            "words": [
                {
                    "start": 1168.23,
                    "end": 1168.67,
                    "word": " returning",
                    "probability": 0.83447265625
                },
                {
                    "start": 1168.67,
                    "end": 1169.35,
                    "word": " basically",
                    "probability": 0.4384765625
                },
                {
                    "start": 1169.35,
                    "end": 1170.13,
                    "word": " there",
                    "probability": 0.7578125
                },
                {
                    "start": 1170.13,
                    "end": 1170.33,
                    "word": " was",
                    "probability": 0.89404296875
                },
                {
                    "start": 1170.33,
                    "end": 1170.49,
                    "word": " an",
                    "probability": 0.908203125
                },
                {
                    "start": 1170.49,
                    "end": 1170.79,
                    "word": " error",
                    "probability": 0.82421875
                },
                {
                    "start": 1170.79,
                    "end": 1172.19,
                    "word": " as",
                    "probability": 0.64404296875
                },
                {
                    "start": 1172.19,
                    "end": 1172.33,
                    "word": " the",
                    "probability": 0.8193359375
                },
                {
                    "start": 1172.33,
                    "end": 1172.57,
                    "word": " error",
                    "probability": 0.83837890625
                },
                {
                    "start": 1172.57,
                    "end": 1173.07,
                    "word": " message",
                    "probability": 0.83056640625
                },
                {
                    "start": 1173.07,
                    "end": 1173.29,
                    "word": ",",
                    "probability": 0.5380859375
                },
                {
                    "start": 1173.29,
                    "end": 1173.47,
                    "word": " it",
                    "probability": 0.8994140625
                },
                {
                    "start": 1173.47,
                    "end": 1173.71,
                    "word": " became",
                    "probability": 0.78125
                },
                {
                    "start": 1173.71,
                    "end": 1173.99,
                    "word": " much",
                    "probability": 0.8017578125
                },
                {
                    "start": 1173.99,
                    "end": 1174.43,
                    "word": " harder",
                    "probability": 0.76171875
                },
                {
                    "start": 1174.43,
                    "end": 1174.71,
                    "word": ",",
                    "probability": 0.607421875
                },
                {
                    "start": 1174.71,
                    "end": 1174.87,
                    "word": " right",
                    "probability": 0.8564453125
                },
                {
                    "start": 1174.87,
                    "end": 1174.91,
                    "word": "?",
                    "probability": 0.8046875
                },
                {
                    "start": 1174.91,
                    "end": 1174.99,
                    "word": " And",
                    "probability": 0.82275390625
                }
            ]
        },
        {
            "id": 165,
            "seek": 117668,
            "start": 1174.99,
            "end": 1180.41,
            "text": " it's still feasible to do, but it became much harder. And it's similar here. If all that was",
            "tokens": [
                51496,
                309,
                311,
                920,
                26648,
                281,
                360,
                11,
                457,
                309,
                3062,
                709,
                6081,
                13,
                400,
                309,
                311,
                2531,
                510,
                13,
                759,
                439,
                300,
                390,
                51768
            ],
            "temperature": 0.0,
            "avg_logprob": -0.16529143794199055,
            "compression_ratio": 1.617391304347826,
            "no_speech_prob": 0.00582122802734375,
            "words": [
                {
                    "start": 1174.99,
                    "end": 1175.15,
                    "word": " it's",
                    "probability": 0.944580078125
                },
                {
                    "start": 1175.15,
                    "end": 1175.33,
                    "word": " still",
                    "probability": 0.904296875
                },
                {
                    "start": 1175.33,
                    "end": 1175.87,
                    "word": " feasible",
                    "probability": 0.80517578125
                },
                {
                    "start": 1175.87,
                    "end": 1176.19,
                    "word": " to",
                    "probability": 0.900390625
                },
                {
                    "start": 1176.19,
                    "end": 1176.37,
                    "word": " do",
                    "probability": 0.912109375
                },
                {
                    "start": 1176.37,
                    "end": 1176.45,
                    "word": ",",
                    "probability": 0.712890625
                },
                {
                    "start": 1176.45,
                    "end": 1176.57,
                    "word": " but",
                    "probability": 0.90087890625
                },
                {
                    "start": 1176.57,
                    "end": 1176.65,
                    "word": " it",
                    "probability": 0.8974609375
                },
                {
                    "start": 1176.65,
                    "end": 1176.95,
                    "word": " became",
                    "probability": 0.78125
                },
                {
                    "start": 1176.95,
                    "end": 1177.19,
                    "word": " much",
                    "probability": 0.80224609375
                },
                {
                    "start": 1177.19,
                    "end": 1177.47,
                    "word": " harder",
                    "probability": 0.76171875
                },
                {
                    "start": 1177.47,
                    "end": 1178.03,
                    "word": ".",
                    "probability": 0.87939453125
                },
                {
                    "start": 1178.21,
                    "end": 1178.43,
                    "word": " And",
                    "probability": 0.89013671875
                },
                {
                    "start": 1178.43,
                    "end": 1178.61,
                    "word": " it's",
                    "probability": 0.94970703125
                },
                {
                    "start": 1178.61,
                    "end": 1178.87,
                    "word": " similar",
                    "probability": 0.92626953125
                },
                {
                    "start": 1178.87,
                    "end": 1179.31,
                    "word": " here",
                    "probability": 0.81396484375
                },
                {
                    "start": 1179.31,
                    "end": 1179.47,
                    "word": ".",
                    "probability": 0.927734375
                },
                {
                    "start": 1179.47,
                    "end": 1179.93,
                    "word": " If",
                    "probability": 0.90185546875
                },
                {
                    "start": 1179.93,
                    "end": 1180.09,
                    "word": " all",
                    "probability": 0.8974609375
                },
                {
                    "start": 1180.09,
                    "end": 1180.23,
                    "word": " that",
                    "probability": 0.8837890625
                },
                {
                    "start": 1180.23,
                    "end": 1180.41,
                    "word": " was",
                    "probability": 0.89306640625
                }
            ]
        },
        {
            "id": 166,
            "seek": 120442,
            "start": 1180.41,
            "end": 1187.95,
            "text": " returned was a blocked or not blocked, that binary signal versus a detailed header,",
            "tokens": [
                50364,
                8752,
                390,
                257,
                15470,
                420,
                406,
                15470,
                11,
                300,
                17434,
                6358,
                5717,
                257,
                9942,
                23117,
                11,
                50696
            ],
            "temperature": 0.0,
            "avg_logprob": -0.1958575593524201,
            "compression_ratio": 1.7028301886792452,
            "no_speech_prob": 0.002471923828125,
            "words": [
                {
                    "start": 1180.41,
                    "end": 1180.99,
                    "word": " returned",
                    "probability": 0.501953125
                },
                {
                    "start": 1180.99,
                    "end": 1181.71,
                    "word": " was",
                    "probability": 0.80615234375
                },
                {
                    "start": 1182.17,
                    "end": 1182.83,
                    "word": " a blocked",
                    "probability": 0.61181640625
                },
                {
                    "start": 1182.83,
                    "end": 1183.03,
                    "word": " or",
                    "probability": 0.88134765625
                },
                {
                    "start": 1183.03,
                    "end": 1183.29,
                    "word": " not",
                    "probability": 0.89111328125
                },
                {
                    "start": 1183.29,
                    "end": 1183.83,
                    "word": " blocked",
                    "probability": 0.61328125
                },
                {
                    "start": 1183.83,
                    "end": 1183.99,
                    "word": ",",
                    "probability": 0.25927734375
                },
                {
                    "start": 1183.99,
                    "end": 1184.21,
                    "word": " that",
                    "probability": 0.7958984375
                },
                {
                    "start": 1184.21,
                    "end": 1184.69,
                    "word": " binary",
                    "probability": 0.7353515625
                },
                {
                    "start": 1184.69,
                    "end": 1185.25,
                    "word": " signal",
                    "probability": 0.8232421875
                },
                {
                    "start": 1185.25,
                    "end": 1185.65,
                    "word": " versus",
                    "probability": 0.52099609375
                },
                {
                    "start": 1185.99,
                    "end": 1186.57,
                    "word": " a detailed",
                    "probability": 0.720703125
                },
                {
                    "start": 1186.57,
                    "end": 1187.09,
                    "word": " header",
                    "probability": 0.79833984375
                },
                {
                    "start": 1187.09,
                    "end": 1187.95,
                    "word": ",",
                    "probability": 0.791015625
                }
            ]
        },
        {
            "id": 167,
            "seek": 120442,
            "start": 1188.51,
            "end": 1194.93,
            "text": " then it would be much more difficult to create a copycat model. And so, you know, that is a",
            "tokens": [
                50772,
                550,
                309,
                576,
                312,
                709,
                544,
                2252,
                281,
                1884,
                257,
                5055,
                18035,
                2316,
                13,
                400,
                370,
                11,
                291,
                458,
                11,
                300,
                307,
                257,
                51076
            ],
            "temperature": 0.0,
            "avg_logprob": -0.1958575593524201,
            "compression_ratio": 1.7028301886792452,
            "no_speech_prob": 0.002471923828125,
            "words": [
                {
                    "start": 1188.51,
                    "end": 1188.85,
                    "word": " then",
                    "probability": 0.74609375
                },
                {
                    "start": 1188.85,
                    "end": 1188.97,
                    "word": " it",
                    "probability": 0.90234375
                },
                {
                    "start": 1188.97,
                    "end": 1189.09,
                    "word": " would",
                    "probability": 0.849609375
                },
                {
                    "start": 1189.09,
                    "end": 1189.45,
                    "word": " be",
                    "probability": 0.9091796875
                },
                {
                    "start": 1189.45,
                    "end": 1190.25,
                    "word": " much",
                    "probability": 0.79541015625
                },
                {
                    "start": 1190.25,
                    "end": 1190.47,
                    "word": " more",
                    "probability": 0.88720703125
                },
                {
                    "start": 1190.47,
                    "end": 1190.81,
                    "word": " difficult",
                    "probability": 0.8466796875
                },
                {
                    "start": 1190.81,
                    "end": 1191.07,
                    "word": " to",
                    "probability": 0.90625
                },
                {
                    "start": 1191.07,
                    "end": 1191.29,
                    "word": " create",
                    "probability": 0.82861328125
                },
                {
                    "start": 1191.45,
                    "end": 1191.95,
                    "word": " a copycat",
                    "probability": 0.855224609375
                },
                {
                    "start": 1191.95,
                    "end": 1192.33,
                    "word": " model",
                    "probability": 0.8232421875
                },
                {
                    "start": 1192.33,
                    "end": 1192.89,
                    "word": ".",
                    "probability": 0.87646484375
                },
                {
                    "start": 1192.89,
                    "end": 1192.97,
                    "word": " And",
                    "probability": 0.7353515625
                },
                {
                    "start": 1192.97,
                    "end": 1193.33,
                    "word": " so",
                    "probability": 0.8818359375
                },
                {
                    "start": 1193.33,
                    "end": 1193.53,
                    "word": ",",
                    "probability": 0.599609375
                },
                {
                    "start": 1193.53,
                    "end": 1193.71,
                    "word": " you",
                    "probability": 0.68994140625
                },
                {
                    "start": 1193.71,
                    "end": 1193.83,
                    "word": " know",
                    "probability": 0.82080078125
                },
                {
                    "start": 1193.83,
                    "end": 1193.91,
                    "word": ",",
                    "probability": 0.98046875
                },
                {
                    "start": 1193.91,
                    "end": 1194.17,
                    "word": " that",
                    "probability": 0.88134765625
                },
                {
                    "start": 1194.17,
                    "end": 1194.37,
                    "word": " is",
                    "probability": 0.90576171875
                },
                {
                    "start": 1194.63,
                    "end": 1194.93,
                    "word": " a real",
                    "probability": 0.8857421875
                }
            ]
        },
        {
            "id": 168,
            "seek": 120442,
            "start": 1194.93,
            "end": 1202.29,
            "text": " real mitigation. And when we think about model theft, kind of the next step is inversion. Model",
            "tokens": [
                51076,
                957,
                32649,
                13,
                400,
                562,
                321,
                519,
                466,
                2316,
                28508,
                11,
                733,
                295,
                264,
                958,
                1823,
                307,
                43576,
                13,
                17105,
                51440
            ],
            "temperature": 0.0,
            "avg_logprob": -0.1958575593524201,
            "compression_ratio": 1.7028301886792452,
            "no_speech_prob": 0.002471923828125,
            "words": [
                {
                    "start": 1194.93,
                    "end": 1195.55,
                    "word": " mitigation",
                    "probability": 0.8291015625
                },
                {
                    "start": 1195.55,
                    "end": 1196.23,
                    "word": ".",
                    "probability": 0.95947265625
                },
                {
                    "start": 1196.47,
                    "end": 1196.53,
                    "word": " And",
                    "probability": 0.884765625
                },
                {
                    "start": 1196.53,
                    "end": 1196.69,
                    "word": " when",
                    "probability": 0.85791015625
                },
                {
                    "start": 1196.69,
                    "end": 1196.85,
                    "word": " we",
                    "probability": 0.90869140625
                },
                {
                    "start": 1196.85,
                    "end": 1197.03,
                    "word": " think",
                    "probability": 0.80810546875
                },
                {
                    "start": 1197.03,
                    "end": 1197.33,
                    "word": " about",
                    "probability": 0.818359375
                },
                {
                    "start": 1197.33,
                    "end": 1197.65,
                    "word": " model",
                    "probability": 0.81298828125
                },
                {
                    "start": 1197.65,
                    "end": 1198.09,
                    "word": " theft",
                    "probability": 0.84130859375
                },
                {
                    "start": 1198.09,
                    "end": 1198.21,
                    "word": ",",
                    "probability": 0.9072265625
                },
                {
                    "start": 1198.21,
                    "end": 1198.35,
                    "word": " kind",
                    "probability": 0.77587890625
                },
                {
                    "start": 1198.35,
                    "end": 1198.41,
                    "word": " of",
                    "probability": 0.90673828125
                },
                {
                    "start": 1198.41,
                    "end": 1198.57,
                    "word": " the",
                    "probability": 0.80419921875
                },
                {
                    "start": 1198.57,
                    "end": 1198.97,
                    "word": " next",
                    "probability": 0.8623046875
                },
                {
                    "start": 1198.97,
                    "end": 1199.79,
                    "word": " step",
                    "probability": 0.818359375
                },
                {
                    "start": 1199.79,
                    "end": 1200.55,
                    "word": " is",
                    "probability": 0.873046875
                },
                {
                    "start": 1200.55,
                    "end": 1201.19,
                    "word": " inversion",
                    "probability": 0.63818359375
                },
                {
                    "start": 1201.19,
                    "end": 1201.67,
                    "word": ".",
                    "probability": 0.2451171875
                },
                {
                    "start": 1201.67,
                    "end": 1201.87,
                    "word": " Model",
                    "probability": 0.76953125
                },
                {
                    "start": 1201.87,
                    "end": 1202.29,
                    "word": " inversion",
                    "probability": 0.642578125
                }
            ]
        },
        {
            "id": 169,
            "seek": 120442,
            "start": 1202.29,
            "end": 1208.15,
            "text": " inversion. And what model inversion is, is recovering training data from a trained model.",
            "tokens": [
                51440,
                43576,
                13,
                400,
                437,
                2316,
                43576,
                307,
                11,
                307,
                29180,
                3097,
                1412,
                490,
                257,
                8895,
                2316,
                13,
                51736
            ],
            "temperature": 0.0,
            "avg_logprob": -0.1958575593524201,
            "compression_ratio": 1.7028301886792452,
            "no_speech_prob": 0.002471923828125,
            "words": [
                {
                    "start": 1202.29,
                    "end": 1202.97,
                    "word": ".",
                    "probability": 0.9609375
                },
                {
                    "start": 1203.27,
                    "end": 1203.39,
                    "word": " And",
                    "probability": 0.84912109375
                },
                {
                    "start": 1203.39,
                    "end": 1203.53,
                    "word": " what",
                    "probability": 0.87890625
                },
                {
                    "start": 1203.53,
                    "end": 1203.81,
                    "word": " model",
                    "probability": 0.84228515625
                },
                {
                    "start": 1203.81,
                    "end": 1204.23,
                    "word": " inversion",
                    "probability": 0.67431640625
                },
                {
                    "start": 1204.23,
                    "end": 1204.63,
                    "word": " is",
                    "probability": 0.90576171875
                },
                {
                    "start": 1204.63,
                    "end": 1204.65,
                    "word": ",",
                    "probability": 0.56494140625
                },
                {
                    "start": 1204.65,
                    "end": 1204.85,
                    "word": " is",
                    "probability": 0.869140625
                },
                {
                    "start": 1204.85,
                    "end": 1205.49,
                    "word": " recovering",
                    "probability": 0.759765625
                },
                {
                    "start": 1205.49,
                    "end": 1206.03,
                    "word": " training",
                    "probability": 0.8076171875
                },
                {
                    "start": 1206.03,
                    "end": 1206.39,
                    "word": " data",
                    "probability": 0.82373046875
                },
                {
                    "start": 1206.39,
                    "end": 1206.85,
                    "word": " from",
                    "probability": 0.8212890625
                },
                {
                    "start": 1207.05,
                    "end": 1207.41,
                    "word": " a trained",
                    "probability": 0.81591796875
                },
                {
                    "start": 1207.41,
                    "end": 1207.81,
                    "word": " model",
                    "probability": 0.8232421875
                },
                {
                    "start": 1207.81,
                    "end": 1208.15,
                    "word": ".",
                    "probability": 0.99169921875
                }
            ]
        },
        {
            "id": 170,
            "seek": 122748,
            "start": 1208.15,
            "end": 1216.15,
            "text": " So, there was a paper by Nick Carlini that came out two years ago, I think, in 2018,",
            "tokens": [
                50380,
                407,
                11,
                456,
                390,
                257,
                3035,
                538,
                9449,
                14256,
                3812,
                300,
                1361,
                484,
                732,
                924,
                2057,
                11,
                286,
                519,
                11,
                294,
                6096,
                11,
                50752
            ],
            "temperature": 0.0,
            "avg_logprob": -0.19127155429330364,
            "compression_ratio": 1.3837837837837839,
            "no_speech_prob": 0.0001055598258972168,
            "words": [
                {
                    "start": 1208.15,
                    "end": 1208.71,
                    "word": " So",
                    "probability": 0.50830078125
                },
                {
                    "start": 1208.71,
                    "end": 1208.89,
                    "word": ",",
                    "probability": 0.261962890625
                },
                {
                    "start": 1208.89,
                    "end": 1209.01,
                    "word": " there",
                    "probability": 0.84521484375
                },
                {
                    "start": 1209.01,
                    "end": 1209.19,
                    "word": " was",
                    "probability": 0.88671875
                },
                {
                    "start": 1209.57,
                    "end": 1209.99,
                    "word": " a paper",
                    "probability": 0.80615234375
                },
                {
                    "start": 1209.99,
                    "end": 1210.39,
                    "word": " by",
                    "probability": 0.87451171875
                },
                {
                    "start": 1210.39,
                    "end": 1210.71,
                    "word": " Nick",
                    "probability": 0.82373046875
                },
                {
                    "start": 1210.71,
                    "end": 1211.27,
                    "word": " Carlini",
                    "probability": 0.862060546875
                },
                {
                    "start": 1211.27,
                    "end": 1211.55,
                    "word": " that",
                    "probability": 0.8388671875
                },
                {
                    "start": 1211.55,
                    "end": 1211.87,
                    "word": " came",
                    "probability": 0.81298828125
                },
                {
                    "start": 1211.87,
                    "end": 1212.35,
                    "word": " out",
                    "probability": 0.81494140625
                },
                {
                    "start": 1212.35,
                    "end": 1213.63,
                    "word": " two",
                    "probability": 0.58203125
                },
                {
                    "start": 1213.63,
                    "end": 1213.93,
                    "word": " years",
                    "probability": 0.880859375
                },
                {
                    "start": 1213.93,
                    "end": 1214.19,
                    "word": " ago",
                    "probability": 0.8232421875
                },
                {
                    "start": 1214.19,
                    "end": 1214.31,
                    "word": ",",
                    "probability": 0.75244140625
                },
                {
                    "start": 1214.31,
                    "end": 1214.43,
                    "word": " I",
                    "probability": 0.9755859375
                },
                {
                    "start": 1214.43,
                    "end": 1214.61,
                    "word": " think",
                    "probability": 0.806640625
                },
                {
                    "start": 1214.61,
                    "end": 1215.11,
                    "word": ",",
                    "probability": 0.771484375
                },
                {
                    "start": 1215.11,
                    "end": 1215.27,
                    "word": " in",
                    "probability": 0.89599609375
                },
                {
                    "start": 1215.27,
                    "end": 1215.83,
                    "word": " 2018",
                    "probability": 0.72314453125
                },
                {
                    "start": 1215.83,
                    "end": 1216.15,
                    "word": ",",
                    "probability": 0.603515625
                }
            ]
        },
        {
            "id": 171,
            "seek": 122748,
            "start": 1216.15,
            "end": 1222.33,
            "text": " about how neural networks unintentionally memorize specific training examples.",
            "tokens": [
                50784,
                466,
                577,
                18161,
                9590,
                45514,
                379,
                27478,
                2685,
                3097,
                5110,
                13,
                51036
            ],
            "temperature": 0.0,
            "avg_logprob": -0.19127155429330364,
            "compression_ratio": 1.3837837837837839,
            "no_speech_prob": 0.0001055598258972168,
            "words": [
                {
                    "start": 1216.15,
                    "end": 1217.01,
                    "word": " about",
                    "probability": 0.82421875
                },
                {
                    "start": 1217.01,
                    "end": 1217.69,
                    "word": " how",
                    "probability": 0.89990234375
                },
                {
                    "start": 1217.69,
                    "end": 1218.11,
                    "word": " neural",
                    "probability": 0.90283203125
                },
                {
                    "start": 1218.11,
                    "end": 1218.57,
                    "word": " networks",
                    "probability": 0.68310546875
                },
                {
                    "start": 1218.57,
                    "end": 1219.37,
                    "word": " unintentionally",
                    "probability": 0.88330078125
                },
                {
                    "start": 1219.37,
                    "end": 1220.01,
                    "word": " memorize",
                    "probability": 0.62451171875
                },
                {
                    "start": 1220.01,
                    "end": 1220.57,
                    "word": " specific",
                    "probability": 0.80078125
                },
                {
                    "start": 1220.57,
                    "end": 1220.97,
                    "word": " training",
                    "probability": 0.82861328125
                },
                {
                    "start": 1220.97,
                    "end": 1221.61,
                    "word": " examples",
                    "probability": 0.78564453125
                },
                {
                    "start": 1221.61,
                    "end": 1222.33,
                    "word": ".",
                    "probability": 0.94580078125
                }
            ]
        },
        {
            "id": 172,
            "seek": 122748,
            "start": 1222.33,
            "end": 1231.21,
            "text": " There have been papers on model inversion, both traditional and using generative adversarial",
            "tokens": [
                51080,
                821,
                362,
                668,
                10577,
                322,
                2316,
                43576,
                11,
                1293,
                5164,
                293,
                1228,
                1337,
                1166,
                17641,
                44745,
                51516
            ],
            "temperature": 0.0,
            "avg_logprob": -0.19127155429330364,
            "compression_ratio": 1.3837837837837839,
            "no_speech_prob": 0.0001055598258972168,
            "words": [
                {
                    "start": 1222.33,
                    "end": 1222.77,
                    "word": " There",
                    "probability": 0.806640625
                },
                {
                    "start": 1222.77,
                    "end": 1223.03,
                    "word": " have",
                    "probability": 0.87646484375
                },
                {
                    "start": 1223.03,
                    "end": 1223.43,
                    "word": " been",
                    "probability": 0.81982421875
                },
                {
                    "start": 1223.43,
                    "end": 1223.99,
                    "word": " papers",
                    "probability": 0.81884765625
                },
                {
                    "start": 1223.99,
                    "end": 1224.59,
                    "word": " on",
                    "probability": 0.9130859375
                },
                {
                    "start": 1224.59,
                    "end": 1225.11,
                    "word": " model",
                    "probability": 0.80029296875
                },
                {
                    "start": 1225.11,
                    "end": 1225.71,
                    "word": " inversion",
                    "probability": 0.64990234375
                },
                {
                    "start": 1225.71,
                    "end": 1226.09,
                    "word": ",",
                    "probability": 0.80029296875
                },
                {
                    "start": 1226.09,
                    "end": 1226.45,
                    "word": " both",
                    "probability": 0.81494140625
                },
                {
                    "start": 1226.45,
                    "end": 1227.63,
                    "word": " traditional",
                    "probability": 0.75341796875
                },
                {
                    "start": 1227.63,
                    "end": 1228.91,
                    "word": " and",
                    "probability": 0.849609375
                },
                {
                    "start": 1228.91,
                    "end": 1229.59,
                    "word": " using",
                    "probability": 0.7998046875
                },
                {
                    "start": 1229.59,
                    "end": 1230.31,
                    "word": " generative",
                    "probability": 0.90087890625
                },
                {
                    "start": 1230.31,
                    "end": 1231.21,
                    "word": " adversarial",
                    "probability": 0.927978515625
                }
            ]
        },
        {
            "id": 173,
            "seek": 125748,
            "start": 1231.21,
            "end": 1240.27,
            "text": " networks that have shown that it's pretty effective. And so, the impact here is the",
            "tokens": [
                50364,
                9590,
                300,
                362,
                4898,
                300,
                309,
                311,
                1238,
                4942,
                13,
                400,
                370,
                11,
                264,
                2712,
                510,
                307,
                264,
                50816
            ],
            "temperature": 0.0,
            "avg_logprob": -0.19714096109283732,
            "compression_ratio": 1.5874439461883407,
            "no_speech_prob": 0.0164031982421875,
            "words": [
                {
                    "start": 1231.21,
                    "end": 1231.83,
                    "word": " networks",
                    "probability": 0.72998046875
                },
                {
                    "start": 1231.83,
                    "end": 1232.91,
                    "word": " that",
                    "probability": 0.78125
                },
                {
                    "start": 1232.91,
                    "end": 1233.05,
                    "word": " have",
                    "probability": 0.87255859375
                },
                {
                    "start": 1233.05,
                    "end": 1233.35,
                    "word": " shown",
                    "probability": 0.87353515625
                },
                {
                    "start": 1233.35,
                    "end": 1233.55,
                    "word": " that",
                    "probability": 0.82763671875
                },
                {
                    "start": 1233.55,
                    "end": 1233.99,
                    "word": " it's",
                    "probability": 0.941650390625
                },
                {
                    "start": 1233.99,
                    "end": 1234.47,
                    "word": " pretty",
                    "probability": 0.68701171875
                },
                {
                    "start": 1234.47,
                    "end": 1234.99,
                    "word": " effective",
                    "probability": 0.744140625
                },
                {
                    "start": 1234.99,
                    "end": 1235.65,
                    "word": ".",
                    "probability": 0.931640625
                },
                {
                    "start": 1237.15,
                    "end": 1237.81,
                    "word": " And",
                    "probability": 0.74658203125
                },
                {
                    "start": 1237.81,
                    "end": 1237.93,
                    "word": " so",
                    "probability": 0.8818359375
                },
                {
                    "start": 1237.93,
                    "end": 1237.97,
                    "word": ",",
                    "probability": 0.494873046875
                },
                {
                    "start": 1237.97,
                    "end": 1238.09,
                    "word": " the",
                    "probability": 0.81298828125
                },
                {
                    "start": 1238.09,
                    "end": 1238.53,
                    "word": " impact",
                    "probability": 0.8798828125
                },
                {
                    "start": 1238.53,
                    "end": 1238.87,
                    "word": " here",
                    "probability": 0.8095703125
                },
                {
                    "start": 1238.87,
                    "end": 1239.51,
                    "word": " is",
                    "probability": 0.9013671875
                },
                {
                    "start": 1239.51,
                    "end": 1240.27,
                    "word": " the",
                    "probability": 0.80859375
                }
            ]
        },
        {
            "id": 174,
            "seek": 125748,
            "start": 1240.27,
            "end": 1246.23,
            "text": " loss of data, right? You're losing training data. And maybe that doesn't matter a ton.",
            "tokens": [
                50816,
                4470,
                295,
                1412,
                11,
                558,
                30,
                509,
                434,
                7027,
                3097,
                1412,
                13,
                400,
                1310,
                300,
                1177,
                380,
                1871,
                257,
                2952,
                13,
                51080
            ],
            "temperature": 0.0,
            "avg_logprob": -0.19714096109283732,
            "compression_ratio": 1.5874439461883407,
            "no_speech_prob": 0.0164031982421875,
            "words": [
                {
                    "start": 1240.27,
                    "end": 1240.61,
                    "word": " loss",
                    "probability": 0.8125
                },
                {
                    "start": 1240.61,
                    "end": 1240.79,
                    "word": " of",
                    "probability": 0.904296875
                },
                {
                    "start": 1240.79,
                    "end": 1241.11,
                    "word": " data",
                    "probability": 0.810546875
                },
                {
                    "start": 1241.11,
                    "end": 1241.37,
                    "word": ",",
                    "probability": 0.71533203125
                },
                {
                    "start": 1241.37,
                    "end": 1241.65,
                    "word": " right",
                    "probability": 0.86572265625
                },
                {
                    "start": 1241.65,
                    "end": 1241.71,
                    "word": "?",
                    "probability": 0.931640625
                },
                {
                    "start": 1241.71,
                    "end": 1241.99,
                    "word": " You're",
                    "probability": 0.91650390625
                },
                {
                    "start": 1241.99,
                    "end": 1242.47,
                    "word": " losing",
                    "probability": 0.81591796875
                },
                {
                    "start": 1242.47,
                    "end": 1242.91,
                    "word": " training",
                    "probability": 0.81689453125
                },
                {
                    "start": 1242.91,
                    "end": 1243.23,
                    "word": " data",
                    "probability": 0.83203125
                },
                {
                    "start": 1243.23,
                    "end": 1243.83,
                    "word": ".",
                    "probability": 0.8671875
                },
                {
                    "start": 1243.83,
                    "end": 1244.01,
                    "word": " And",
                    "probability": 0.876953125
                },
                {
                    "start": 1244.01,
                    "end": 1244.23,
                    "word": " maybe",
                    "probability": 0.86669921875
                },
                {
                    "start": 1244.23,
                    "end": 1244.43,
                    "word": " that",
                    "probability": 0.8837890625
                },
                {
                    "start": 1244.43,
                    "end": 1244.73,
                    "word": " doesn't",
                    "probability": 0.933837890625
                },
                {
                    "start": 1244.73,
                    "end": 1244.97,
                    "word": " matter",
                    "probability": 0.79833984375
                },
                {
                    "start": 1245.27,
                    "end": 1245.69,
                    "word": " a ton",
                    "probability": 0.8798828125
                },
                {
                    "start": 1245.69,
                    "end": 1246.23,
                    "word": ".",
                    "probability": 0.83642578125
                }
            ]
        },
        {
            "id": 175,
            "seek": 125748,
            "start": 1246.23,
            "end": 1253.77,
            "text": " So, like at Rapid7, we have our open dataset, right? And that's free and open for anyone to",
            "tokens": [
                51108,
                407,
                11,
                411,
                412,
                44580,
                22,
                11,
                321,
                362,
                527,
                1269,
                28872,
                11,
                558,
                30,
                400,
                300,
                311,
                1737,
                293,
                1269,
                337,
                2878,
                281,
                51492
            ],
            "temperature": 0.0,
            "avg_logprob": -0.19714096109283732,
            "compression_ratio": 1.5874439461883407,
            "no_speech_prob": 0.0164031982421875,
            "words": [
                {
                    "start": 1246.23,
                    "end": 1246.39,
                    "word": " So",
                    "probability": 0.857421875
                },
                {
                    "start": 1246.39,
                    "end": 1246.47,
                    "word": ",",
                    "probability": 0.95556640625
                },
                {
                    "start": 1246.47,
                    "end": 1246.65,
                    "word": " like",
                    "probability": 0.82568359375
                },
                {
                    "start": 1246.65,
                    "end": 1246.81,
                    "word": " at",
                    "probability": 0.47998046875
                },
                {
                    "start": 1246.81,
                    "end": 1247.49,
                    "word": " Rapid7",
                    "probability": 0.697998046875
                },
                {
                    "start": 1247.49,
                    "end": 1247.67,
                    "word": ",",
                    "probability": 0.751953125
                },
                {
                    "start": 1247.67,
                    "end": 1247.85,
                    "word": " we",
                    "probability": 0.89892578125
                },
                {
                    "start": 1247.85,
                    "end": 1248.23,
                    "word": " have",
                    "probability": 0.8828125
                },
                {
                    "start": 1248.23,
                    "end": 1248.63,
                    "word": " our",
                    "probability": 0.7939453125
                },
                {
                    "start": 1248.63,
                    "end": 1249.03,
                    "word": " open",
                    "probability": 0.7890625
                },
                {
                    "start": 1249.03,
                    "end": 1249.47,
                    "word": " dataset",
                    "probability": 0.4130859375
                },
                {
                    "start": 1249.47,
                    "end": 1249.83,
                    "word": ",",
                    "probability": 0.83349609375
                },
                {
                    "start": 1249.83,
                    "end": 1250.13,
                    "word": " right",
                    "probability": 0.8564453125
                },
                {
                    "start": 1250.13,
                    "end": 1250.47,
                    "word": "?",
                    "probability": 0.97314453125
                },
                {
                    "start": 1250.47,
                    "end": 1251.07,
                    "word": " And",
                    "probability": 0.87744140625
                },
                {
                    "start": 1251.07,
                    "end": 1251.41,
                    "word": " that's",
                    "probability": 0.93798828125
                },
                {
                    "start": 1251.41,
                    "end": 1251.99,
                    "word": " free",
                    "probability": 0.8720703125
                },
                {
                    "start": 1251.99,
                    "end": 1252.27,
                    "word": " and",
                    "probability": 0.89404296875
                },
                {
                    "start": 1252.27,
                    "end": 1252.61,
                    "word": " open",
                    "probability": 0.822265625
                },
                {
                    "start": 1252.61,
                    "end": 1252.91,
                    "word": " for",
                    "probability": 0.89501953125
                },
                {
                    "start": 1252.91,
                    "end": 1253.35,
                    "word": " anyone",
                    "probability": 0.904296875
                },
                {
                    "start": 1253.35,
                    "end": 1253.77,
                    "word": " to",
                    "probability": 0.9033203125
                }
            ]
        },
        {
            "id": 176,
            "seek": 125748,
            "start": 1253.77,
            "end": 1260.95,
            "text": " use for noncommercial purposes, and we make it available for, you know, other stuff, right?",
            "tokens": [
                51492,
                764,
                337,
                2107,
                1112,
                39260,
                9932,
                11,
                293,
                321,
                652,
                309,
                2435,
                337,
                11,
                291,
                458,
                11,
                661,
                1507,
                11,
                558,
                30,
                51816
            ],
            "temperature": 0.0,
            "avg_logprob": -0.19714096109283732,
            "compression_ratio": 1.5874439461883407,
            "no_speech_prob": 0.0164031982421875,
            "words": [
                {
                    "start": 1253.77,
                    "end": 1254.25,
                    "word": " use",
                    "probability": 0.826171875
                },
                {
                    "start": 1254.25,
                    "end": 1254.61,
                    "word": " for",
                    "probability": 0.87646484375
                },
                {
                    "start": 1254.61,
                    "end": 1255.49,
                    "word": " noncommercial",
                    "probability": 0.8343098958333334
                },
                {
                    "start": 1255.49,
                    "end": 1256.03,
                    "word": " purposes",
                    "probability": 0.85791015625
                },
                {
                    "start": 1256.03,
                    "end": 1256.31,
                    "word": ",",
                    "probability": 0.336181640625
                },
                {
                    "start": 1256.31,
                    "end": 1256.51,
                    "word": " and",
                    "probability": 0.89599609375
                },
                {
                    "start": 1256.51,
                    "end": 1256.67,
                    "word": " we",
                    "probability": 0.9033203125
                },
                {
                    "start": 1256.67,
                    "end": 1256.93,
                    "word": " make",
                    "probability": 0.87060546875
                },
                {
                    "start": 1256.93,
                    "end": 1257.21,
                    "word": " it",
                    "probability": 0.90625
                },
                {
                    "start": 1257.21,
                    "end": 1258.25,
                    "word": " available",
                    "probability": 0.73583984375
                },
                {
                    "start": 1258.25,
                    "end": 1258.63,
                    "word": " for",
                    "probability": 0.89794921875
                },
                {
                    "start": 1258.63,
                    "end": 1258.77,
                    "word": ",",
                    "probability": 0.8330078125
                },
                {
                    "start": 1258.77,
                    "end": 1259.17,
                    "word": " you",
                    "probability": 0.90087890625
                },
                {
                    "start": 1259.17,
                    "end": 1259.27,
                    "word": " know",
                    "probability": 0.81640625
                },
                {
                    "start": 1259.27,
                    "end": 1259.31,
                    "word": ",",
                    "probability": 0.9970703125
                },
                {
                    "start": 1259.31,
                    "end": 1259.73,
                    "word": " other",
                    "probability": 0.771484375
                },
                {
                    "start": 1259.73,
                    "end": 1260.03,
                    "word": " stuff",
                    "probability": 0.802734375
                },
                {
                    "start": 1260.03,
                    "end": 1260.09,
                    "word": ",",
                    "probability": 0.8740234375
                },
                {
                    "start": 1260.09,
                    "end": 1260.29,
                    "word": " right",
                    "probability": 0.880859375
                },
                {
                    "start": 1260.29,
                    "end": 1260.95,
                    "word": "?",
                    "probability": 0.98974609375
                }
            ]
        },
        {
            "id": 177,
            "seek": 128532,
            "start": 1261.67,
            "end": 1268.61,
            "text": " And so, like, if our model was trained on open data, and somebody inverted our model,",
            "tokens": [
                50404,
                400,
                370,
                11,
                411,
                11,
                498,
                527,
                2316,
                390,
                8895,
                322,
                1269,
                1412,
                11,
                293,
                2618,
                38969,
                527,
                2316,
                11,
                50704
            ],
            "temperature": 0.0,
            "avg_logprob": -0.187152784731653,
            "compression_ratio": 1.6948356807511737,
            "no_speech_prob": 0.00015592575073242188,
            "words": [
                {
                    "start": 1261.67,
                    "end": 1262.27,
                    "word": " And",
                    "probability": 0.45263671875
                },
                {
                    "start": 1262.27,
                    "end": 1262.43,
                    "word": " so",
                    "probability": 0.8759765625
                },
                {
                    "start": 1262.43,
                    "end": 1262.51,
                    "word": ",",
                    "probability": 0.251953125
                },
                {
                    "start": 1262.51,
                    "end": 1262.65,
                    "word": " like",
                    "probability": 0.72705078125
                },
                {
                    "start": 1262.65,
                    "end": 1262.77,
                    "word": ",",
                    "probability": 0.7568359375
                },
                {
                    "start": 1262.77,
                    "end": 1262.81,
                    "word": " if",
                    "probability": 0.89697265625
                },
                {
                    "start": 1262.81,
                    "end": 1263.49,
                    "word": " our",
                    "probability": 0.8193359375
                },
                {
                    "start": 1263.49,
                    "end": 1264.07,
                    "word": " model",
                    "probability": 0.81982421875
                },
                {
                    "start": 1264.07,
                    "end": 1264.31,
                    "word": " was",
                    "probability": 0.89892578125
                },
                {
                    "start": 1264.31,
                    "end": 1264.61,
                    "word": " trained",
                    "probability": 0.8154296875
                },
                {
                    "start": 1264.61,
                    "end": 1264.83,
                    "word": " on",
                    "probability": 0.908203125
                },
                {
                    "start": 1264.83,
                    "end": 1265.15,
                    "word": " open",
                    "probability": 0.7939453125
                },
                {
                    "start": 1265.15,
                    "end": 1265.49,
                    "word": " data",
                    "probability": 0.822265625
                },
                {
                    "start": 1265.49,
                    "end": 1265.99,
                    "word": ",",
                    "probability": 0.255126953125
                },
                {
                    "start": 1265.99,
                    "end": 1266.39,
                    "word": " and",
                    "probability": 0.89013671875
                },
                {
                    "start": 1266.39,
                    "end": 1266.85,
                    "word": " somebody",
                    "probability": 0.77685546875
                },
                {
                    "start": 1266.85,
                    "end": 1267.41,
                    "word": " inverted",
                    "probability": 0.73828125
                },
                {
                    "start": 1267.41,
                    "end": 1267.67,
                    "word": " our",
                    "probability": 0.8154296875
                },
                {
                    "start": 1267.67,
                    "end": 1268.09,
                    "word": " model",
                    "probability": 0.81982421875
                },
                {
                    "start": 1268.09,
                    "end": 1268.61,
                    "word": ",",
                    "probability": 0.9677734375
                }
            ]
        },
        {
            "id": 178,
            "seek": 128532,
            "start": 1268.61,
            "end": 1275.11,
            "text": " we probably wouldn't care that much, because, like, that data's out there, right? But if you",
            "tokens": [
                50736,
                321,
                1391,
                2759,
                380,
                1127,
                300,
                709,
                11,
                570,
                11,
                411,
                11,
                300,
                1412,
                311,
                484,
                456,
                11,
                558,
                30,
                583,
                498,
                291,
                51056
            ],
            "temperature": 0.0,
            "avg_logprob": -0.187152784731653,
            "compression_ratio": 1.6948356807511737,
            "no_speech_prob": 0.00015592575073242188,
            "words": [
                {
                    "start": 1268.61,
                    "end": 1268.93,
                    "word": " we",
                    "probability": 0.89599609375
                },
                {
                    "start": 1268.93,
                    "end": 1269.13,
                    "word": " probably",
                    "probability": 0.90185546875
                },
                {
                    "start": 1269.13,
                    "end": 1269.31,
                    "word": " wouldn't",
                    "probability": 0.903564453125
                },
                {
                    "start": 1269.31,
                    "end": 1269.49,
                    "word": " care",
                    "probability": 0.818359375
                },
                {
                    "start": 1269.49,
                    "end": 1269.67,
                    "word": " that",
                    "probability": 0.87646484375
                },
                {
                    "start": 1269.67,
                    "end": 1269.97,
                    "word": " much",
                    "probability": 0.8115234375
                },
                {
                    "start": 1269.97,
                    "end": 1270.15,
                    "word": ",",
                    "probability": 0.394287109375
                },
                {
                    "start": 1270.15,
                    "end": 1270.35,
                    "word": " because",
                    "probability": 0.708984375
                },
                {
                    "start": 1270.35,
                    "end": 1270.45,
                    "word": ",",
                    "probability": 0.77099609375
                },
                {
                    "start": 1270.45,
                    "end": 1270.59,
                    "word": " like",
                    "probability": 0.8798828125
                },
                {
                    "start": 1270.59,
                    "end": 1271.49,
                    "word": ",",
                    "probability": 0.9404296875
                },
                {
                    "start": 1271.49,
                    "end": 1272.15,
                    "word": " that",
                    "probability": 0.80322265625
                },
                {
                    "start": 1272.15,
                    "end": 1272.63,
                    "word": " data's",
                    "probability": 0.587646484375
                },
                {
                    "start": 1272.63,
                    "end": 1272.79,
                    "word": " out",
                    "probability": 0.82568359375
                },
                {
                    "start": 1272.79,
                    "end": 1273.07,
                    "word": " there",
                    "probability": 0.849609375
                },
                {
                    "start": 1273.07,
                    "end": 1273.13,
                    "word": ",",
                    "probability": 0.634765625
                },
                {
                    "start": 1273.13,
                    "end": 1273.39,
                    "word": " right",
                    "probability": 0.8720703125
                },
                {
                    "start": 1273.39,
                    "end": 1273.99,
                    "word": "?",
                    "probability": 0.92822265625
                },
                {
                    "start": 1274.17,
                    "end": 1274.77,
                    "word": " But",
                    "probability": 0.88623046875
                },
                {
                    "start": 1274.77,
                    "end": 1274.91,
                    "word": " if",
                    "probability": 0.87890625
                },
                {
                    "start": 1274.91,
                    "end": 1275.11,
                    "word": " you",
                    "probability": 0.896484375
                }
            ]
        },
        {
            "id": 179,
            "seek": 128532,
            "start": 1275.11,
            "end": 1282.47,
            "text": " are training a model for, like, a large medical device manufacturer, and it's trained on sensitive",
            "tokens": [
                51056,
                366,
                3097,
                257,
                2316,
                337,
                11,
                411,
                11,
                257,
                2416,
                4625,
                4302,
                18022,
                11,
                293,
                309,
                311,
                8895,
                322,
                9477,
                51428
            ],
            "temperature": 0.0,
            "avg_logprob": -0.187152784731653,
            "compression_ratio": 1.6948356807511737,
            "no_speech_prob": 0.00015592575073242188,
            "words": [
                {
                    "start": 1275.11,
                    "end": 1275.31,
                    "word": " are",
                    "probability": 0.8857421875
                },
                {
                    "start": 1275.31,
                    "end": 1275.71,
                    "word": " training",
                    "probability": 0.79150390625
                },
                {
                    "start": 1275.99,
                    "end": 1276.31,
                    "word": " a model",
                    "probability": 0.81884765625
                },
                {
                    "start": 1276.31,
                    "end": 1276.83,
                    "word": " for",
                    "probability": 0.89501953125
                },
                {
                    "start": 1276.83,
                    "end": 1277.19,
                    "word": ",",
                    "probability": 0.87890625
                },
                {
                    "start": 1277.19,
                    "end": 1277.45,
                    "word": " like",
                    "probability": 0.87548828125
                },
                {
                    "start": 1277.45,
                    "end": 1277.75,
                    "word": ",",
                    "probability": 0.98876953125
                },
                {
                    "start": 1278.05,
                    "end": 1278.65,
                    "word": " a large",
                    "probability": 0.93115234375
                },
                {
                    "start": 1278.65,
                    "end": 1279.17,
                    "word": " medical",
                    "probability": 0.90869140625
                },
                {
                    "start": 1279.17,
                    "end": 1279.65,
                    "word": " device",
                    "probability": 0.88134765625
                },
                {
                    "start": 1279.65,
                    "end": 1280.45,
                    "word": " manufacturer",
                    "probability": 0.75
                },
                {
                    "start": 1280.45,
                    "end": 1280.97,
                    "word": ",",
                    "probability": 0.78271484375
                },
                {
                    "start": 1280.97,
                    "end": 1281.21,
                    "word": " and",
                    "probability": 0.89453125
                },
                {
                    "start": 1281.21,
                    "end": 1281.43,
                    "word": " it's",
                    "probability": 0.947998046875
                },
                {
                    "start": 1281.43,
                    "end": 1281.71,
                    "word": " trained",
                    "probability": 0.81982421875
                },
                {
                    "start": 1281.71,
                    "end": 1281.99,
                    "word": " on",
                    "probability": 0.90380859375
                },
                {
                    "start": 1281.99,
                    "end": 1282.47,
                    "word": " sensitive",
                    "probability": 0.77978515625
                }
            ]
        },
        {
            "id": 180,
            "seek": 128532,
            "start": 1282.47,
            "end": 1289.05,
            "text": " medical information, and that information gets recovered, well, that's a lot worse.",
            "tokens": [
                51428,
                4625,
                1589,
                11,
                293,
                300,
                1589,
                2170,
                19542,
                11,
                731,
                11,
                300,
                311,
                257,
                688,
                5324,
                13,
                51720
            ],
            "temperature": 0.0,
            "avg_logprob": -0.187152784731653,
            "compression_ratio": 1.6948356807511737,
            "no_speech_prob": 0.00015592575073242188,
            "words": [
                {
                    "start": 1282.47,
                    "end": 1282.93,
                    "word": " medical",
                    "probability": 0.8984375
                },
                {
                    "start": 1282.93,
                    "end": 1283.71,
                    "word": " information",
                    "probability": 0.72509765625
                },
                {
                    "start": 1283.71,
                    "end": 1284.45,
                    "word": ",",
                    "probability": 0.92724609375
                },
                {
                    "start": 1284.45,
                    "end": 1284.85,
                    "word": " and",
                    "probability": 0.896484375
                },
                {
                    "start": 1284.85,
                    "end": 1285.13,
                    "word": " that",
                    "probability": 0.87646484375
                },
                {
                    "start": 1285.13,
                    "end": 1285.77,
                    "word": " information",
                    "probability": 0.7373046875
                },
                {
                    "start": 1285.77,
                    "end": 1286.03,
                    "word": " gets",
                    "probability": 0.7783203125
                },
                {
                    "start": 1286.03,
                    "end": 1286.45,
                    "word": " recovered",
                    "probability": 0.5849609375
                },
                {
                    "start": 1286.45,
                    "end": 1286.91,
                    "word": ",",
                    "probability": 0.9423828125
                },
                {
                    "start": 1286.91,
                    "end": 1287.11,
                    "word": " well",
                    "probability": 0.86181640625
                },
                {
                    "start": 1287.11,
                    "end": 1287.15,
                    "word": ",",
                    "probability": 0.95947265625
                },
                {
                    "start": 1287.15,
                    "end": 1287.71,
                    "word": " that's",
                    "probability": 0.8779296875
                },
                {
                    "start": 1287.85,
                    "end": 1287.97,
                    "word": " a lot",
                    "probability": 0.8896484375
                },
                {
                    "start": 1287.97,
                    "end": 1288.47,
                    "word": " worse",
                    "probability": 0.89404296875
                },
                {
                    "start": 1288.47,
                    "end": 1289.05,
                    "word": ".",
                    "probability": 0.8212890625
                }
            ]
        },
        {
            "id": 181,
            "seek": 131158,
            "start": 1289.05,
            "end": 1295.15,
            "text": " If you're a data scientist at Equifax or some other credit rating bureau,",
            "tokens": [
                50372,
                759,
                291,
                434,
                257,
                1412,
                12662,
                412,
                15624,
                351,
                2797,
                420,
                512,
                661,
                5397,
                10990,
                35343,
                11,
                50672
            ],
            "temperature": 0.0,
            "avg_logprob": -0.16634908373036036,
            "compression_ratio": 1.5613207547169812,
            "no_speech_prob": 0.0013885498046875,
            "words": [
                {
                    "start": 1289.05,
                    "end": 1289.47,
                    "word": " If",
                    "probability": 0.658203125
                },
                {
                    "start": 1289.47,
                    "end": 1289.67,
                    "word": " you're",
                    "probability": 0.921875
                },
                {
                    "start": 1289.97,
                    "end": 1290.23,
                    "word": " a data",
                    "probability": 0.8056640625
                },
                {
                    "start": 1290.23,
                    "end": 1290.75,
                    "word": " scientist",
                    "probability": 0.7373046875
                },
                {
                    "start": 1290.75,
                    "end": 1291.49,
                    "word": " at",
                    "probability": 0.90234375
                },
                {
                    "start": 1291.49,
                    "end": 1292.29,
                    "word": " Equifax",
                    "probability": 0.9627278645833334
                },
                {
                    "start": 1292.29,
                    "end": 1292.95,
                    "word": " or",
                    "probability": 0.7998046875
                },
                {
                    "start": 1292.95,
                    "end": 1293.37,
                    "word": " some",
                    "probability": 0.81640625
                },
                {
                    "start": 1293.37,
                    "end": 1293.67,
                    "word": " other",
                    "probability": 0.77685546875
                },
                {
                    "start": 1293.67,
                    "end": 1294.05,
                    "word": " credit",
                    "probability": 0.78466796875
                },
                {
                    "start": 1294.05,
                    "end": 1294.39,
                    "word": " rating",
                    "probability": 0.89208984375
                },
                {
                    "start": 1294.39,
                    "end": 1294.75,
                    "word": " bureau",
                    "probability": 0.62255859375
                },
                {
                    "start": 1294.75,
                    "end": 1295.15,
                    "word": ",",
                    "probability": 0.54150390625
                }
            ]
        },
        {
            "id": 182,
            "seek": 131158,
            "start": 1295.15,
            "end": 1299.09,
            "text": " and you have a bunch of sensitive financial data that you train a model on,",
            "tokens": [
                50672,
                293,
                291,
                362,
                257,
                3840,
                295,
                9477,
                4669,
                1412,
                300,
                291,
                3847,
                257,
                2316,
                322,
                11,
                50836
            ],
            "temperature": 0.0,
            "avg_logprob": -0.16634908373036036,
            "compression_ratio": 1.5613207547169812,
            "no_speech_prob": 0.0013885498046875,
            "words": [
                {
                    "start": 1295.15,
                    "end": 1295.55,
                    "word": " and",
                    "probability": 0.89794921875
                },
                {
                    "start": 1295.55,
                    "end": 1295.67,
                    "word": " you",
                    "probability": 0.880859375
                },
                {
                    "start": 1295.67,
                    "end": 1295.81,
                    "word": " have",
                    "probability": 0.8857421875
                },
                {
                    "start": 1295.95,
                    "end": 1296.13,
                    "word": " a bunch",
                    "probability": 0.80712890625
                },
                {
                    "start": 1296.13,
                    "end": 1296.29,
                    "word": " of",
                    "probability": 0.90625
                },
                {
                    "start": 1296.29,
                    "end": 1296.59,
                    "word": " sensitive",
                    "probability": 0.81005859375
                },
                {
                    "start": 1296.59,
                    "end": 1297.05,
                    "word": " financial",
                    "probability": 0.7958984375
                },
                {
                    "start": 1297.05,
                    "end": 1297.39,
                    "word": " data",
                    "probability": 0.8173828125
                },
                {
                    "start": 1297.39,
                    "end": 1297.57,
                    "word": " that",
                    "probability": 0.87939453125
                },
                {
                    "start": 1297.57,
                    "end": 1297.71,
                    "word": " you",
                    "probability": 0.90478515625
                },
                {
                    "start": 1297.71,
                    "end": 1297.93,
                    "word": " train",
                    "probability": 0.84130859375
                },
                {
                    "start": 1298.07,
                    "end": 1298.29,
                    "word": " a model",
                    "probability": 0.8193359375
                },
                {
                    "start": 1298.29,
                    "end": 1298.59,
                    "word": " on",
                    "probability": 0.90576171875
                },
                {
                    "start": 1298.59,
                    "end": 1299.09,
                    "word": ",",
                    "probability": 0.85107421875
                }
            ]
        },
        {
            "id": 183,
            "seek": 131158,
            "start": 1299.09,
            "end": 1308.57,
            "text": " that's pretty terrifying, right? The penalties for losing everybody's data at Equifax could be huge.",
            "tokens": [
                50876,
                300,
                311,
                1238,
                18106,
                11,
                558,
                30,
                440,
                35389,
                337,
                7027,
                2201,
                311,
                1412,
                412,
                15624,
                351,
                2797,
                727,
                312,
                2603,
                13,
                51332
            ],
            "temperature": 0.0,
            "avg_logprob": -0.16634908373036036,
            "compression_ratio": 1.5613207547169812,
            "no_speech_prob": 0.0013885498046875,
            "words": [
                {
                    "start": 1299.09,
                    "end": 1299.79,
                    "word": " that's",
                    "probability": 0.93994140625
                },
                {
                    "start": 1299.79,
                    "end": 1300.11,
                    "word": " pretty",
                    "probability": 0.88037109375
                },
                {
                    "start": 1300.11,
                    "end": 1300.75,
                    "word": " terrifying",
                    "probability": 0.8193359375
                },
                {
                    "start": 1300.75,
                    "end": 1301.29,
                    "word": ",",
                    "probability": 0.6357421875
                },
                {
                    "start": 1301.29,
                    "end": 1302.37,
                    "word": " right",
                    "probability": 0.8837890625
                },
                {
                    "start": 1302.37,
                    "end": 1302.43,
                    "word": "?",
                    "probability": 0.919921875
                },
                {
                    "start": 1302.43,
                    "end": 1302.71,
                    "word": " The",
                    "probability": 0.81689453125
                },
                {
                    "start": 1302.71,
                    "end": 1303.15,
                    "word": " penalties",
                    "probability": 0.89306640625
                },
                {
                    "start": 1303.15,
                    "end": 1303.63,
                    "word": " for",
                    "probability": 0.89990234375
                },
                {
                    "start": 1303.63,
                    "end": 1304.31,
                    "word": " losing",
                    "probability": 0.796875
                },
                {
                    "start": 1304.31,
                    "end": 1305.37,
                    "word": " everybody's",
                    "probability": 0.8740234375
                },
                {
                    "start": 1305.37,
                    "end": 1305.79,
                    "word": " data",
                    "probability": 0.8193359375
                },
                {
                    "start": 1305.79,
                    "end": 1306.45,
                    "word": " at",
                    "probability": 0.89599609375
                },
                {
                    "start": 1306.45,
                    "end": 1307.17,
                    "word": " Equifax",
                    "probability": 0.95068359375
                },
                {
                    "start": 1307.17,
                    "end": 1307.71,
                    "word": " could",
                    "probability": 0.810546875
                },
                {
                    "start": 1307.71,
                    "end": 1307.95,
                    "word": " be",
                    "probability": 0.9013671875
                },
                {
                    "start": 1307.95,
                    "end": 1308.47,
                    "word": " huge",
                    "probability": 0.8115234375
                },
                {
                    "start": 1308.47,
                    "end": 1308.57,
                    "word": ".",
                    "probability": 0.9169921875
                }
            ]
        },
        {
            "id": 184,
            "seek": 131158,
            "start": 1308.57,
            "end": 1315.31,
            "text": " You could have to pay a bunch of people's credit monitoring service for a while,",
            "tokens": [
                51332,
                509,
                727,
                362,
                281,
                1689,
                257,
                3840,
                295,
                561,
                311,
                5397,
                11028,
                2643,
                337,
                257,
                1339,
                11,
                51672
            ],
            "temperature": 0.0,
            "avg_logprob": -0.16634908373036036,
            "compression_ratio": 1.5613207547169812,
            "no_speech_prob": 0.0013885498046875,
            "words": [
                {
                    "start": 1308.57,
                    "end": 1308.67,
                    "word": " You",
                    "probability": 0.8935546875
                },
                {
                    "start": 1308.67,
                    "end": 1308.81,
                    "word": " could",
                    "probability": 0.814453125
                },
                {
                    "start": 1308.81,
                    "end": 1309.01,
                    "word": " have",
                    "probability": 0.8759765625
                },
                {
                    "start": 1309.01,
                    "end": 1309.25,
                    "word": " to",
                    "probability": 0.90380859375
                },
                {
                    "start": 1309.25,
                    "end": 1309.65,
                    "word": " pay",
                    "probability": 0.88671875
                },
                {
                    "start": 1309.81,
                    "end": 1310.03,
                    "word": " a bunch",
                    "probability": 0.81689453125
                },
                {
                    "start": 1310.03,
                    "end": 1310.25,
                    "word": " of",
                    "probability": 0.90625
                },
                {
                    "start": 1310.25,
                    "end": 1311.33,
                    "word": " people's",
                    "probability": 0.946044921875
                },
                {
                    "start": 1311.33,
                    "end": 1312.25,
                    "word": " credit",
                    "probability": 0.8134765625
                },
                {
                    "start": 1312.25,
                    "end": 1312.77,
                    "word": " monitoring",
                    "probability": 0.8212890625
                },
                {
                    "start": 1312.77,
                    "end": 1313.53,
                    "word": " service",
                    "probability": 0.787109375
                },
                {
                    "start": 1313.53,
                    "end": 1314.41,
                    "word": " for",
                    "probability": 0.892578125
                },
                {
                    "start": 1314.61,
                    "end": 1314.93,
                    "word": " a while",
                    "probability": 0.88134765625
                },
                {
                    "start": 1314.93,
                    "end": 1315.31,
                    "word": ",",
                    "probability": 0.421630859375
                }
            ]
        },
        {
            "id": 185,
            "seek": 134072,
            "start": 1315.31,
            "end": 1322.29,
            "text": " and that would be bad for your business. So, when we think about mitigations,",
            "tokens": [
                50364,
                293,
                300,
                576,
                312,
                1578,
                337,
                428,
                1606,
                13,
                407,
                11,
                562,
                321,
                519,
                466,
                15699,
                763,
                11,
                50704
            ],
            "temperature": 0.0,
            "avg_logprob": -0.20615897797843785,
            "compression_ratio": 1.8918918918918919,
            "no_speech_prob": 8.219480514526367e-05,
            "words": [
                {
                    "start": 1315.31,
                    "end": 1315.43,
                    "word": " and",
                    "probability": 0.19140625
                },
                {
                    "start": 1315.43,
                    "end": 1315.53,
                    "word": " that",
                    "probability": 0.85986328125
                },
                {
                    "start": 1315.53,
                    "end": 1315.67,
                    "word": " would",
                    "probability": 0.8310546875
                },
                {
                    "start": 1315.67,
                    "end": 1315.81,
                    "word": " be",
                    "probability": 0.896484375
                },
                {
                    "start": 1315.81,
                    "end": 1316.19,
                    "word": " bad",
                    "probability": 0.89111328125
                },
                {
                    "start": 1316.19,
                    "end": 1316.83,
                    "word": " for",
                    "probability": 0.87939453125
                },
                {
                    "start": 1316.83,
                    "end": 1317.01,
                    "word": " your",
                    "probability": 0.80908203125
                },
                {
                    "start": 1317.01,
                    "end": 1317.57,
                    "word": " business",
                    "probability": 0.89013671875
                },
                {
                    "start": 1317.57,
                    "end": 1318.05,
                    "word": ".",
                    "probability": 0.95703125
                },
                {
                    "start": 1319.69,
                    "end": 1320.17,
                    "word": " So",
                    "probability": 0.8095703125
                },
                {
                    "start": 1320.17,
                    "end": 1320.43,
                    "word": ",",
                    "probability": 0.497314453125
                },
                {
                    "start": 1320.43,
                    "end": 1320.79,
                    "word": " when",
                    "probability": 0.8759765625
                },
                {
                    "start": 1320.79,
                    "end": 1320.89,
                    "word": " we",
                    "probability": 0.90576171875
                },
                {
                    "start": 1320.89,
                    "end": 1321.03,
                    "word": " think",
                    "probability": 0.8017578125
                },
                {
                    "start": 1321.03,
                    "end": 1321.23,
                    "word": " about",
                    "probability": 0.82421875
                },
                {
                    "start": 1321.23,
                    "end": 1322.05,
                    "word": " mitigations",
                    "probability": 0.856201171875
                },
                {
                    "start": 1322.05,
                    "end": 1322.29,
                    "word": ",",
                    "probability": 0.88037109375
                }
            ]
        },
        {
            "id": 186,
            "seek": 134072,
            "start": 1323.35,
            "end": 1328.33,
            "text": " really, the first one is to protect access to the model. So, if they have direct access to",
            "tokens": [
                50776,
                534,
                11,
                264,
                700,
                472,
                307,
                281,
                2371,
                2105,
                281,
                264,
                2316,
                13,
                407,
                11,
                498,
                436,
                362,
                2047,
                2105,
                281,
                51012
            ],
            "temperature": 0.0,
            "avg_logprob": -0.20615897797843785,
            "compression_ratio": 1.8918918918918919,
            "no_speech_prob": 8.219480514526367e-05,
            "words": [
                {
                    "start": 1323.35,
                    "end": 1323.83,
                    "word": " really",
                    "probability": 0.73046875
                },
                {
                    "start": 1323.83,
                    "end": 1323.99,
                    "word": ",",
                    "probability": 0.33203125
                },
                {
                    "start": 1323.99,
                    "end": 1324.07,
                    "word": " the",
                    "probability": 0.8154296875
                },
                {
                    "start": 1324.07,
                    "end": 1324.29,
                    "word": " first",
                    "probability": 0.8115234375
                },
                {
                    "start": 1324.29,
                    "end": 1324.45,
                    "word": " one",
                    "probability": 0.81005859375
                },
                {
                    "start": 1324.45,
                    "end": 1324.57,
                    "word": " is",
                    "probability": 0.9013671875
                },
                {
                    "start": 1324.57,
                    "end": 1324.71,
                    "word": " to",
                    "probability": 0.89697265625
                },
                {
                    "start": 1324.71,
                    "end": 1324.99,
                    "word": " protect",
                    "probability": 0.79638671875
                },
                {
                    "start": 1324.99,
                    "end": 1325.41,
                    "word": " access",
                    "probability": 0.79150390625
                },
                {
                    "start": 1325.41,
                    "end": 1325.67,
                    "word": " to",
                    "probability": 0.8984375
                },
                {
                    "start": 1325.67,
                    "end": 1325.81,
                    "word": " the",
                    "probability": 0.80712890625
                },
                {
                    "start": 1325.81,
                    "end": 1326.09,
                    "word": " model",
                    "probability": 0.81005859375
                },
                {
                    "start": 1326.09,
                    "end": 1326.49,
                    "word": ".",
                    "probability": 0.90283203125
                },
                {
                    "start": 1326.49,
                    "end": 1326.67,
                    "word": " So",
                    "probability": 0.81396484375
                },
                {
                    "start": 1326.67,
                    "end": 1326.79,
                    "word": ",",
                    "probability": 0.60498046875
                },
                {
                    "start": 1326.79,
                    "end": 1326.91,
                    "word": " if",
                    "probability": 0.900390625
                },
                {
                    "start": 1326.91,
                    "end": 1327.03,
                    "word": " they",
                    "probability": 0.7333984375
                },
                {
                    "start": 1327.03,
                    "end": 1327.19,
                    "word": " have",
                    "probability": 0.873046875
                },
                {
                    "start": 1327.19,
                    "end": 1327.75,
                    "word": " direct",
                    "probability": 0.82470703125
                },
                {
                    "start": 1327.75,
                    "end": 1328.11,
                    "word": " access",
                    "probability": 0.7890625
                },
                {
                    "start": 1328.11,
                    "end": 1328.33,
                    "word": " to",
                    "probability": 0.90185546875
                }
            ]
        },
        {
            "id": 187,
            "seek": 134072,
            "start": 1328.33,
            "end": 1332.69,
            "text": " the model, it's much easier to do inversion. It's much, much easier to do inversion.",
            "tokens": [
                51012,
                264,
                2316,
                11,
                309,
                311,
                709,
                3571,
                281,
                360,
                43576,
                13,
                467,
                311,
                709,
                11,
                709,
                3571,
                281,
                360,
                43576,
                13,
                51216
            ],
            "temperature": 0.0,
            "avg_logprob": -0.20615897797843785,
            "compression_ratio": 1.8918918918918919,
            "no_speech_prob": 8.219480514526367e-05,
            "words": [
                {
                    "start": 1328.33,
                    "end": 1328.49,
                    "word": " the",
                    "probability": 0.8115234375
                },
                {
                    "start": 1328.49,
                    "end": 1328.71,
                    "word": " model",
                    "probability": 0.8115234375
                },
                {
                    "start": 1328.71,
                    "end": 1328.87,
                    "word": ",",
                    "probability": 0.95947265625
                },
                {
                    "start": 1328.87,
                    "end": 1329.03,
                    "word": " it's",
                    "probability": 0.942138671875
                },
                {
                    "start": 1329.03,
                    "end": 1329.41,
                    "word": " much",
                    "probability": 0.81103515625
                },
                {
                    "start": 1329.41,
                    "end": 1329.71,
                    "word": " easier",
                    "probability": 0.92529296875
                },
                {
                    "start": 1329.71,
                    "end": 1329.89,
                    "word": " to",
                    "probability": 0.90576171875
                },
                {
                    "start": 1329.89,
                    "end": 1330.03,
                    "word": " do",
                    "probability": 0.9052734375
                },
                {
                    "start": 1330.03,
                    "end": 1330.41,
                    "word": " inversion",
                    "probability": 0.5537109375
                },
                {
                    "start": 1330.41,
                    "end": 1330.77,
                    "word": ".",
                    "probability": 0.89990234375
                },
                {
                    "start": 1330.77,
                    "end": 1330.89,
                    "word": " It's",
                    "probability": 0.921142578125
                },
                {
                    "start": 1330.89,
                    "end": 1331.17,
                    "word": " much",
                    "probability": 0.8212890625
                },
                {
                    "start": 1331.17,
                    "end": 1331.31,
                    "word": ",",
                    "probability": 0.91796875
                },
                {
                    "start": 1331.31,
                    "end": 1331.35,
                    "word": " much",
                    "probability": 0.80615234375
                },
                {
                    "start": 1331.35,
                    "end": 1331.61,
                    "word": " easier",
                    "probability": 0.9248046875
                },
                {
                    "start": 1331.61,
                    "end": 1331.79,
                    "word": " to",
                    "probability": 0.90478515625
                },
                {
                    "start": 1331.79,
                    "end": 1331.93,
                    "word": " do",
                    "probability": 0.91259765625
                },
                {
                    "start": 1331.93,
                    "end": 1332.21,
                    "word": " inversion",
                    "probability": 0.6875
                },
                {
                    "start": 1332.21,
                    "end": 1332.69,
                    "word": ".",
                    "probability": 0.99560546875
                }
            ]
        },
        {
            "id": 188,
            "seek": 134072,
            "start": 1334.11,
            "end": 1338.37,
            "text": " There are different techniques you can use if you have direct access to the model.",
            "tokens": [
                51308,
                821,
                366,
                819,
                7512,
                291,
                393,
                764,
                498,
                291,
                362,
                2047,
                2105,
                281,
                264,
                2316,
                13,
                51492
            ],
            "temperature": 0.0,
            "avg_logprob": -0.20615897797843785,
            "compression_ratio": 1.8918918918918919,
            "no_speech_prob": 8.219480514526367e-05,
            "words": [
                {
                    "start": 1334.11,
                    "end": 1334.45,
                    "word": " There",
                    "probability": 0.78515625
                },
                {
                    "start": 1334.45,
                    "end": 1334.53,
                    "word": " are",
                    "probability": 0.88232421875
                },
                {
                    "start": 1334.53,
                    "end": 1334.77,
                    "word": " different",
                    "probability": 0.6962890625
                },
                {
                    "start": 1334.77,
                    "end": 1335.15,
                    "word": " techniques",
                    "probability": 0.84033203125
                },
                {
                    "start": 1335.15,
                    "end": 1335.33,
                    "word": " you",
                    "probability": 0.88916015625
                },
                {
                    "start": 1335.33,
                    "end": 1335.47,
                    "word": " can",
                    "probability": 0.88916015625
                },
                {
                    "start": 1335.47,
                    "end": 1335.71,
                    "word": " use",
                    "probability": 0.8154296875
                },
                {
                    "start": 1335.71,
                    "end": 1335.85,
                    "word": " if",
                    "probability": 0.87451171875
                },
                {
                    "start": 1335.85,
                    "end": 1335.93,
                    "word": " you",
                    "probability": 0.8896484375
                },
                {
                    "start": 1335.93,
                    "end": 1336.05,
                    "word": " have",
                    "probability": 0.86962890625
                },
                {
                    "start": 1336.05,
                    "end": 1336.55,
                    "word": " direct",
                    "probability": 0.7958984375
                },
                {
                    "start": 1336.55,
                    "end": 1337.07,
                    "word": " access",
                    "probability": 0.78076171875
                },
                {
                    "start": 1337.07,
                    "end": 1337.43,
                    "word": " to",
                    "probability": 0.90380859375
                },
                {
                    "start": 1337.43,
                    "end": 1337.61,
                    "word": " the",
                    "probability": 0.81787109375
                },
                {
                    "start": 1337.61,
                    "end": 1337.89,
                    "word": " model",
                    "probability": 0.8212890625
                },
                {
                    "start": 1337.89,
                    "end": 1338.37,
                    "word": ".",
                    "probability": 0.8681640625
                }
            ]
        },
        {
            "id": 189,
            "seek": 134072,
            "start": 1338.95,
            "end": 1344.45,
            "text": " And then the other mitigations for model inversion are essentially don't let people",
            "tokens": [
                51560,
                400,
                550,
                264,
                661,
                15699,
                763,
                337,
                2316,
                43576,
                366,
                4476,
                500,
                380,
                718,
                561,
                51820
            ],
            "temperature": 0.0,
            "avg_logprob": -0.20615897797843785,
            "compression_ratio": 1.8918918918918919,
            "no_speech_prob": 8.219480514526367e-05,
            "words": [
                {
                    "start": 1338.95,
                    "end": 1339.43,
                    "word": " And",
                    "probability": 0.8583984375
                },
                {
                    "start": 1339.43,
                    "end": 1339.55,
                    "word": " then",
                    "probability": 0.7158203125
                },
                {
                    "start": 1339.55,
                    "end": 1339.67,
                    "word": " the",
                    "probability": 0.6796875
                },
                {
                    "start": 1339.67,
                    "end": 1339.87,
                    "word": " other",
                    "probability": 0.7705078125
                },
                {
                    "start": 1339.87,
                    "end": 1340.73,
                    "word": " mitigations",
                    "probability": 0.88330078125
                },
                {
                    "start": 1340.73,
                    "end": 1340.97,
                    "word": " for",
                    "probability": 0.89892578125
                },
                {
                    "start": 1340.97,
                    "end": 1341.31,
                    "word": " model",
                    "probability": 0.80078125
                },
                {
                    "start": 1341.31,
                    "end": 1341.77,
                    "word": " inversion",
                    "probability": 0.66943359375
                },
                {
                    "start": 1341.77,
                    "end": 1342.49,
                    "word": " are",
                    "probability": 0.84521484375
                },
                {
                    "start": 1342.49,
                    "end": 1342.93,
                    "word": " essentially",
                    "probability": 0.58154296875
                },
                {
                    "start": 1342.93,
                    "end": 1343.79,
                    "word": " don't",
                    "probability": 0.7607421875
                },
                {
                    "start": 1343.79,
                    "end": 1344.11,
                    "word": " let",
                    "probability": 0.892578125
                },
                {
                    "start": 1344.11,
                    "end": 1344.45,
                    "word": " people",
                    "probability": 0.92138671875
                }
            ]
        },
        {
            "id": 190,
            "seek": 137072,
            "start": 1344.45,
            "end": 1354.89,
            "text": " steal your model. Because then it's easier to use it as sort of the discriminator in the GAN.",
            "tokens": [
                50364,
                11009,
                428,
                2316,
                13,
                1436,
                550,
                309,
                311,
                3571,
                281,
                764,
                309,
                382,
                1333,
                295,
                264,
                20828,
                1639,
                294,
                264,
                460,
                1770,
                13,
                50856
            ],
            "temperature": 0.0,
            "avg_logprob": -0.2223165692842525,
            "compression_ratio": 1.5473251028806585,
            "no_speech_prob": 0.0009546279907226562,
            "words": [
                {
                    "start": 1344.45,
                    "end": 1344.85,
                    "word": " steal",
                    "probability": 0.7236328125
                },
                {
                    "start": 1344.85,
                    "end": 1345.27,
                    "word": " your",
                    "probability": 0.80615234375
                },
                {
                    "start": 1345.27,
                    "end": 1345.57,
                    "word": " model",
                    "probability": 0.82275390625
                },
                {
                    "start": 1345.57,
                    "end": 1346.17,
                    "word": ".",
                    "probability": 0.240966796875
                },
                {
                    "start": 1346.63,
                    "end": 1347.07,
                    "word": " Because",
                    "probability": 0.8623046875
                },
                {
                    "start": 1347.07,
                    "end": 1347.43,
                    "word": " then",
                    "probability": 0.71533203125
                },
                {
                    "start": 1347.43,
                    "end": 1349.05,
                    "word": " it's",
                    "probability": 0.868896484375
                },
                {
                    "start": 1349.05,
                    "end": 1349.49,
                    "word": " easier",
                    "probability": 0.92529296875
                },
                {
                    "start": 1349.49,
                    "end": 1349.81,
                    "word": " to",
                    "probability": 0.90673828125
                },
                {
                    "start": 1349.81,
                    "end": 1350.21,
                    "word": " use",
                    "probability": 0.81982421875
                },
                {
                    "start": 1350.21,
                    "end": 1350.41,
                    "word": " it",
                    "probability": 0.904296875
                },
                {
                    "start": 1350.41,
                    "end": 1350.69,
                    "word": " as",
                    "probability": 0.90087890625
                },
                {
                    "start": 1350.69,
                    "end": 1350.89,
                    "word": " sort",
                    "probability": 0.765625
                },
                {
                    "start": 1350.89,
                    "end": 1351.09,
                    "word": " of",
                    "probability": 0.90625
                },
                {
                    "start": 1351.09,
                    "end": 1352.35,
                    "word": " the",
                    "probability": 0.76904296875
                },
                {
                    "start": 1352.35,
                    "end": 1353.33,
                    "word": " discriminator",
                    "probability": 0.84423828125
                },
                {
                    "start": 1353.33,
                    "end": 1353.83,
                    "word": " in",
                    "probability": 0.8896484375
                },
                {
                    "start": 1353.83,
                    "end": 1353.99,
                    "word": " the",
                    "probability": 0.82177734375
                },
                {
                    "start": 1353.99,
                    "end": 1354.29,
                    "word": " GAN",
                    "probability": 0.71435546875
                },
                {
                    "start": 1354.29,
                    "end": 1354.89,
                    "word": ".",
                    "probability": 0.83935546875
                }
            ]
        },
        {
            "id": 191,
            "seek": 137072,
            "start": 1354.99,
            "end": 1360.37,
            "text": " The architecture is a little more complicated than a traditional generative adversarial network,",
            "tokens": [
                50916,
                440,
                9482,
                307,
                257,
                707,
                544,
                6179,
                813,
                257,
                5164,
                1337,
                1166,
                17641,
                44745,
                3209,
                11,
                51152
            ],
            "temperature": 0.0,
            "avg_logprob": -0.2223165692842525,
            "compression_ratio": 1.5473251028806585,
            "no_speech_prob": 0.0009546279907226562,
            "words": [
                {
                    "start": 1354.99,
                    "end": 1355.53,
                    "word": " The",
                    "probability": 0.429931640625
                },
                {
                    "start": 1355.53,
                    "end": 1356.17,
                    "word": " architecture",
                    "probability": 0.83544921875
                },
                {
                    "start": 1356.17,
                    "end": 1356.35,
                    "word": " is",
                    "probability": 0.70947265625
                },
                {
                    "start": 1356.45,
                    "end": 1356.53,
                    "word": " a little",
                    "probability": 0.80615234375
                },
                {
                    "start": 1356.53,
                    "end": 1356.79,
                    "word": " more",
                    "probability": 0.8427734375
                },
                {
                    "start": 1356.79,
                    "end": 1357.35,
                    "word": " complicated",
                    "probability": 0.8056640625
                },
                {
                    "start": 1357.35,
                    "end": 1357.63,
                    "word": " than",
                    "probability": 0.88134765625
                },
                {
                    "start": 1357.77,
                    "end": 1358.31,
                    "word": " a traditional",
                    "probability": 0.79443359375
                },
                {
                    "start": 1358.31,
                    "end": 1359.07,
                    "word": " generative",
                    "probability": 0.855224609375
                },
                {
                    "start": 1359.07,
                    "end": 1359.73,
                    "word": " adversarial",
                    "probability": 0.932861328125
                },
                {
                    "start": 1359.73,
                    "end": 1360.13,
                    "word": " network",
                    "probability": 0.83203125
                },
                {
                    "start": 1360.13,
                    "end": 1360.37,
                    "word": ",",
                    "probability": 0.50390625
                }
            ]
        },
        {
            "id": 192,
            "seek": 137072,
            "start": 1360.37,
            "end": 1366.29,
            "text": " but the idea is pretty similar. You basically show it an example and say, does this look familiar?",
            "tokens": [
                51152,
                457,
                264,
                1558,
                307,
                1238,
                2531,
                13,
                509,
                1936,
                855,
                309,
                364,
                1365,
                293,
                584,
                11,
                775,
                341,
                574,
                4963,
                30,
                51456
            ],
            "temperature": 0.0,
            "avg_logprob": -0.2223165692842525,
            "compression_ratio": 1.5473251028806585,
            "no_speech_prob": 0.0009546279907226562,
            "words": [
                {
                    "start": 1360.37,
                    "end": 1360.47,
                    "word": " but",
                    "probability": 0.8935546875
                },
                {
                    "start": 1360.47,
                    "end": 1360.61,
                    "word": " the",
                    "probability": 0.81005859375
                },
                {
                    "start": 1360.61,
                    "end": 1360.87,
                    "word": " idea",
                    "probability": 0.78076171875
                },
                {
                    "start": 1360.87,
                    "end": 1361.03,
                    "word": " is",
                    "probability": 0.88232421875
                },
                {
                    "start": 1361.03,
                    "end": 1361.33,
                    "word": " pretty",
                    "probability": 0.8740234375
                },
                {
                    "start": 1361.33,
                    "end": 1361.73,
                    "word": " similar",
                    "probability": 0.91064453125
                },
                {
                    "start": 1361.73,
                    "end": 1362.33,
                    "word": ".",
                    "probability": 0.99462890625
                },
                {
                    "start": 1362.71,
                    "end": 1363.27,
                    "word": " You",
                    "probability": 0.90185546875
                },
                {
                    "start": 1363.27,
                    "end": 1363.65,
                    "word": " basically",
                    "probability": 0.89208984375
                },
                {
                    "start": 1363.65,
                    "end": 1363.93,
                    "word": " show",
                    "probability": 0.87451171875
                },
                {
                    "start": 1363.93,
                    "end": 1364.03,
                    "word": " it",
                    "probability": 0.80419921875
                },
                {
                    "start": 1364.03,
                    "end": 1364.15,
                    "word": " an",
                    "probability": 0.8427734375
                },
                {
                    "start": 1364.15,
                    "end": 1364.49,
                    "word": " example",
                    "probability": 0.9326171875
                },
                {
                    "start": 1364.49,
                    "end": 1364.71,
                    "word": " and",
                    "probability": 0.8701171875
                },
                {
                    "start": 1364.71,
                    "end": 1364.85,
                    "word": " say",
                    "probability": 0.8935546875
                },
                {
                    "start": 1364.85,
                    "end": 1364.93,
                    "word": ",",
                    "probability": 0.69140625
                },
                {
                    "start": 1364.93,
                    "end": 1365.03,
                    "word": " does",
                    "probability": 0.44873046875
                },
                {
                    "start": 1365.03,
                    "end": 1365.21,
                    "word": " this",
                    "probability": 0.87744140625
                },
                {
                    "start": 1365.21,
                    "end": 1365.37,
                    "word": " look",
                    "probability": 0.91455078125
                },
                {
                    "start": 1365.37,
                    "end": 1365.71,
                    "word": " familiar",
                    "probability": 0.77099609375
                },
                {
                    "start": 1365.71,
                    "end": 1366.29,
                    "word": "?",
                    "probability": 0.861328125
                }
            ]
        },
        {
            "id": 193,
            "seek": 137072,
            "start": 1366.29,
            "end": 1372.97,
            "text": " And it goes yes or no. And if it says yes, then you may have inverted the information.",
            "tokens": [
                51456,
                400,
                309,
                1709,
                2086,
                420,
                572,
                13,
                400,
                498,
                309,
                1619,
                2086,
                11,
                550,
                291,
                815,
                362,
                38969,
                264,
                1536,
                687,
                399,
                13,
                51768
            ],
            "temperature": 0.0,
            "avg_logprob": -0.2223165692842525,
            "compression_ratio": 1.5473251028806585,
            "no_speech_prob": 0.0009546279907226562,
            "words": [
                {
                    "start": 1366.29,
                    "end": 1366.57,
                    "word": " And",
                    "probability": 0.85986328125
                },
                {
                    "start": 1366.57,
                    "end": 1366.71,
                    "word": " it",
                    "probability": 0.88720703125
                },
                {
                    "start": 1366.71,
                    "end": 1367.05,
                    "word": " goes",
                    "probability": 0.84716796875
                },
                {
                    "start": 1367.05,
                    "end": 1367.33,
                    "word": " yes",
                    "probability": 0.63134765625
                },
                {
                    "start": 1367.33,
                    "end": 1367.49,
                    "word": " or",
                    "probability": 0.908203125
                },
                {
                    "start": 1367.49,
                    "end": 1367.83,
                    "word": " no",
                    "probability": 0.9091796875
                },
                {
                    "start": 1367.83,
                    "end": 1368.43,
                    "word": ".",
                    "probability": 0.78271484375
                },
                {
                    "start": 1368.57,
                    "end": 1369.07,
                    "word": " And",
                    "probability": 0.86279296875
                },
                {
                    "start": 1369.07,
                    "end": 1369.17,
                    "word": " if",
                    "probability": 0.9033203125
                },
                {
                    "start": 1369.17,
                    "end": 1369.29,
                    "word": " it",
                    "probability": 0.9052734375
                },
                {
                    "start": 1369.29,
                    "end": 1369.47,
                    "word": " says",
                    "probability": 0.783203125
                },
                {
                    "start": 1369.47,
                    "end": 1369.91,
                    "word": " yes",
                    "probability": 0.857421875
                },
                {
                    "start": 1369.91,
                    "end": 1370.03,
                    "word": ",",
                    "probability": 0.9658203125
                },
                {
                    "start": 1370.03,
                    "end": 1370.25,
                    "word": " then",
                    "probability": 0.71142578125
                },
                {
                    "start": 1370.25,
                    "end": 1370.49,
                    "word": " you",
                    "probability": 0.8916015625
                },
                {
                    "start": 1370.49,
                    "end": 1370.87,
                    "word": " may",
                    "probability": 0.8994140625
                },
                {
                    "start": 1370.87,
                    "end": 1371.01,
                    "word": " have",
                    "probability": 0.88037109375
                },
                {
                    "start": 1371.01,
                    "end": 1371.39,
                    "word": " inverted",
                    "probability": 0.666015625
                },
                {
                    "start": 1371.39,
                    "end": 1371.99,
                    "word": " the",
                    "probability": 0.791015625
                },
                {
                    "start": 1371.99,
                    "end": 1372.47,
                    "word": " information",
                    "probability": 0.746337890625
                },
                {
                    "start": 1372.47,
                    "end": 1372.97,
                    "word": ".",
                    "probability": 0.99755859375
                }
            ]
        },
        {
            "id": 194,
            "seek": 139642,
            "start": 1375.05,
            "end": 1382.79,
            "text": " So, our next threat is poisoning. Data poisoning. And data poisoning is when malicious users",
            "tokens": [
                50400,
                407,
                11,
                527,
                958,
                4734,
                307,
                36778,
                13,
                11888,
                36778,
                13,
                400,
                1412,
                36778,
                307,
                562,
                33496,
                5022,
                50780
            ],
            "temperature": 0.0,
            "avg_logprob": -0.21298076923076922,
            "compression_ratio": 1.6420454545454546,
            "no_speech_prob": 0.0003514289855957031,
            "words": [
                {
                    "start": 1375.05,
                    "end": 1375.47,
                    "word": " So",
                    "probability": 0.71435546875
                },
                {
                    "start": 1375.47,
                    "end": 1375.55,
                    "word": ",",
                    "probability": 0.2254638671875
                },
                {
                    "start": 1375.55,
                    "end": 1375.61,
                    "word": " our",
                    "probability": 0.80078125
                },
                {
                    "start": 1375.61,
                    "end": 1375.83,
                    "word": " next",
                    "probability": 0.833984375
                },
                {
                    "start": 1375.83,
                    "end": 1376.21,
                    "word": " threat",
                    "probability": 0.83056640625
                },
                {
                    "start": 1376.21,
                    "end": 1377.07,
                    "word": " is",
                    "probability": 0.90478515625
                },
                {
                    "start": 1377.07,
                    "end": 1377.97,
                    "word": " poisoning",
                    "probability": 0.5751953125
                },
                {
                    "start": 1377.97,
                    "end": 1378.31,
                    "word": ".",
                    "probability": 0.2322998046875
                },
                {
                    "start": 1378.31,
                    "end": 1378.49,
                    "word": " Data",
                    "probability": 0.9169921875
                },
                {
                    "start": 1378.49,
                    "end": 1378.87,
                    "word": " poisoning",
                    "probability": 0.79638671875
                },
                {
                    "start": 1378.87,
                    "end": 1379.71,
                    "word": ".",
                    "probability": 0.94677734375
                },
                {
                    "start": 1379.93,
                    "end": 1380.49,
                    "word": " And",
                    "probability": 0.7568359375
                },
                {
                    "start": 1380.49,
                    "end": 1380.73,
                    "word": " data",
                    "probability": 0.720703125
                },
                {
                    "start": 1380.73,
                    "end": 1381.05,
                    "word": " poisoning",
                    "probability": 0.78759765625
                },
                {
                    "start": 1381.05,
                    "end": 1381.53,
                    "word": " is",
                    "probability": 0.90869140625
                },
                {
                    "start": 1381.53,
                    "end": 1381.81,
                    "word": " when",
                    "probability": 0.86962890625
                },
                {
                    "start": 1381.81,
                    "end": 1382.25,
                    "word": " malicious",
                    "probability": 0.83642578125
                },
                {
                    "start": 1382.25,
                    "end": 1382.79,
                    "word": " users",
                    "probability": 0.8330078125
                }
            ]
        },
        {
            "id": 195,
            "seek": 139642,
            "start": 1382.79,
            "end": 1390.65,
            "text": " inject bad training data into a model to corrupt it. So, usually, this happens in online learning",
            "tokens": [
                50780,
                10711,
                1578,
                3097,
                1412,
                666,
                257,
                2316,
                281,
                17366,
                309,
                13,
                407,
                11,
                2673,
                11,
                341,
                2314,
                294,
                2950,
                2539,
                51172
            ],
            "temperature": 0.0,
            "avg_logprob": -0.21298076923076922,
            "compression_ratio": 1.6420454545454546,
            "no_speech_prob": 0.0003514289855957031,
            "words": [
                {
                    "start": 1382.79,
                    "end": 1383.21,
                    "word": " inject",
                    "probability": 0.6728515625
                },
                {
                    "start": 1383.21,
                    "end": 1383.61,
                    "word": " bad",
                    "probability": 0.91650390625
                },
                {
                    "start": 1383.61,
                    "end": 1384.01,
                    "word": " training",
                    "probability": 0.814453125
                },
                {
                    "start": 1384.01,
                    "end": 1384.45,
                    "word": " data",
                    "probability": 0.81494140625
                },
                {
                    "start": 1384.45,
                    "end": 1384.95,
                    "word": " into",
                    "probability": 0.74365234375
                },
                {
                    "start": 1385.19,
                    "end": 1385.43,
                    "word": " a model",
                    "probability": 0.79345703125
                },
                {
                    "start": 1385.43,
                    "end": 1385.67,
                    "word": " to",
                    "probability": 0.88720703125
                },
                {
                    "start": 1385.67,
                    "end": 1386.01,
                    "word": " corrupt",
                    "probability": 0.82275390625
                },
                {
                    "start": 1386.01,
                    "end": 1386.27,
                    "word": " it",
                    "probability": 0.90185546875
                },
                {
                    "start": 1386.27,
                    "end": 1387.07,
                    "word": ".",
                    "probability": 0.9833984375
                },
                {
                    "start": 1387.07,
                    "end": 1387.31,
                    "word": " So",
                    "probability": 0.83935546875
                },
                {
                    "start": 1387.31,
                    "end": 1387.37,
                    "word": ",",
                    "probability": 0.87451171875
                },
                {
                    "start": 1387.37,
                    "end": 1387.97,
                    "word": " usually",
                    "probability": 0.82568359375
                },
                {
                    "start": 1387.97,
                    "end": 1388.47,
                    "word": ",",
                    "probability": 0.25634765625
                },
                {
                    "start": 1388.47,
                    "end": 1388.65,
                    "word": " this",
                    "probability": 0.8837890625
                },
                {
                    "start": 1388.65,
                    "end": 1389.13,
                    "word": " happens",
                    "probability": 0.90869140625
                },
                {
                    "start": 1389.13,
                    "end": 1389.57,
                    "word": " in",
                    "probability": 0.9013671875
                },
                {
                    "start": 1389.57,
                    "end": 1390.13,
                    "word": " online",
                    "probability": 0.77880859375
                },
                {
                    "start": 1390.13,
                    "end": 1390.65,
                    "word": " learning",
                    "probability": 0.841796875
                }
            ]
        },
        {
            "id": 196,
            "seek": 139642,
            "start": 1391.25,
            "end": 1400.15,
            "text": " where the model is continuously trained on the input to the model. And that is definitely a threat",
            "tokens": [
                51204,
                689,
                264,
                2316,
                307,
                15684,
                8895,
                322,
                264,
                4846,
                281,
                264,
                2316,
                13,
                400,
                300,
                307,
                2138,
                257,
                4734,
                51640
            ],
            "temperature": 0.0,
            "avg_logprob": -0.21298076923076922,
            "compression_ratio": 1.6420454545454546,
            "no_speech_prob": 0.0003514289855957031,
            "words": [
                {
                    "start": 1391.25,
                    "end": 1391.67,
                    "word": " where",
                    "probability": 0.406005859375
                },
                {
                    "start": 1391.67,
                    "end": 1392.11,
                    "word": " the",
                    "probability": 0.8173828125
                },
                {
                    "start": 1392.11,
                    "end": 1392.35,
                    "word": " model",
                    "probability": 0.8115234375
                },
                {
                    "start": 1392.35,
                    "end": 1392.57,
                    "word": " is",
                    "probability": 0.90771484375
                },
                {
                    "start": 1392.57,
                    "end": 1393.21,
                    "word": " continuously",
                    "probability": 0.837890625
                },
                {
                    "start": 1393.21,
                    "end": 1393.87,
                    "word": " trained",
                    "probability": 0.82861328125
                },
                {
                    "start": 1393.87,
                    "end": 1394.07,
                    "word": " on",
                    "probability": 0.89453125
                },
                {
                    "start": 1394.07,
                    "end": 1394.21,
                    "word": " the",
                    "probability": 0.80419921875
                },
                {
                    "start": 1394.21,
                    "end": 1394.57,
                    "word": " input",
                    "probability": 0.88330078125
                },
                {
                    "start": 1394.57,
                    "end": 1395.01,
                    "word": " to",
                    "probability": 0.90185546875
                },
                {
                    "start": 1395.01,
                    "end": 1395.23,
                    "word": " the",
                    "probability": 0.8134765625
                },
                {
                    "start": 1395.23,
                    "end": 1395.51,
                    "word": " model",
                    "probability": 0.8046875
                },
                {
                    "start": 1395.51,
                    "end": 1396.35,
                    "word": ".",
                    "probability": 0.728515625
                },
                {
                    "start": 1396.39,
                    "end": 1397.09,
                    "word": " And",
                    "probability": 0.876953125
                },
                {
                    "start": 1397.09,
                    "end": 1397.33,
                    "word": " that",
                    "probability": 0.880859375
                },
                {
                    "start": 1397.33,
                    "end": 1397.75,
                    "word": " is",
                    "probability": 0.90771484375
                },
                {
                    "start": 1397.75,
                    "end": 1398.75,
                    "word": " definitely",
                    "probability": 0.51904296875
                },
                {
                    "start": 1399.47,
                    "end": 1400.15,
                    "word": " a threat",
                    "probability": 0.82373046875
                }
            ]
        },
        {
            "id": 197,
            "seek": 142004,
            "start": 1401.65,
            "end": 1408.17,
            "text": " that needs to be considered. The other option is that they can, you know, get access to wherever",
            "tokens": [
                50400,
                300,
                2203,
                281,
                312,
                4888,
                13,
                440,
                661,
                3614,
                307,
                300,
                436,
                393,
                11,
                291,
                458,
                11,
                483,
                2105,
                281,
                8660,
                50764
            ],
            "temperature": 0.0,
            "avg_logprob": -0.20052082815032074,
            "compression_ratio": 1.559782608695652,
            "no_speech_prob": 0.0005359649658203125,
            "words": [
                {
                    "start": 1401.65,
                    "end": 1402.25,
                    "word": " that",
                    "probability": 0.525390625
                },
                {
                    "start": 1402.25,
                    "end": 1402.49,
                    "word": " needs",
                    "probability": 0.77392578125
                },
                {
                    "start": 1402.49,
                    "end": 1402.63,
                    "word": " to",
                    "probability": 0.90966796875
                },
                {
                    "start": 1402.63,
                    "end": 1403.01,
                    "word": " be",
                    "probability": 0.912109375
                },
                {
                    "start": 1403.01,
                    "end": 1403.75,
                    "word": " considered",
                    "probability": 0.6279296875
                },
                {
                    "start": 1403.75,
                    "end": 1404.19,
                    "word": ".",
                    "probability": 0.95556640625
                },
                {
                    "start": 1404.19,
                    "end": 1404.51,
                    "word": " The",
                    "probability": 0.80419921875
                },
                {
                    "start": 1404.51,
                    "end": 1404.71,
                    "word": " other",
                    "probability": 0.7763671875
                },
                {
                    "start": 1404.71,
                    "end": 1405.11,
                    "word": " option",
                    "probability": 0.787109375
                },
                {
                    "start": 1405.11,
                    "end": 1405.39,
                    "word": " is",
                    "probability": 0.90380859375
                },
                {
                    "start": 1405.39,
                    "end": 1405.67,
                    "word": " that",
                    "probability": 0.8623046875
                },
                {
                    "start": 1405.67,
                    "end": 1405.97,
                    "word": " they",
                    "probability": 0.740234375
                },
                {
                    "start": 1405.97,
                    "end": 1406.31,
                    "word": " can",
                    "probability": 0.89892578125
                },
                {
                    "start": 1406.31,
                    "end": 1406.75,
                    "word": ",",
                    "probability": 0.31298828125
                },
                {
                    "start": 1406.75,
                    "end": 1406.95,
                    "word": " you",
                    "probability": 0.83740234375
                },
                {
                    "start": 1406.95,
                    "end": 1407.07,
                    "word": " know",
                    "probability": 0.81494140625
                },
                {
                    "start": 1407.07,
                    "end": 1407.11,
                    "word": ",",
                    "probability": 0.96875
                },
                {
                    "start": 1407.11,
                    "end": 1407.29,
                    "word": " get",
                    "probability": 0.8994140625
                },
                {
                    "start": 1407.29,
                    "end": 1407.73,
                    "word": " access",
                    "probability": 0.7744140625
                },
                {
                    "start": 1407.73,
                    "end": 1407.97,
                    "word": " to",
                    "probability": 0.90185546875
                },
                {
                    "start": 1407.97,
                    "end": 1408.17,
                    "word": " wherever",
                    "probability": 0.9111328125
                }
            ]
        },
        {
            "id": 198,
            "seek": 142004,
            "start": 1408.17,
            "end": 1415.63,
            "text": " you store your data and just inject bad examples. So, our impact is misclassification. It sort of",
            "tokens": [
                50764,
                291,
                3531,
                428,
                1412,
                293,
                445,
                10711,
                1578,
                5110,
                13,
                407,
                11,
                527,
                2712,
                307,
                3346,
                11665,
                3774,
                13,
                467,
                1333,
                295,
                51140
            ],
            "temperature": 0.0,
            "avg_logprob": -0.20052082815032074,
            "compression_ratio": 1.559782608695652,
            "no_speech_prob": 0.0005359649658203125,
            "words": [
                {
                    "start": 1408.17,
                    "end": 1408.43,
                    "word": " you",
                    "probability": 0.8984375
                },
                {
                    "start": 1408.43,
                    "end": 1408.67,
                    "word": " store",
                    "probability": 0.7890625
                },
                {
                    "start": 1408.67,
                    "end": 1408.85,
                    "word": " your",
                    "probability": 0.806640625
                },
                {
                    "start": 1408.85,
                    "end": 1409.13,
                    "word": " data",
                    "probability": 0.818359375
                },
                {
                    "start": 1409.13,
                    "end": 1409.31,
                    "word": " and",
                    "probability": 0.84228515625
                },
                {
                    "start": 1409.31,
                    "end": 1409.51,
                    "word": " just",
                    "probability": 0.90380859375
                },
                {
                    "start": 1409.51,
                    "end": 1409.95,
                    "word": " inject",
                    "probability": 0.75244140625
                },
                {
                    "start": 1409.95,
                    "end": 1410.35,
                    "word": " bad",
                    "probability": 0.88525390625
                },
                {
                    "start": 1410.35,
                    "end": 1410.93,
                    "word": " examples",
                    "probability": 0.7861328125
                },
                {
                    "start": 1410.93,
                    "end": 1411.53,
                    "word": ".",
                    "probability": 0.9638671875
                },
                {
                    "start": 1411.65,
                    "end": 1411.93,
                    "word": " So",
                    "probability": 0.87939453125
                },
                {
                    "start": 1411.93,
                    "end": 1411.99,
                    "word": ",",
                    "probability": 0.2366943359375
                },
                {
                    "start": 1411.99,
                    "end": 1412.13,
                    "word": " our",
                    "probability": 0.794921875
                },
                {
                    "start": 1412.13,
                    "end": 1412.55,
                    "word": " impact",
                    "probability": 0.87451171875
                },
                {
                    "start": 1412.55,
                    "end": 1412.75,
                    "word": " is",
                    "probability": 0.9033203125
                },
                {
                    "start": 1412.75,
                    "end": 1413.85,
                    "word": " misclassification",
                    "probability": 0.9607747395833334
                },
                {
                    "start": 1413.85,
                    "end": 1414.45,
                    "word": ".",
                    "probability": 0.9365234375
                },
                {
                    "start": 1415.07,
                    "end": 1415.33,
                    "word": " It",
                    "probability": 0.884765625
                },
                {
                    "start": 1415.33,
                    "end": 1415.49,
                    "word": " sort",
                    "probability": 0.77734375
                },
                {
                    "start": 1415.49,
                    "end": 1415.63,
                    "word": " of",
                    "probability": 0.9111328125
                }
            ]
        },
        {
            "id": 199,
            "seek": 142004,
            "start": 1415.63,
            "end": 1423.77,
            "text": " shifts our decision boundary to deliberately misclassify samples. So, you know, letting spam",
            "tokens": [
                51140,
                19201,
                527,
                3537,
                12866,
                281,
                23506,
                3346,
                11665,
                2505,
                10938,
                13,
                407,
                11,
                291,
                458,
                11,
                8295,
                24028,
                51548
            ],
            "temperature": 0.0,
            "avg_logprob": -0.20052082815032074,
            "compression_ratio": 1.559782608695652,
            "no_speech_prob": 0.0005359649658203125,
            "words": [
                {
                    "start": 1415.63,
                    "end": 1416.09,
                    "word": " shifts",
                    "probability": 0.8740234375
                },
                {
                    "start": 1416.09,
                    "end": 1416.47,
                    "word": " our",
                    "probability": 0.8056640625
                },
                {
                    "start": 1416.47,
                    "end": 1416.97,
                    "word": " decision",
                    "probability": 0.857421875
                },
                {
                    "start": 1416.97,
                    "end": 1417.53,
                    "word": " boundary",
                    "probability": 0.90869140625
                },
                {
                    "start": 1417.53,
                    "end": 1418.45,
                    "word": " to",
                    "probability": 0.8974609375
                },
                {
                    "start": 1418.45,
                    "end": 1419.29,
                    "word": " deliberately",
                    "probability": 0.83056640625
                },
                {
                    "start": 1419.29,
                    "end": 1420.25,
                    "word": " misclassify",
                    "probability": 0.9251302083333334
                },
                {
                    "start": 1420.25,
                    "end": 1420.83,
                    "word": " samples",
                    "probability": 0.75732421875
                },
                {
                    "start": 1420.83,
                    "end": 1421.43,
                    "word": ".",
                    "probability": 0.84912109375
                },
                {
                    "start": 1421.61,
                    "end": 1421.73,
                    "word": " So",
                    "probability": 0.88720703125
                },
                {
                    "start": 1421.73,
                    "end": 1421.97,
                    "word": ",",
                    "probability": 0.97216796875
                },
                {
                    "start": 1421.97,
                    "end": 1422.13,
                    "word": " you",
                    "probability": 0.85888671875
                },
                {
                    "start": 1422.13,
                    "end": 1422.29,
                    "word": " know",
                    "probability": 0.82177734375
                },
                {
                    "start": 1422.29,
                    "end": 1422.85,
                    "word": ",",
                    "probability": 0.97607421875
                },
                {
                    "start": 1422.85,
                    "end": 1423.25,
                    "word": " letting",
                    "probability": 0.83349609375
                },
                {
                    "start": 1423.25,
                    "end": 1423.77,
                    "word": " spam",
                    "probability": 0.9345703125
                }
            ]
        },
        {
            "id": 200,
            "seek": 144742,
            "start": 1423.77,
            "end": 1431.47,
            "text": " through because now your threshold for what is spam is so high because you've just been doing",
            "tokens": [
                50368,
                807,
                570,
                586,
                428,
                14678,
                337,
                437,
                307,
                24028,
                307,
                370,
                1090,
                570,
                291,
                600,
                445,
                668,
                884,
                50740
            ],
            "temperature": 0.0,
            "avg_logprob": -0.19750702381134033,
            "compression_ratio": 1.6488888888888888,
            "no_speech_prob": 0.0087127685546875,
            "words": [
                {
                    "start": 1423.77,
                    "end": 1424.17,
                    "word": " through",
                    "probability": 0.2578125
                },
                {
                    "start": 1424.17,
                    "end": 1424.73,
                    "word": " because",
                    "probability": 0.71142578125
                },
                {
                    "start": 1424.73,
                    "end": 1425.73,
                    "word": " now",
                    "probability": 0.8662109375
                },
                {
                    "start": 1425.73,
                    "end": 1426.51,
                    "word": " your",
                    "probability": 0.77783203125
                },
                {
                    "start": 1426.51,
                    "end": 1427.13,
                    "word": " threshold",
                    "probability": 0.6279296875
                },
                {
                    "start": 1427.13,
                    "end": 1427.47,
                    "word": " for",
                    "probability": 0.8828125
                },
                {
                    "start": 1427.47,
                    "end": 1427.77,
                    "word": " what",
                    "probability": 0.85546875
                },
                {
                    "start": 1427.77,
                    "end": 1428.01,
                    "word": " is",
                    "probability": 0.90185546875
                },
                {
                    "start": 1428.01,
                    "end": 1428.45,
                    "word": " spam",
                    "probability": 0.9052734375
                },
                {
                    "start": 1428.45,
                    "end": 1428.73,
                    "word": " is",
                    "probability": 0.9013671875
                },
                {
                    "start": 1428.73,
                    "end": 1429.03,
                    "word": " so",
                    "probability": 0.90380859375
                },
                {
                    "start": 1429.03,
                    "end": 1429.61,
                    "word": " high",
                    "probability": 0.86669921875
                },
                {
                    "start": 1429.61,
                    "end": 1430.23,
                    "word": " because",
                    "probability": 0.70703125
                },
                {
                    "start": 1430.23,
                    "end": 1430.57,
                    "word": " you've",
                    "probability": 0.90869140625
                },
                {
                    "start": 1430.57,
                    "end": 1430.79,
                    "word": " just",
                    "probability": 0.87353515625
                },
                {
                    "start": 1430.79,
                    "end": 1431.09,
                    "word": " been",
                    "probability": 0.8095703125
                },
                {
                    "start": 1431.09,
                    "end": 1431.47,
                    "word": " doing",
                    "probability": 0.89892578125
                }
            ]
        },
        {
            "id": 201,
            "seek": 144742,
            "start": 1432.17,
            "end": 1439.59,
            "text": " online learning, right? And so, when we mitigate it, right, we want to go back and think about",
            "tokens": [
                50788,
                2950,
                2539,
                11,
                558,
                30,
                400,
                370,
                11,
                562,
                321,
                27336,
                309,
                11,
                558,
                11,
                321,
                528,
                281,
                352,
                646,
                293,
                519,
                466,
                51152
            ],
            "temperature": 0.0,
            "avg_logprob": -0.19750702381134033,
            "compression_ratio": 1.6488888888888888,
            "no_speech_prob": 0.0087127685546875,
            "words": [
                {
                    "start": 1432.17,
                    "end": 1432.77,
                    "word": " online",
                    "probability": 0.7451171875
                },
                {
                    "start": 1432.77,
                    "end": 1433.37,
                    "word": " learning",
                    "probability": 0.85400390625
                },
                {
                    "start": 1433.37,
                    "end": 1433.61,
                    "word": ",",
                    "probability": 0.415771484375
                },
                {
                    "start": 1433.61,
                    "end": 1433.93,
                    "word": " right",
                    "probability": 0.85693359375
                },
                {
                    "start": 1433.93,
                    "end": 1434.53,
                    "word": "?",
                    "probability": 0.7822265625
                },
                {
                    "start": 1435.37,
                    "end": 1435.97,
                    "word": " And",
                    "probability": 0.78662109375
                },
                {
                    "start": 1435.97,
                    "end": 1436.11,
                    "word": " so",
                    "probability": 0.8955078125
                },
                {
                    "start": 1436.11,
                    "end": 1436.23,
                    "word": ",",
                    "probability": 0.12005615234375
                },
                {
                    "start": 1436.23,
                    "end": 1436.29,
                    "word": " when",
                    "probability": 0.859375
                },
                {
                    "start": 1436.29,
                    "end": 1436.43,
                    "word": " we",
                    "probability": 0.90283203125
                },
                {
                    "start": 1436.43,
                    "end": 1436.81,
                    "word": " mitigate",
                    "probability": 0.85546875
                },
                {
                    "start": 1436.81,
                    "end": 1437.39,
                    "word": " it",
                    "probability": 0.90283203125
                },
                {
                    "start": 1437.39,
                    "end": 1437.67,
                    "word": ",",
                    "probability": 0.9599609375
                },
                {
                    "start": 1437.67,
                    "end": 1437.89,
                    "word": " right",
                    "probability": 0.62451171875
                },
                {
                    "start": 1437.89,
                    "end": 1437.93,
                    "word": ",",
                    "probability": 0.75830078125
                },
                {
                    "start": 1437.93,
                    "end": 1438.17,
                    "word": " we",
                    "probability": 0.90673828125
                },
                {
                    "start": 1438.17,
                    "end": 1438.35,
                    "word": " want",
                    "probability": 0.73486328125
                },
                {
                    "start": 1438.35,
                    "end": 1438.45,
                    "word": " to",
                    "probability": 0.9091796875
                },
                {
                    "start": 1438.45,
                    "end": 1438.61,
                    "word": " go",
                    "probability": 0.90576171875
                },
                {
                    "start": 1438.61,
                    "end": 1438.95,
                    "word": " back",
                    "probability": 0.822265625
                },
                {
                    "start": 1438.95,
                    "end": 1439.13,
                    "word": " and",
                    "probability": 0.896484375
                },
                {
                    "start": 1439.13,
                    "end": 1439.31,
                    "word": " think",
                    "probability": 0.81689453125
                },
                {
                    "start": 1439.31,
                    "end": 1439.59,
                    "word": " about",
                    "probability": 0.82470703125
                }
            ]
        },
        {
            "id": 202,
            "seek": 144742,
            "start": 1439.59,
            "end": 1446.27,
            "text": " protecting access to our data. So, we want data integrity, right? So, having data versioning",
            "tokens": [
                51152,
                12316,
                2105,
                281,
                527,
                1412,
                13,
                407,
                11,
                321,
                528,
                1412,
                16000,
                11,
                558,
                30,
                407,
                11,
                1419,
                1412,
                3037,
                278,
                51492
            ],
            "temperature": 0.0,
            "avg_logprob": -0.19750702381134033,
            "compression_ratio": 1.6488888888888888,
            "no_speech_prob": 0.0087127685546875,
            "words": [
                {
                    "start": 1439.59,
                    "end": 1440.07,
                    "word": " protecting",
                    "probability": 0.7236328125
                },
                {
                    "start": 1440.07,
                    "end": 1440.57,
                    "word": " access",
                    "probability": 0.79052734375
                },
                {
                    "start": 1440.57,
                    "end": 1440.79,
                    "word": " to",
                    "probability": 0.8974609375
                },
                {
                    "start": 1440.79,
                    "end": 1440.93,
                    "word": " our",
                    "probability": 0.818359375
                },
                {
                    "start": 1440.93,
                    "end": 1441.27,
                    "word": " data",
                    "probability": 0.810546875
                },
                {
                    "start": 1441.27,
                    "end": 1441.67,
                    "word": ".",
                    "probability": 0.8505859375
                },
                {
                    "start": 1441.67,
                    "end": 1441.77,
                    "word": " So",
                    "probability": 0.88916015625
                },
                {
                    "start": 1441.77,
                    "end": 1441.85,
                    "word": ",",
                    "probability": 0.74169921875
                },
                {
                    "start": 1441.85,
                    "end": 1441.93,
                    "word": " we",
                    "probability": 0.9111328125
                },
                {
                    "start": 1441.93,
                    "end": 1442.17,
                    "word": " want",
                    "probability": 0.82080078125
                },
                {
                    "start": 1442.17,
                    "end": 1442.57,
                    "word": " data",
                    "probability": 0.81640625
                },
                {
                    "start": 1442.57,
                    "end": 1443.37,
                    "word": " integrity",
                    "probability": 0.71875
                },
                {
                    "start": 1443.37,
                    "end": 1443.61,
                    "word": ",",
                    "probability": 0.9306640625
                },
                {
                    "start": 1443.61,
                    "end": 1443.87,
                    "word": " right",
                    "probability": 0.87353515625
                },
                {
                    "start": 1443.87,
                    "end": 1444.33,
                    "word": "?",
                    "probability": 0.94482421875
                },
                {
                    "start": 1444.33,
                    "end": 1444.57,
                    "word": " So",
                    "probability": 0.90087890625
                },
                {
                    "start": 1444.57,
                    "end": 1444.63,
                    "word": ",",
                    "probability": 0.9228515625
                },
                {
                    "start": 1444.63,
                    "end": 1444.95,
                    "word": " having",
                    "probability": 0.814453125
                },
                {
                    "start": 1444.95,
                    "end": 1445.51,
                    "word": " data",
                    "probability": 0.82861328125
                },
                {
                    "start": 1445.51,
                    "end": 1446.27,
                    "word": " versioning",
                    "probability": 0.845458984375
                }
            ]
        },
        {
            "id": 203,
            "seek": 144742,
            "start": 1446.27,
            "end": 1451.15,
            "text": " is really important. Making sure that your training data has a version, has an associated",
            "tokens": [
                51492,
                307,
                534,
                1021,
                13,
                14595,
                988,
                300,
                428,
                3097,
                1412,
                575,
                257,
                3037,
                11,
                575,
                364,
                6615,
                51740
            ],
            "temperature": 0.0,
            "avg_logprob": -0.19750702381134033,
            "compression_ratio": 1.6488888888888888,
            "no_speech_prob": 0.0087127685546875,
            "words": [
                {
                    "start": 1446.27,
                    "end": 1446.59,
                    "word": " is",
                    "probability": 0.9033203125
                },
                {
                    "start": 1446.59,
                    "end": 1446.85,
                    "word": " really",
                    "probability": 0.77734375
                },
                {
                    "start": 1446.85,
                    "end": 1447.31,
                    "word": " important",
                    "probability": 0.765625
                },
                {
                    "start": 1447.31,
                    "end": 1447.57,
                    "word": ".",
                    "probability": 0.3173828125
                },
                {
                    "start": 1447.57,
                    "end": 1447.77,
                    "word": " Making",
                    "probability": 0.83447265625
                },
                {
                    "start": 1447.77,
                    "end": 1448.01,
                    "word": " sure",
                    "probability": 0.8134765625
                },
                {
                    "start": 1448.01,
                    "end": 1448.15,
                    "word": " that",
                    "probability": 0.8623046875
                },
                {
                    "start": 1448.15,
                    "end": 1448.27,
                    "word": " your",
                    "probability": 0.71240234375
                },
                {
                    "start": 1448.27,
                    "end": 1448.63,
                    "word": " training",
                    "probability": 0.802734375
                },
                {
                    "start": 1448.63,
                    "end": 1448.97,
                    "word": " data",
                    "probability": 0.82080078125
                },
                {
                    "start": 1448.97,
                    "end": 1449.21,
                    "word": " has",
                    "probability": 0.9013671875
                },
                {
                    "start": 1449.35,
                    "end": 1449.75,
                    "word": " a version",
                    "probability": 0.6982421875
                },
                {
                    "start": 1449.75,
                    "end": 1450.27,
                    "word": ",",
                    "probability": 0.916015625
                },
                {
                    "start": 1450.27,
                    "end": 1450.49,
                    "word": " has",
                    "probability": 0.8916015625
                },
                {
                    "start": 1450.49,
                    "end": 1450.63,
                    "word": " an",
                    "probability": 0.904296875
                },
                {
                    "start": 1450.63,
                    "end": 1451.15,
                    "word": " associated",
                    "probability": 0.82861328125
                }
            ]
        },
        {
            "id": 204,
            "seek": 147654,
            "start": 1451.15,
            "end": 1459.35,
            "text": " hash, has a label. That's super important. And then our next is we don't want to just allow",
            "tokens": [
                50372,
                22019,
                11,
                575,
                257,
                7645,
                13,
                663,
                311,
                1687,
                1021,
                13,
                400,
                550,
                527,
                958,
                307,
                321,
                500,
                380,
                528,
                281,
                445,
                2089,
                50784
            ],
            "temperature": 0.0,
            "avg_logprob": -0.1951462821440494,
            "compression_ratio": 1.64,
            "no_speech_prob": 0.00048041343688964844,
            "words": [
                {
                    "start": 1451.15,
                    "end": 1451.79,
                    "word": " hash",
                    "probability": 0.2496337890625
                },
                {
                    "start": 1451.79,
                    "end": 1451.95,
                    "word": ",",
                    "probability": 0.0660400390625
                },
                {
                    "start": 1451.95,
                    "end": 1452.05,
                    "word": " has",
                    "probability": 0.587890625
                },
                {
                    "start": 1452.23,
                    "end": 1452.61,
                    "word": " a label",
                    "probability": 0.8212890625
                },
                {
                    "start": 1452.61,
                    "end": 1453.25,
                    "word": ".",
                    "probability": 0.37548828125
                },
                {
                    "start": 1453.35,
                    "end": 1453.75,
                    "word": " That's",
                    "probability": 0.886474609375
                },
                {
                    "start": 1453.75,
                    "end": 1454.05,
                    "word": " super",
                    "probability": 0.88623046875
                },
                {
                    "start": 1454.05,
                    "end": 1454.65,
                    "word": " important",
                    "probability": 0.7431640625
                },
                {
                    "start": 1454.65,
                    "end": 1455.29,
                    "word": ".",
                    "probability": 0.9287109375
                },
                {
                    "start": 1455.87,
                    "end": 1455.89,
                    "word": " And",
                    "probability": 0.79345703125
                },
                {
                    "start": 1455.89,
                    "end": 1456.01,
                    "word": " then",
                    "probability": 0.6943359375
                },
                {
                    "start": 1456.01,
                    "end": 1456.15,
                    "word": " our",
                    "probability": 0.7783203125
                },
                {
                    "start": 1456.15,
                    "end": 1456.49,
                    "word": " next",
                    "probability": 0.86376953125
                },
                {
                    "start": 1456.49,
                    "end": 1456.75,
                    "word": " is",
                    "probability": 0.88818359375
                },
                {
                    "start": 1456.75,
                    "end": 1457.33,
                    "word": " we",
                    "probability": 0.64794921875
                },
                {
                    "start": 1457.33,
                    "end": 1457.55,
                    "word": " don't",
                    "probability": 0.945556640625
                },
                {
                    "start": 1457.55,
                    "end": 1458.21,
                    "word": " want",
                    "probability": 0.8125
                },
                {
                    "start": 1458.21,
                    "end": 1458.37,
                    "word": " to",
                    "probability": 0.90869140625
                },
                {
                    "start": 1458.37,
                    "end": 1458.71,
                    "word": " just",
                    "probability": 0.849609375
                },
                {
                    "start": 1458.71,
                    "end": 1459.35,
                    "word": " allow",
                    "probability": 0.8759765625
                }
            ]
        },
        {
            "id": 205,
            "seek": 147654,
            "start": 1460.03,
            "end": 1466.25,
            "text": " unvalidated, untrusted input to our model. That's something we see a ton is people who just",
            "tokens": [
                50784,
                517,
                3337,
                327,
                770,
                11,
                1701,
                81,
                6589,
                4846,
                281,
                527,
                2316,
                13,
                663,
                311,
                746,
                321,
                536,
                257,
                2952,
                307,
                561,
                567,
                445,
                51124
            ],
            "temperature": 0.0,
            "avg_logprob": -0.1951462821440494,
            "compression_ratio": 1.64,
            "no_speech_prob": 0.00048041343688964844,
            "words": [
                {
                    "start": 1460.03,
                    "end": 1460.67,
                    "word": " unvalidated",
                    "probability": 0.9296875
                },
                {
                    "start": 1460.67,
                    "end": 1460.83,
                    "word": ",",
                    "probability": 0.7041015625
                },
                {
                    "start": 1460.83,
                    "end": 1461.37,
                    "word": " untrusted",
                    "probability": 0.91015625
                },
                {
                    "start": 1461.37,
                    "end": 1461.77,
                    "word": " input",
                    "probability": 0.86669921875
                },
                {
                    "start": 1461.77,
                    "end": 1462.07,
                    "word": " to",
                    "probability": 0.89453125
                },
                {
                    "start": 1462.07,
                    "end": 1462.25,
                    "word": " our",
                    "probability": 0.822265625
                },
                {
                    "start": 1462.25,
                    "end": 1462.59,
                    "word": " model",
                    "probability": 0.80615234375
                },
                {
                    "start": 1462.59,
                    "end": 1462.83,
                    "word": ".",
                    "probability": 0.97021484375
                },
                {
                    "start": 1462.83,
                    "end": 1463.01,
                    "word": " That's",
                    "probability": 0.9013671875
                },
                {
                    "start": 1463.01,
                    "end": 1463.21,
                    "word": " something",
                    "probability": 0.75048828125
                },
                {
                    "start": 1463.21,
                    "end": 1463.41,
                    "word": " we",
                    "probability": 0.8935546875
                },
                {
                    "start": 1463.41,
                    "end": 1463.59,
                    "word": " see",
                    "probability": 0.81787109375
                },
                {
                    "start": 1463.79,
                    "end": 1464.33,
                    "word": " a ton",
                    "probability": 0.89794921875
                },
                {
                    "start": 1464.33,
                    "end": 1465.29,
                    "word": " is",
                    "probability": 0.35546875
                },
                {
                    "start": 1465.29,
                    "end": 1465.67,
                    "word": " people",
                    "probability": 0.91357421875
                },
                {
                    "start": 1465.67,
                    "end": 1465.95,
                    "word": " who",
                    "probability": 0.81103515625
                },
                {
                    "start": 1465.95,
                    "end": 1466.25,
                    "word": " just",
                    "probability": 0.87890625
                }
            ]
        },
        {
            "id": 206,
            "seek": 147654,
            "start": 1466.25,
            "end": 1472.75,
            "text": " accept whatever input with no validation on the front end. It's just whatever gets input to the",
            "tokens": [
                51124,
                3241,
                2035,
                4846,
                365,
                572,
                24071,
                322,
                264,
                1868,
                917,
                13,
                467,
                311,
                445,
                2035,
                2170,
                4846,
                281,
                264,
                51440
            ],
            "temperature": 0.0,
            "avg_logprob": -0.1951462821440494,
            "compression_ratio": 1.64,
            "no_speech_prob": 0.00048041343688964844,
            "words": [
                {
                    "start": 1466.25,
                    "end": 1466.97,
                    "word": " accept",
                    "probability": 0.70263671875
                },
                {
                    "start": 1466.97,
                    "end": 1467.63,
                    "word": " whatever",
                    "probability": 0.7900390625
                },
                {
                    "start": 1467.63,
                    "end": 1468.27,
                    "word": " input",
                    "probability": 0.83447265625
                },
                {
                    "start": 1468.27,
                    "end": 1468.77,
                    "word": " with",
                    "probability": 0.78759765625
                },
                {
                    "start": 1468.77,
                    "end": 1469.13,
                    "word": " no",
                    "probability": 0.90576171875
                },
                {
                    "start": 1469.13,
                    "end": 1469.63,
                    "word": " validation",
                    "probability": 0.84619140625
                },
                {
                    "start": 1469.63,
                    "end": 1469.85,
                    "word": " on",
                    "probability": 0.8994140625
                },
                {
                    "start": 1469.85,
                    "end": 1469.97,
                    "word": " the",
                    "probability": 0.81591796875
                },
                {
                    "start": 1469.97,
                    "end": 1470.25,
                    "word": " front",
                    "probability": 0.91748046875
                },
                {
                    "start": 1470.25,
                    "end": 1470.51,
                    "word": " end",
                    "probability": 0.5595703125
                },
                {
                    "start": 1470.51,
                    "end": 1470.69,
                    "word": ".",
                    "probability": 0.7294921875
                },
                {
                    "start": 1470.69,
                    "end": 1470.87,
                    "word": " It's",
                    "probability": 0.94677734375
                },
                {
                    "start": 1470.87,
                    "end": 1471.15,
                    "word": " just",
                    "probability": 0.88671875
                },
                {
                    "start": 1471.15,
                    "end": 1471.73,
                    "word": " whatever",
                    "probability": 0.77587890625
                },
                {
                    "start": 1471.73,
                    "end": 1472.01,
                    "word": " gets",
                    "probability": 0.79638671875
                },
                {
                    "start": 1472.01,
                    "end": 1472.41,
                    "word": " input",
                    "probability": 0.884765625
                },
                {
                    "start": 1472.41,
                    "end": 1472.59,
                    "word": " to",
                    "probability": 0.900390625
                },
                {
                    "start": 1472.59,
                    "end": 1472.75,
                    "word": " the",
                    "probability": 0.8154296875
                }
            ]
        },
        {
            "id": 207,
            "seek": 147654,
            "start": 1472.75,
            "end": 1480.27,
            "text": " API gets pushed into the model. And that can really cause problems when you're retraining",
            "tokens": [
                51440,
                9362,
                2170,
                9152,
                666,
                264,
                2316,
                13,
                400,
                300,
                393,
                534,
                3082,
                2740,
                562,
                291,
                434,
                49356,
                1760,
                51836
            ],
            "temperature": 0.0,
            "avg_logprob": -0.1951462821440494,
            "compression_ratio": 1.64,
            "no_speech_prob": 0.00048041343688964844,
            "words": [
                {
                    "start": 1472.75,
                    "end": 1473.15,
                    "word": " API",
                    "probability": 0.81884765625
                },
                {
                    "start": 1473.15,
                    "end": 1473.51,
                    "word": " gets",
                    "probability": 0.744140625
                },
                {
                    "start": 1473.51,
                    "end": 1473.93,
                    "word": " pushed",
                    "probability": 0.73486328125
                },
                {
                    "start": 1473.93,
                    "end": 1474.19,
                    "word": " into",
                    "probability": 0.7119140625
                },
                {
                    "start": 1474.19,
                    "end": 1474.41,
                    "word": " the",
                    "probability": 0.8125
                },
                {
                    "start": 1474.41,
                    "end": 1474.73,
                    "word": " model",
                    "probability": 0.8203125
                },
                {
                    "start": 1474.73,
                    "end": 1475.37,
                    "word": ".",
                    "probability": 0.82470703125
                },
                {
                    "start": 1476.13,
                    "end": 1476.77,
                    "word": " And",
                    "probability": 0.8642578125
                },
                {
                    "start": 1476.77,
                    "end": 1477.09,
                    "word": " that",
                    "probability": 0.87744140625
                },
                {
                    "start": 1477.09,
                    "end": 1477.45,
                    "word": " can",
                    "probability": 0.9033203125
                },
                {
                    "start": 1477.45,
                    "end": 1478.23,
                    "word": " really",
                    "probability": 0.7685546875
                },
                {
                    "start": 1478.23,
                    "end": 1478.73,
                    "word": " cause",
                    "probability": 0.8115234375
                },
                {
                    "start": 1478.73,
                    "end": 1479.21,
                    "word": " problems",
                    "probability": 0.6953125
                },
                {
                    "start": 1479.21,
                    "end": 1479.53,
                    "word": " when",
                    "probability": 0.87451171875
                },
                {
                    "start": 1479.53,
                    "end": 1479.73,
                    "word": " you're",
                    "probability": 0.92333984375
                },
                {
                    "start": 1479.73,
                    "end": 1480.27,
                    "word": " retraining",
                    "probability": 0.8583984375
                }
            ]
        },
        {
            "id": 208,
            "seek": 150348,
            "start": 1480.27,
            "end": 1489.43,
            "text": " if you just take, you know, those inputs that you get and sort of store them off for training later.",
            "tokens": [
                50376,
                498,
                291,
                445,
                747,
                11,
                291,
                458,
                11,
                729,
                15743,
                300,
                291,
                483,
                293,
                1333,
                295,
                3531,
                552,
                766,
                337,
                3097,
                1780,
                13,
                50800
            ],
            "temperature": 0.0,
            "avg_logprob": -0.18078270582395178,
            "compression_ratio": 1.6444444444444444,
            "no_speech_prob": 0.00016605854034423828,
            "words": [
                {
                    "start": 1480.27,
                    "end": 1480.77,
                    "word": " if",
                    "probability": 0.63671875
                },
                {
                    "start": 1480.77,
                    "end": 1480.91,
                    "word": " you",
                    "probability": 0.896484375
                },
                {
                    "start": 1480.91,
                    "end": 1481.39,
                    "word": " just",
                    "probability": 0.88037109375
                },
                {
                    "start": 1481.39,
                    "end": 1482.27,
                    "word": " take",
                    "probability": 0.8017578125
                },
                {
                    "start": 1482.27,
                    "end": 1482.49,
                    "word": ",",
                    "probability": 0.330810546875
                },
                {
                    "start": 1482.49,
                    "end": 1483.21,
                    "word": " you",
                    "probability": 0.80419921875
                },
                {
                    "start": 1483.21,
                    "end": 1483.37,
                    "word": " know",
                    "probability": 0.82080078125
                },
                {
                    "start": 1483.37,
                    "end": 1483.51,
                    "word": ",",
                    "probability": 0.92333984375
                },
                {
                    "start": 1483.51,
                    "end": 1484.67,
                    "word": " those",
                    "probability": 0.84716796875
                },
                {
                    "start": 1484.67,
                    "end": 1485.29,
                    "word": " inputs",
                    "probability": 0.7734375
                },
                {
                    "start": 1485.29,
                    "end": 1485.51,
                    "word": " that",
                    "probability": 0.87744140625
                },
                {
                    "start": 1485.51,
                    "end": 1485.65,
                    "word": " you",
                    "probability": 0.89599609375
                },
                {
                    "start": 1485.65,
                    "end": 1485.91,
                    "word": " get",
                    "probability": 0.89892578125
                },
                {
                    "start": 1485.91,
                    "end": 1486.11,
                    "word": " and",
                    "probability": 0.830078125
                },
                {
                    "start": 1486.11,
                    "end": 1486.29,
                    "word": " sort",
                    "probability": 0.802734375
                },
                {
                    "start": 1486.29,
                    "end": 1486.39,
                    "word": " of",
                    "probability": 0.90625
                },
                {
                    "start": 1486.39,
                    "end": 1486.75,
                    "word": " store",
                    "probability": 0.80859375
                },
                {
                    "start": 1486.75,
                    "end": 1486.95,
                    "word": " them",
                    "probability": 0.72705078125
                },
                {
                    "start": 1486.95,
                    "end": 1487.35,
                    "word": " off",
                    "probability": 0.75048828125
                },
                {
                    "start": 1487.35,
                    "end": 1487.65,
                    "word": " for",
                    "probability": 0.87158203125
                },
                {
                    "start": 1487.65,
                    "end": 1488.57,
                    "word": " training",
                    "probability": 0.68017578125
                },
                {
                    "start": 1488.57,
                    "end": 1489.03,
                    "word": " later",
                    "probability": 0.87841796875
                },
                {
                    "start": 1489.03,
                    "end": 1489.43,
                    "word": ".",
                    "probability": 0.95947265625
                }
            ]
        },
        {
            "id": 209,
            "seek": 150348,
            "start": 1490.69,
            "end": 1497.15,
            "text": " So, you know, we're coming up on the end of this talk. And so, we want to talk about,",
            "tokens": [
                50880,
                407,
                11,
                291,
                458,
                11,
                321,
                434,
                1348,
                493,
                322,
                264,
                917,
                295,
                341,
                751,
                13,
                400,
                370,
                11,
                321,
                528,
                281,
                751,
                466,
                11,
                51196
            ],
            "temperature": 0.0,
            "avg_logprob": -0.18078270582395178,
            "compression_ratio": 1.6444444444444444,
            "no_speech_prob": 0.00016605854034423828,
            "words": [
                {
                    "start": 1490.69,
                    "end": 1491.09,
                    "word": " So",
                    "probability": 0.87109375
                },
                {
                    "start": 1491.09,
                    "end": 1491.49,
                    "word": ",",
                    "probability": 0.93310546875
                },
                {
                    "start": 1491.49,
                    "end": 1491.83,
                    "word": " you",
                    "probability": 0.84765625
                },
                {
                    "start": 1491.83,
                    "end": 1491.93,
                    "word": " know",
                    "probability": 0.81689453125
                },
                {
                    "start": 1491.93,
                    "end": 1492.09,
                    "word": ",",
                    "probability": 0.97900390625
                },
                {
                    "start": 1492.09,
                    "end": 1492.75,
                    "word": " we're",
                    "probability": 0.9287109375
                },
                {
                    "start": 1492.75,
                    "end": 1492.93,
                    "word": " coming",
                    "probability": 0.81201171875
                },
                {
                    "start": 1492.93,
                    "end": 1493.11,
                    "word": " up",
                    "probability": 0.90869140625
                },
                {
                    "start": 1493.11,
                    "end": 1493.25,
                    "word": " on",
                    "probability": 0.90576171875
                },
                {
                    "start": 1493.25,
                    "end": 1493.37,
                    "word": " the",
                    "probability": 0.81591796875
                },
                {
                    "start": 1493.37,
                    "end": 1493.53,
                    "word": " end",
                    "probability": 0.79931640625
                },
                {
                    "start": 1493.53,
                    "end": 1493.63,
                    "word": " of",
                    "probability": 0.90087890625
                },
                {
                    "start": 1493.63,
                    "end": 1493.85,
                    "word": " this",
                    "probability": 0.88134765625
                },
                {
                    "start": 1493.85,
                    "end": 1494.33,
                    "word": " talk",
                    "probability": 0.8203125
                },
                {
                    "start": 1494.33,
                    "end": 1494.65,
                    "word": ".",
                    "probability": 0.486572265625
                },
                {
                    "start": 1494.65,
                    "end": 1494.85,
                    "word": " And",
                    "probability": 0.8212890625
                },
                {
                    "start": 1494.85,
                    "end": 1496.07,
                    "word": " so",
                    "probability": 0.85986328125
                },
                {
                    "start": 1496.07,
                    "end": 1496.31,
                    "word": ",",
                    "probability": 0.072998046875
                },
                {
                    "start": 1496.31,
                    "end": 1496.55,
                    "word": " we",
                    "probability": 0.8994140625
                },
                {
                    "start": 1496.55,
                    "end": 1496.69,
                    "word": " want",
                    "probability": 0.45849609375
                },
                {
                    "start": 1496.69,
                    "end": 1496.77,
                    "word": " to",
                    "probability": 0.90869140625
                },
                {
                    "start": 1496.77,
                    "end": 1496.93,
                    "word": " talk",
                    "probability": 0.81884765625
                },
                {
                    "start": 1496.93,
                    "end": 1497.07,
                    "word": " about",
                    "probability": 0.81787109375
                },
                {
                    "start": 1497.07,
                    "end": 1497.15,
                    "word": ",",
                    "probability": 0.2315673828125
                }
            ]
        },
        {
            "id": 210,
            "seek": 150348,
            "start": 1497.15,
            "end": 1501.37,
            "text": " like, what can I do, right? And this is sort of a choose your own adventure story. It depends",
            "tokens": [
                51196,
                411,
                11,
                437,
                393,
                286,
                360,
                11,
                558,
                30,
                400,
                341,
                307,
                1333,
                295,
                257,
                2826,
                428,
                1065,
                9868,
                1657,
                13,
                467,
                5946,
                51416
            ],
            "temperature": 0.0,
            "avg_logprob": -0.18078270582395178,
            "compression_ratio": 1.6444444444444444,
            "no_speech_prob": 0.00016605854034423828,
            "words": [
                {
                    "start": 1497.15,
                    "end": 1497.21,
                    "word": " like",
                    "probability": 0.82958984375
                },
                {
                    "start": 1497.21,
                    "end": 1497.25,
                    "word": ",",
                    "probability": 0.92578125
                },
                {
                    "start": 1497.25,
                    "end": 1497.33,
                    "word": " what",
                    "probability": 0.8642578125
                },
                {
                    "start": 1497.33,
                    "end": 1497.47,
                    "word": " can",
                    "probability": 0.8955078125
                },
                {
                    "start": 1497.47,
                    "end": 1497.71,
                    "word": " I",
                    "probability": 0.99951171875
                },
                {
                    "start": 1497.71,
                    "end": 1498.01,
                    "word": " do",
                    "probability": 0.9091796875
                },
                {
                    "start": 1498.01,
                    "end": 1498.27,
                    "word": ",",
                    "probability": 0.66015625
                },
                {
                    "start": 1498.27,
                    "end": 1498.51,
                    "word": " right",
                    "probability": 0.876953125
                },
                {
                    "start": 1498.51,
                    "end": 1498.71,
                    "word": "?",
                    "probability": 0.9189453125
                },
                {
                    "start": 1498.71,
                    "end": 1498.99,
                    "word": " And",
                    "probability": 0.89306640625
                },
                {
                    "start": 1498.99,
                    "end": 1499.15,
                    "word": " this",
                    "probability": 0.87890625
                },
                {
                    "start": 1499.15,
                    "end": 1499.29,
                    "word": " is",
                    "probability": 0.912109375
                },
                {
                    "start": 1499.29,
                    "end": 1499.45,
                    "word": " sort",
                    "probability": 0.80029296875
                },
                {
                    "start": 1499.45,
                    "end": 1499.53,
                    "word": " of",
                    "probability": 0.904296875
                },
                {
                    "start": 1499.67,
                    "end": 1499.93,
                    "word": " a choose",
                    "probability": 0.76953125
                },
                {
                    "start": 1499.93,
                    "end": 1500.07,
                    "word": " your",
                    "probability": 0.6494140625
                },
                {
                    "start": 1500.07,
                    "end": 1500.21,
                    "word": " own",
                    "probability": 0.81787109375
                },
                {
                    "start": 1500.21,
                    "end": 1500.57,
                    "word": " adventure",
                    "probability": 0.712890625
                },
                {
                    "start": 1500.57,
                    "end": 1500.93,
                    "word": " story",
                    "probability": 0.8974609375
                },
                {
                    "start": 1500.93,
                    "end": 1501.05,
                    "word": ".",
                    "probability": 0.83349609375
                },
                {
                    "start": 1501.05,
                    "end": 1501.09,
                    "word": " It",
                    "probability": 0.896484375
                },
                {
                    "start": 1501.09,
                    "end": 1501.37,
                    "word": " depends",
                    "probability": 0.763671875
                }
            ]
        },
        {
            "id": 211,
            "seek": 150348,
            "start": 1501.37,
            "end": 1507.21,
            "text": " on your role. So, if you're a hacker, and by hacker, I mean this in the colloquial sense,",
            "tokens": [
                51416,
                322,
                428,
                3090,
                13,
                407,
                11,
                498,
                291,
                434,
                257,
                38155,
                11,
                293,
                538,
                38155,
                11,
                286,
                914,
                341,
                294,
                264,
                1263,
                29826,
                831,
                2020,
                11,
                51704
            ],
            "temperature": 0.0,
            "avg_logprob": -0.18078270582395178,
            "compression_ratio": 1.6444444444444444,
            "no_speech_prob": 0.00016605854034423828,
            "words": [
                {
                    "start": 1501.37,
                    "end": 1501.51,
                    "word": " on",
                    "probability": 0.90673828125
                },
                {
                    "start": 1501.51,
                    "end": 1501.65,
                    "word": " your",
                    "probability": 0.80224609375
                },
                {
                    "start": 1501.65,
                    "end": 1502.01,
                    "word": " role",
                    "probability": 0.77197265625
                },
                {
                    "start": 1502.01,
                    "end": 1502.41,
                    "word": ".",
                    "probability": 0.9990234375
                },
                {
                    "start": 1502.83,
                    "end": 1503.09,
                    "word": " So",
                    "probability": 0.8994140625
                },
                {
                    "start": 1503.09,
                    "end": 1503.23,
                    "word": ",",
                    "probability": 0.483154296875
                },
                {
                    "start": 1503.23,
                    "end": 1503.29,
                    "word": " if",
                    "probability": 0.90673828125
                },
                {
                    "start": 1503.29,
                    "end": 1503.51,
                    "word": " you're",
                    "probability": 0.923828125
                },
                {
                    "start": 1503.75,
                    "end": 1504.11,
                    "word": " a hacker",
                    "probability": 0.6845703125
                },
                {
                    "start": 1504.11,
                    "end": 1504.55,
                    "word": ",",
                    "probability": 0.68994140625
                },
                {
                    "start": 1504.55,
                    "end": 1504.95,
                    "word": " and",
                    "probability": 0.8974609375
                },
                {
                    "start": 1504.95,
                    "end": 1505.13,
                    "word": " by",
                    "probability": 0.90283203125
                },
                {
                    "start": 1505.13,
                    "end": 1505.49,
                    "word": " hacker",
                    "probability": 0.71484375
                },
                {
                    "start": 1505.49,
                    "end": 1505.63,
                    "word": ",",
                    "probability": 0.759765625
                },
                {
                    "start": 1505.63,
                    "end": 1505.75,
                    "word": " I",
                    "probability": 0.9990234375
                },
                {
                    "start": 1505.75,
                    "end": 1505.87,
                    "word": " mean",
                    "probability": 0.92919921875
                },
                {
                    "start": 1505.87,
                    "end": 1506.05,
                    "word": " this",
                    "probability": 0.77490234375
                },
                {
                    "start": 1506.05,
                    "end": 1506.15,
                    "word": " in",
                    "probability": 0.880859375
                },
                {
                    "start": 1506.15,
                    "end": 1506.27,
                    "word": " the",
                    "probability": 0.8037109375
                },
                {
                    "start": 1506.27,
                    "end": 1506.81,
                    "word": " colloquial",
                    "probability": 0.923828125
                },
                {
                    "start": 1506.81,
                    "end": 1507.13,
                    "word": " sense",
                    "probability": 0.6787109375
                },
                {
                    "start": 1507.13,
                    "end": 1507.21,
                    "word": ",",
                    "probability": 0.88037109375
                }
            ]
        },
        {
            "id": 212,
            "seek": 153144,
            "start": 1507.37,
            "end": 1514.29,
            "text": " a black hat, or a red teamer, right? Black hats are already doing some of this stuff.",
            "tokens": [
                50364,
                257,
                2211,
                2385,
                11,
                420,
                257,
                2182,
                1469,
                260,
                11,
                558,
                30,
                4076,
                20549,
                366,
                1217,
                884,
                512,
                295,
                341,
                1507,
                13,
                50684
            ],
            "temperature": 0.0,
            "avg_logprob": -0.20984543074843703,
            "compression_ratio": 1.5991189427312775,
            "no_speech_prob": 0.00243377685546875,
            "words": [
                {
                    "start": 1507.37,
                    "end": 1507.73,
                    "word": " a black",
                    "probability": 0.53076171875
                },
                {
                    "start": 1507.73,
                    "end": 1508.05,
                    "word": " hat",
                    "probability": 0.8017578125
                },
                {
                    "start": 1508.05,
                    "end": 1508.51,
                    "word": ",",
                    "probability": 0.2548828125
                },
                {
                    "start": 1508.51,
                    "end": 1509.37,
                    "word": " or",
                    "probability": 0.85888671875
                },
                {
                    "start": 1509.77,
                    "end": 1509.99,
                    "word": " a red",
                    "probability": 0.81396484375
                },
                {
                    "start": 1509.99,
                    "end": 1510.47,
                    "word": " teamer",
                    "probability": 0.7724609375
                },
                {
                    "start": 1510.47,
                    "end": 1510.53,
                    "word": ",",
                    "probability": 0.305419921875
                },
                {
                    "start": 1510.53,
                    "end": 1510.75,
                    "word": " right",
                    "probability": 0.79833984375
                },
                {
                    "start": 1510.75,
                    "end": 1510.85,
                    "word": "?",
                    "probability": 0.8203125
                },
                {
                    "start": 1510.85,
                    "end": 1511.05,
                    "word": " Black",
                    "probability": 0.8984375
                },
                {
                    "start": 1511.05,
                    "end": 1511.25,
                    "word": " hats",
                    "probability": 0.80029296875
                },
                {
                    "start": 1511.25,
                    "end": 1511.39,
                    "word": " are",
                    "probability": 0.88623046875
                },
                {
                    "start": 1511.39,
                    "end": 1511.73,
                    "word": " already",
                    "probability": 0.82861328125
                },
                {
                    "start": 1511.73,
                    "end": 1512.53,
                    "word": " doing",
                    "probability": 0.90380859375
                },
                {
                    "start": 1512.53,
                    "end": 1512.89,
                    "word": " some",
                    "probability": 0.81640625
                },
                {
                    "start": 1512.89,
                    "end": 1513.07,
                    "word": " of",
                    "probability": 0.90771484375
                },
                {
                    "start": 1513.07,
                    "end": 1513.29,
                    "word": " this",
                    "probability": 0.87646484375
                },
                {
                    "start": 1513.29,
                    "end": 1513.71,
                    "word": " stuff",
                    "probability": 0.80029296875
                },
                {
                    "start": 1513.71,
                    "end": 1514.29,
                    "word": ".",
                    "probability": 0.75390625
                }
            ]
        },
        {
            "id": 213,
            "seek": 153144,
            "start": 1514.69,
            "end": 1519.51,
            "text": " So, I really mean, you know, if you're a red teamer, if you're catching up to the elite",
            "tokens": [
                50736,
                407,
                11,
                286,
                534,
                914,
                11,
                291,
                458,
                11,
                498,
                291,
                434,
                257,
                2182,
                1469,
                260,
                11,
                498,
                291,
                434,
                16124,
                493,
                281,
                264,
                17801,
                50972
            ],
            "temperature": 0.0,
            "avg_logprob": -0.20984543074843703,
            "compression_ratio": 1.5991189427312775,
            "no_speech_prob": 0.00243377685546875,
            "words": [
                {
                    "start": 1514.69,
                    "end": 1514.91,
                    "word": " So",
                    "probability": 0.81005859375
                },
                {
                    "start": 1514.91,
                    "end": 1514.99,
                    "word": ",",
                    "probability": 0.272705078125
                },
                {
                    "start": 1514.99,
                    "end": 1515.05,
                    "word": " I",
                    "probability": 0.951171875
                },
                {
                    "start": 1515.05,
                    "end": 1515.33,
                    "word": " really",
                    "probability": 0.76416015625
                },
                {
                    "start": 1515.33,
                    "end": 1515.73,
                    "word": " mean",
                    "probability": 0.921875
                },
                {
                    "start": 1515.73,
                    "end": 1516.05,
                    "word": ",",
                    "probability": 0.70654296875
                },
                {
                    "start": 1516.05,
                    "end": 1516.19,
                    "word": " you",
                    "probability": 0.59716796875
                },
                {
                    "start": 1516.19,
                    "end": 1516.33,
                    "word": " know",
                    "probability": 0.8125
                },
                {
                    "start": 1516.33,
                    "end": 1516.39,
                    "word": ",",
                    "probability": 0.96435546875
                },
                {
                    "start": 1516.39,
                    "end": 1516.53,
                    "word": " if",
                    "probability": 0.89599609375
                },
                {
                    "start": 1516.53,
                    "end": 1516.67,
                    "word": " you're",
                    "probability": 0.90869140625
                },
                {
                    "start": 1516.83,
                    "end": 1517.03,
                    "word": " a red",
                    "probability": 0.8349609375
                },
                {
                    "start": 1517.03,
                    "end": 1517.41,
                    "word": " teamer",
                    "probability": 0.863037109375
                },
                {
                    "start": 1517.41,
                    "end": 1517.45,
                    "word": ",",
                    "probability": 0.92724609375
                },
                {
                    "start": 1517.45,
                    "end": 1517.55,
                    "word": " if",
                    "probability": 0.86767578125
                },
                {
                    "start": 1517.55,
                    "end": 1517.65,
                    "word": " you're",
                    "probability": 0.92626953125
                },
                {
                    "start": 1517.65,
                    "end": 1518.09,
                    "word": " catching",
                    "probability": 0.70751953125
                },
                {
                    "start": 1518.09,
                    "end": 1518.45,
                    "word": " up",
                    "probability": 0.91162109375
                },
                {
                    "start": 1518.45,
                    "end": 1518.65,
                    "word": " to",
                    "probability": 0.9013671875
                },
                {
                    "start": 1518.65,
                    "end": 1518.89,
                    "word": " the",
                    "probability": 0.8173828125
                },
                {
                    "start": 1518.89,
                    "end": 1519.51,
                    "word": " elite",
                    "probability": 0.84033203125
                }
            ]
        },
        {
            "id": 214,
            "seek": 153144,
            "start": 1520.81,
            "end": 1527.31,
            "text": " APT black hat hackers, right? Take advantage of the lack of defenses on machine learning systems.",
            "tokens": [
                51036,
                5372,
                51,
                2211,
                2385,
                39766,
                11,
                558,
                30,
                3664,
                5002,
                295,
                264,
                5011,
                295,
                35989,
                322,
                3479,
                2539,
                3652,
                13,
                51364
            ],
            "temperature": 0.0,
            "avg_logprob": -0.20984543074843703,
            "compression_ratio": 1.5991189427312775,
            "no_speech_prob": 0.00243377685546875,
            "words": [
                {
                    "start": 1520.81,
                    "end": 1521.39,
                    "word": " APT",
                    "probability": 0.7139892578125
                },
                {
                    "start": 1521.39,
                    "end": 1521.79,
                    "word": " black",
                    "probability": 0.6708984375
                },
                {
                    "start": 1521.79,
                    "end": 1522.05,
                    "word": " hat",
                    "probability": 0.83935546875
                },
                {
                    "start": 1522.05,
                    "end": 1522.49,
                    "word": " hackers",
                    "probability": 0.72802734375
                },
                {
                    "start": 1522.49,
                    "end": 1522.75,
                    "word": ",",
                    "probability": 0.96337890625
                },
                {
                    "start": 1522.75,
                    "end": 1522.95,
                    "word": " right",
                    "probability": 0.82275390625
                },
                {
                    "start": 1522.95,
                    "end": 1523.41,
                    "word": "?",
                    "probability": 0.51806640625
                },
                {
                    "start": 1523.41,
                    "end": 1523.95,
                    "word": " Take",
                    "probability": 0.8251953125
                },
                {
                    "start": 1523.95,
                    "end": 1524.35,
                    "word": " advantage",
                    "probability": 0.7548828125
                },
                {
                    "start": 1524.35,
                    "end": 1524.53,
                    "word": " of",
                    "probability": 0.90283203125
                },
                {
                    "start": 1524.53,
                    "end": 1524.65,
                    "word": " the",
                    "probability": 0.81005859375
                },
                {
                    "start": 1524.65,
                    "end": 1524.85,
                    "word": " lack",
                    "probability": 0.81005859375
                },
                {
                    "start": 1524.85,
                    "end": 1524.97,
                    "word": " of",
                    "probability": 0.90380859375
                },
                {
                    "start": 1524.97,
                    "end": 1525.31,
                    "word": " defenses",
                    "probability": 0.794921875
                },
                {
                    "start": 1525.31,
                    "end": 1525.59,
                    "word": " on",
                    "probability": 0.90673828125
                },
                {
                    "start": 1525.59,
                    "end": 1525.87,
                    "word": " machine",
                    "probability": 0.7021484375
                },
                {
                    "start": 1525.87,
                    "end": 1526.17,
                    "word": " learning",
                    "probability": 0.81005859375
                },
                {
                    "start": 1526.17,
                    "end": 1526.73,
                    "word": " systems",
                    "probability": 0.8076171875
                },
                {
                    "start": 1526.73,
                    "end": 1527.31,
                    "word": ".",
                    "probability": 0.97998046875
                }
            ]
        },
        {
            "id": 215,
            "seek": 153144,
            "start": 1527.35,
            "end": 1535.17,
            "text": " Microsoft put out a report earlier this year saying that in their survey of large companies",
            "tokens": [
                51364,
                8116,
                829,
                484,
                257,
                2275,
                3071,
                341,
                1064,
                1566,
                300,
                294,
                641,
                8984,
                295,
                2416,
                3431,
                51764
            ],
            "temperature": 0.0,
            "avg_logprob": -0.20984543074843703,
            "compression_ratio": 1.5991189427312775,
            "no_speech_prob": 0.00243377685546875,
            "words": [
                {
                    "start": 1527.35,
                    "end": 1527.69,
                    "word": " Microsoft",
                    "probability": 0.798828125
                },
                {
                    "start": 1527.69,
                    "end": 1528.13,
                    "word": " put",
                    "probability": 0.8916015625
                },
                {
                    "start": 1528.13,
                    "end": 1528.27,
                    "word": " out",
                    "probability": 0.82861328125
                },
                {
                    "start": 1528.41,
                    "end": 1528.81,
                    "word": " a report",
                    "probability": 0.8466796875
                },
                {
                    "start": 1528.81,
                    "end": 1530.17,
                    "word": " earlier",
                    "probability": 0.8857421875
                },
                {
                    "start": 1530.17,
                    "end": 1530.55,
                    "word": " this",
                    "probability": 0.88427734375
                },
                {
                    "start": 1530.55,
                    "end": 1530.81,
                    "word": " year",
                    "probability": 0.87939453125
                },
                {
                    "start": 1530.81,
                    "end": 1531.73,
                    "word": " saying",
                    "probability": 0.61669921875
                },
                {
                    "start": 1531.73,
                    "end": 1532.17,
                    "word": " that",
                    "probability": 0.8642578125
                },
                {
                    "start": 1532.17,
                    "end": 1532.83,
                    "word": " in",
                    "probability": 0.8037109375
                },
                {
                    "start": 1532.83,
                    "end": 1533.05,
                    "word": " their",
                    "probability": 0.9052734375
                },
                {
                    "start": 1533.05,
                    "end": 1533.51,
                    "word": " survey",
                    "probability": 0.7900390625
                },
                {
                    "start": 1533.51,
                    "end": 1534.13,
                    "word": " of",
                    "probability": 0.9052734375
                },
                {
                    "start": 1534.13,
                    "end": 1534.69,
                    "word": " large",
                    "probability": 0.9287109375
                },
                {
                    "start": 1534.69,
                    "end": 1535.17,
                    "word": " companies",
                    "probability": 0.8388671875
                }
            ]
        },
        {
            "id": 216,
            "seek": 155926,
            "start": 1535.17,
            "end": 1544.41,
            "text": " and government organizations out of, I think it was 28, only three had any meaningful defenses",
            "tokens": [
                50364,
                293,
                2463,
                6150,
                484,
                295,
                11,
                286,
                519,
                309,
                390,
                7562,
                11,
                787,
                1045,
                632,
                604,
                10995,
                35989,
                50828
            ],
            "temperature": 0.0,
            "avg_logprob": -0.19374999801317852,
            "compression_ratio": 1.5584415584415585,
            "no_speech_prob": 0.0011692047119140625,
            "words": [
                {
                    "start": 1535.17,
                    "end": 1535.39,
                    "word": " and",
                    "probability": 0.67236328125
                },
                {
                    "start": 1535.39,
                    "end": 1535.75,
                    "word": " government",
                    "probability": 0.8134765625
                },
                {
                    "start": 1535.75,
                    "end": 1536.73,
                    "word": " organizations",
                    "probability": 0.74169921875
                },
                {
                    "start": 1536.73,
                    "end": 1537.23,
                    "word": " out",
                    "probability": 0.53515625
                },
                {
                    "start": 1537.23,
                    "end": 1537.45,
                    "word": " of",
                    "probability": 0.89990234375
                },
                {
                    "start": 1537.45,
                    "end": 1537.57,
                    "word": ",",
                    "probability": 0.358154296875
                },
                {
                    "start": 1537.57,
                    "end": 1537.61,
                    "word": " I",
                    "probability": 0.97021484375
                },
                {
                    "start": 1537.61,
                    "end": 1537.73,
                    "word": " think",
                    "probability": 0.8076171875
                },
                {
                    "start": 1537.73,
                    "end": 1537.81,
                    "word": " it",
                    "probability": 0.82373046875
                },
                {
                    "start": 1537.81,
                    "end": 1537.93,
                    "word": " was",
                    "probability": 0.89697265625
                },
                {
                    "start": 1537.93,
                    "end": 1538.69,
                    "word": " 28",
                    "probability": 0.8154296875
                },
                {
                    "start": 1538.69,
                    "end": 1539.63,
                    "word": ",",
                    "probability": 0.58642578125
                },
                {
                    "start": 1539.63,
                    "end": 1540.69,
                    "word": " only",
                    "probability": 0.8017578125
                },
                {
                    "start": 1540.69,
                    "end": 1541.55,
                    "word": " three",
                    "probability": 0.74169921875
                },
                {
                    "start": 1541.55,
                    "end": 1542.83,
                    "word": " had",
                    "probability": 0.7998046875
                },
                {
                    "start": 1542.83,
                    "end": 1543.33,
                    "word": " any",
                    "probability": 0.826171875
                },
                {
                    "start": 1543.33,
                    "end": 1543.91,
                    "word": " meaningful",
                    "probability": 0.8623046875
                },
                {
                    "start": 1543.91,
                    "end": 1544.41,
                    "word": " defenses",
                    "probability": 0.86279296875
                }
            ]
        },
        {
            "id": 217,
            "seek": 155926,
            "start": 1544.41,
            "end": 1551.23,
            "text": " on their machine learning systems. So, if you're a red teamer, like, check it out.",
            "tokens": [
                50828,
                322,
                641,
                3479,
                2539,
                3652,
                13,
                407,
                11,
                498,
                291,
                434,
                257,
                2182,
                1469,
                260,
                11,
                411,
                11,
                1520,
                309,
                484,
                13,
                51160
            ],
            "temperature": 0.0,
            "avg_logprob": -0.19374999801317852,
            "compression_ratio": 1.5584415584415585,
            "no_speech_prob": 0.0011692047119140625,
            "words": [
                {
                    "start": 1544.41,
                    "end": 1544.73,
                    "word": " on",
                    "probability": 0.908203125
                },
                {
                    "start": 1544.73,
                    "end": 1544.87,
                    "word": " their",
                    "probability": 0.9052734375
                },
                {
                    "start": 1544.87,
                    "end": 1545.13,
                    "word": " machine",
                    "probability": 0.71630859375
                },
                {
                    "start": 1545.13,
                    "end": 1545.43,
                    "word": " learning",
                    "probability": 0.8310546875
                },
                {
                    "start": 1545.43,
                    "end": 1545.97,
                    "word": " systems",
                    "probability": 0.7890625
                },
                {
                    "start": 1545.97,
                    "end": 1546.55,
                    "word": ".",
                    "probability": 0.9599609375
                },
                {
                    "start": 1547.79,
                    "end": 1548.37,
                    "word": " So",
                    "probability": 0.8359375
                },
                {
                    "start": 1548.37,
                    "end": 1548.43,
                    "word": ",",
                    "probability": 0.211181640625
                },
                {
                    "start": 1548.43,
                    "end": 1548.47,
                    "word": " if",
                    "probability": 0.9111328125
                },
                {
                    "start": 1548.47,
                    "end": 1548.57,
                    "word": " you're",
                    "probability": 0.91796875
                },
                {
                    "start": 1548.73,
                    "end": 1548.87,
                    "word": " a red",
                    "probability": 0.6884765625
                },
                {
                    "start": 1548.87,
                    "end": 1549.39,
                    "word": " teamer",
                    "probability": 0.83251953125
                },
                {
                    "start": 1549.39,
                    "end": 1549.75,
                    "word": ",",
                    "probability": 0.86767578125
                },
                {
                    "start": 1549.75,
                    "end": 1550.01,
                    "word": " like",
                    "probability": 0.64111328125
                },
                {
                    "start": 1550.01,
                    "end": 1550.25,
                    "word": ",",
                    "probability": 0.7265625
                },
                {
                    "start": 1550.25,
                    "end": 1550.53,
                    "word": " check",
                    "probability": 0.82568359375
                },
                {
                    "start": 1550.53,
                    "end": 1550.67,
                    "word": " it",
                    "probability": 0.9013671875
                },
                {
                    "start": 1550.67,
                    "end": 1550.91,
                    "word": " out",
                    "probability": 0.82470703125
                },
                {
                    "start": 1550.91,
                    "end": 1551.23,
                    "word": ".",
                    "probability": 0.873046875
                }
            ]
        },
        {
            "id": 218,
            "seek": 155926,
            "start": 1551.23,
            "end": 1557.65,
            "text": " If you're a vulnerability discovery person, like, check it out. You can definitely find stuff.",
            "tokens": [
                51160,
                759,
                291,
                434,
                257,
                24210,
                12114,
                954,
                11,
                411,
                11,
                1520,
                309,
                484,
                13,
                509,
                393,
                2138,
                915,
                1507,
                13,
                51456
            ],
            "temperature": 0.0,
            "avg_logprob": -0.19374999801317852,
            "compression_ratio": 1.5584415584415585,
            "no_speech_prob": 0.0011692047119140625,
            "words": [
                {
                    "start": 1551.23,
                    "end": 1551.37,
                    "word": " If",
                    "probability": 0.89501953125
                },
                {
                    "start": 1551.37,
                    "end": 1551.63,
                    "word": " you're",
                    "probability": 0.92236328125
                },
                {
                    "start": 1552.29,
                    "end": 1552.69,
                    "word": " a vulnerability",
                    "probability": 0.869140625
                },
                {
                    "start": 1552.69,
                    "end": 1553.41,
                    "word": " discovery",
                    "probability": 0.755859375
                },
                {
                    "start": 1553.41,
                    "end": 1553.89,
                    "word": " person",
                    "probability": 0.79736328125
                },
                {
                    "start": 1553.89,
                    "end": 1554.11,
                    "word": ",",
                    "probability": 0.90185546875
                },
                {
                    "start": 1554.11,
                    "end": 1554.25,
                    "word": " like",
                    "probability": 0.8203125
                },
                {
                    "start": 1554.25,
                    "end": 1554.85,
                    "word": ",",
                    "probability": 0.94775390625
                },
                {
                    "start": 1554.85,
                    "end": 1555.15,
                    "word": " check",
                    "probability": 0.82275390625
                },
                {
                    "start": 1555.15,
                    "end": 1555.29,
                    "word": " it",
                    "probability": 0.9013671875
                },
                {
                    "start": 1555.29,
                    "end": 1555.45,
                    "word": " out",
                    "probability": 0.82470703125
                },
                {
                    "start": 1555.45,
                    "end": 1555.57,
                    "word": ".",
                    "probability": 0.87255859375
                },
                {
                    "start": 1555.57,
                    "end": 1555.63,
                    "word": " You",
                    "probability": 0.8828125
                },
                {
                    "start": 1555.63,
                    "end": 1555.79,
                    "word": " can",
                    "probability": 0.48486328125
                },
                {
                    "start": 1555.79,
                    "end": 1556.35,
                    "word": " definitely",
                    "probability": 0.46435546875
                },
                {
                    "start": 1556.35,
                    "end": 1556.77,
                    "word": " find",
                    "probability": 0.82373046875
                },
                {
                    "start": 1556.77,
                    "end": 1557.07,
                    "word": " stuff",
                    "probability": 0.80029296875
                },
                {
                    "start": 1557.07,
                    "end": 1557.65,
                    "word": ".",
                    "probability": 0.822265625
                }
            ]
        },
        {
            "id": 219,
            "seek": 155926,
            "start": 1557.93,
            "end": 1562.99,
            "text": " And that's because nobody knows what you're doing, and nobody's looking for you, right?",
            "tokens": [
                51496,
                400,
                300,
                311,
                570,
                5079,
                3255,
                437,
                291,
                434,
                884,
                11,
                293,
                5079,
                311,
                1237,
                337,
                291,
                11,
                558,
                30,
                51724
            ],
            "temperature": 0.0,
            "avg_logprob": -0.19374999801317852,
            "compression_ratio": 1.5584415584415585,
            "no_speech_prob": 0.0011692047119140625,
            "words": [
                {
                    "start": 1557.93,
                    "end": 1558.11,
                    "word": " And",
                    "probability": 0.85986328125
                },
                {
                    "start": 1558.11,
                    "end": 1558.29,
                    "word": " that's",
                    "probability": 0.935791015625
                },
                {
                    "start": 1558.29,
                    "end": 1558.45,
                    "word": " because",
                    "probability": 0.6455078125
                },
                {
                    "start": 1558.45,
                    "end": 1558.95,
                    "word": " nobody",
                    "probability": 0.8134765625
                },
                {
                    "start": 1558.95,
                    "end": 1559.17,
                    "word": " knows",
                    "probability": 0.73583984375
                },
                {
                    "start": 1559.17,
                    "end": 1559.31,
                    "word": " what",
                    "probability": 0.8857421875
                },
                {
                    "start": 1559.31,
                    "end": 1559.41,
                    "word": " you're",
                    "probability": 0.92041015625
                },
                {
                    "start": 1559.41,
                    "end": 1559.77,
                    "word": " doing",
                    "probability": 0.91650390625
                },
                {
                    "start": 1559.77,
                    "end": 1560.07,
                    "word": ",",
                    "probability": 0.343017578125
                },
                {
                    "start": 1560.07,
                    "end": 1560.29,
                    "word": " and",
                    "probability": 0.88916015625
                },
                {
                    "start": 1560.29,
                    "end": 1560.65,
                    "word": " nobody's",
                    "probability": 0.847900390625
                },
                {
                    "start": 1560.65,
                    "end": 1560.93,
                    "word": " looking",
                    "probability": 0.79296875
                },
                {
                    "start": 1560.93,
                    "end": 1561.23,
                    "word": " for",
                    "probability": 0.89892578125
                },
                {
                    "start": 1561.23,
                    "end": 1561.49,
                    "word": " you",
                    "probability": 0.89306640625
                },
                {
                    "start": 1561.49,
                    "end": 1562.11,
                    "word": ",",
                    "probability": 0.56689453125
                },
                {
                    "start": 1562.11,
                    "end": 1562.41,
                    "word": " right",
                    "probability": 0.87255859375
                },
                {
                    "start": 1562.41,
                    "end": 1562.99,
                    "word": "?",
                    "probability": 0.93115234375
                }
            ]
        },
        {
            "id": 220,
            "seek": 158710,
            "start": 1562.99,
            "end": 1569.27,
            "text": " Nobody is looking for black hat hackers, right? APTs, whoever.",
            "tokens": [
                50380,
                9297,
                307,
                1237,
                337,
                2211,
                2385,
                39766,
                11,
                558,
                30,
                5372,
                33424,
                11,
                11387,
                13,
                50652
            ],
            "temperature": 0.0,
            "avg_logprob": -0.1879734800319479,
            "compression_ratio": 1.5358851674641147,
            "no_speech_prob": 0.00028228759765625,
            "words": [
                {
                    "start": 1562.99,
                    "end": 1563.67,
                    "word": " Nobody",
                    "probability": 0.50390625
                },
                {
                    "start": 1563.67,
                    "end": 1563.91,
                    "word": " is",
                    "probability": 0.8935546875
                },
                {
                    "start": 1563.91,
                    "end": 1564.17,
                    "word": " looking",
                    "probability": 0.81005859375
                },
                {
                    "start": 1564.17,
                    "end": 1564.75,
                    "word": " for",
                    "probability": 0.892578125
                },
                {
                    "start": 1564.75,
                    "end": 1566.05,
                    "word": " black",
                    "probability": 0.607421875
                },
                {
                    "start": 1566.05,
                    "end": 1566.35,
                    "word": " hat",
                    "probability": 0.73046875
                },
                {
                    "start": 1566.35,
                    "end": 1567.21,
                    "word": " hackers",
                    "probability": 0.669921875
                },
                {
                    "start": 1567.21,
                    "end": 1567.61,
                    "word": ",",
                    "probability": 0.662109375
                },
                {
                    "start": 1567.61,
                    "end": 1567.61,
                    "word": " right",
                    "probability": 0.51416015625
                },
                {
                    "start": 1567.61,
                    "end": 1567.73,
                    "word": "?",
                    "probability": 0.73779296875
                },
                {
                    "start": 1567.77,
                    "end": 1568.29,
                    "word": " APTs",
                    "probability": 0.75146484375
                },
                {
                    "start": 1568.29,
                    "end": 1568.45,
                    "word": ",",
                    "probability": 0.853515625
                },
                {
                    "start": 1568.45,
                    "end": 1568.75,
                    "word": " whoever",
                    "probability": 0.8681640625
                },
                {
                    "start": 1568.75,
                    "end": 1569.27,
                    "word": ".",
                    "probability": 0.8359375
                }
            ]
        },
        {
            "id": 221,
            "seek": 158710,
            "start": 1574.34,
            "end": 1579.94,
            "text": " And if you are, let me know. Yell at me in the discord, and I'll be like, oh damn, cool.",
            "tokens": [
                50736,
                400,
                498,
                291,
                366,
                11,
                718,
                385,
                458,
                13,
                835,
                285,
                412,
                385,
                294,
                264,
                32989,
                11,
                293,
                286,
                603,
                312,
                411,
                11,
                1954,
                8151,
                11,
                1627,
                13,
                50988
            ],
            "temperature": 0.0,
            "avg_logprob": -0.1879734800319479,
            "compression_ratio": 1.5358851674641147,
            "no_speech_prob": 0.00028228759765625,
            "words": [
                {
                    "start": 1574.34,
                    "end": 1574.56,
                    "word": " And",
                    "probability": 0.7880859375
                },
                {
                    "start": 1574.56,
                    "end": 1574.66,
                    "word": " if",
                    "probability": 0.90380859375
                },
                {
                    "start": 1574.66,
                    "end": 1574.78,
                    "word": " you",
                    "probability": 0.896484375
                },
                {
                    "start": 1574.78,
                    "end": 1574.96,
                    "word": " are",
                    "probability": 0.8818359375
                },
                {
                    "start": 1574.96,
                    "end": 1575.32,
                    "word": ",",
                    "probability": 0.8232421875
                },
                {
                    "start": 1575.32,
                    "end": 1575.56,
                    "word": " let",
                    "probability": 0.89599609375
                },
                {
                    "start": 1575.56,
                    "end": 1575.66,
                    "word": " me",
                    "probability": 0.904296875
                },
                {
                    "start": 1575.66,
                    "end": 1575.92,
                    "word": " know",
                    "probability": 0.81396484375
                },
                {
                    "start": 1575.92,
                    "end": 1576.14,
                    "word": ".",
                    "probability": 0.441162109375
                },
                {
                    "start": 1576.14,
                    "end": 1576.22,
                    "word": " Yell",
                    "probability": 0.834716796875
                },
                {
                    "start": 1576.22,
                    "end": 1576.32,
                    "word": " at",
                    "probability": 0.90185546875
                },
                {
                    "start": 1576.32,
                    "end": 1576.42,
                    "word": " me",
                    "probability": 0.9013671875
                },
                {
                    "start": 1576.42,
                    "end": 1576.54,
                    "word": " in",
                    "probability": 0.89404296875
                },
                {
                    "start": 1576.54,
                    "end": 1576.64,
                    "word": " the",
                    "probability": 0.80859375
                },
                {
                    "start": 1576.64,
                    "end": 1576.96,
                    "word": " discord",
                    "probability": 0.418701171875
                },
                {
                    "start": 1576.96,
                    "end": 1577.82,
                    "word": ",",
                    "probability": 0.237548828125
                },
                {
                    "start": 1577.82,
                    "end": 1578.1,
                    "word": " and",
                    "probability": 0.88916015625
                },
                {
                    "start": 1578.1,
                    "end": 1578.28,
                    "word": " I'll",
                    "probability": 0.96337890625
                },
                {
                    "start": 1578.28,
                    "end": 1578.42,
                    "word": " be",
                    "probability": 0.89794921875
                },
                {
                    "start": 1578.42,
                    "end": 1578.58,
                    "word": " like",
                    "probability": 0.86474609375
                },
                {
                    "start": 1578.58,
                    "end": 1578.64,
                    "word": ",",
                    "probability": 0.86279296875
                },
                {
                    "start": 1578.64,
                    "end": 1578.74,
                    "word": " oh",
                    "probability": 0.2978515625
                },
                {
                    "start": 1578.74,
                    "end": 1579.06,
                    "word": " damn",
                    "probability": 0.482666015625
                },
                {
                    "start": 1579.06,
                    "end": 1579.22,
                    "word": ",",
                    "probability": 0.73193359375
                },
                {
                    "start": 1579.22,
                    "end": 1579.42,
                    "word": " cool",
                    "probability": 0.89111328125
                },
                {
                    "start": 1579.42,
                    "end": 1579.94,
                    "word": ".",
                    "probability": 0.94482421875
                }
            ]
        },
        {
            "id": 222,
            "seek": 158710,
            "start": 1581.58,
            "end": 1587.72,
            "text": " If you're a defender, if you're a blue teamer, right? If you're a sock monkey, if you're",
            "tokens": [
                51112,
                759,
                291,
                434,
                257,
                26537,
                11,
                498,
                291,
                434,
                257,
                3344,
                1469,
                260,
                11,
                558,
                30,
                759,
                291,
                434,
                257,
                35302,
                17847,
                11,
                498,
                291,
                434,
                51392
            ],
            "temperature": 0.0,
            "avg_logprob": -0.1879734800319479,
            "compression_ratio": 1.5358851674641147,
            "no_speech_prob": 0.00028228759765625,
            "words": [
                {
                    "start": 1581.58,
                    "end": 1582.1,
                    "word": " If",
                    "probability": 0.880859375
                },
                {
                    "start": 1582.1,
                    "end": 1582.24,
                    "word": " you're",
                    "probability": 0.923095703125
                },
                {
                    "start": 1582.36,
                    "end": 1582.78,
                    "word": " a defender",
                    "probability": 0.755859375
                },
                {
                    "start": 1582.78,
                    "end": 1583.1,
                    "word": ",",
                    "probability": 0.94873046875
                },
                {
                    "start": 1583.1,
                    "end": 1583.24,
                    "word": " if",
                    "probability": 0.8896484375
                },
                {
                    "start": 1583.24,
                    "end": 1583.34,
                    "word": " you're",
                    "probability": 0.925048828125
                },
                {
                    "start": 1583.46,
                    "end": 1583.7,
                    "word": " a blue",
                    "probability": 0.85546875
                },
                {
                    "start": 1583.7,
                    "end": 1584.26,
                    "word": " teamer",
                    "probability": 0.8173828125
                },
                {
                    "start": 1584.26,
                    "end": 1584.44,
                    "word": ",",
                    "probability": 0.9609375
                },
                {
                    "start": 1584.44,
                    "end": 1584.64,
                    "word": " right",
                    "probability": 0.71630859375
                },
                {
                    "start": 1584.64,
                    "end": 1584.7,
                    "word": "?",
                    "probability": 0.75390625
                },
                {
                    "start": 1584.7,
                    "end": 1584.9,
                    "word": " If",
                    "probability": 0.82763671875
                },
                {
                    "start": 1584.9,
                    "end": 1585.76,
                    "word": " you're",
                    "probability": 0.926513671875
                },
                {
                    "start": 1586.18,
                    "end": 1586.46,
                    "word": " a sock",
                    "probability": 0.6484375
                },
                {
                    "start": 1586.46,
                    "end": 1586.78,
                    "word": " monkey",
                    "probability": 0.88623046875
                },
                {
                    "start": 1586.78,
                    "end": 1587.12,
                    "word": ",",
                    "probability": 0.95556640625
                },
                {
                    "start": 1587.12,
                    "end": 1587.22,
                    "word": " if",
                    "probability": 0.89501953125
                },
                {
                    "start": 1587.22,
                    "end": 1587.72,
                    "word": " you're",
                    "probability": 0.924560546875
                }
            ]
        },
        {
            "id": 223,
            "seek": 158710,
            "start": 1588.96,
            "end": 1594.7,
            "text": " a researcher, if you're at DEF CON, you're probably one of these two categories.",
            "tokens": [
                51456,
                257,
                21751,
                11,
                498,
                291,
                434,
                412,
                10113,
                37,
                16596,
                11,
                291,
                434,
                1391,
                472,
                295,
                613,
                732,
                10479,
                13,
                51732
            ],
            "temperature": 0.0,
            "avg_logprob": -0.1879734800319479,
            "compression_ratio": 1.5358851674641147,
            "no_speech_prob": 0.00028228759765625,
            "words": [
                {
                    "start": 1588.96,
                    "end": 1589.48,
                    "word": " a researcher",
                    "probability": 0.78515625
                },
                {
                    "start": 1589.48,
                    "end": 1589.74,
                    "word": ",",
                    "probability": 0.91259765625
                },
                {
                    "start": 1589.74,
                    "end": 1589.92,
                    "word": " if",
                    "probability": 0.892578125
                },
                {
                    "start": 1589.92,
                    "end": 1590.18,
                    "word": " you're",
                    "probability": 0.878173828125
                },
                {
                    "start": 1590.18,
                    "end": 1590.5,
                    "word": " at",
                    "probability": 0.89013671875
                },
                {
                    "start": 1590.5,
                    "end": 1591.02,
                    "word": " DEF",
                    "probability": 0.5833740234375
                },
                {
                    "start": 1591.02,
                    "end": 1591.36,
                    "word": " CON",
                    "probability": 0.84130859375
                },
                {
                    "start": 1591.36,
                    "end": 1591.52,
                    "word": ",",
                    "probability": 0.90087890625
                },
                {
                    "start": 1591.52,
                    "end": 1591.72,
                    "word": " you're",
                    "probability": 0.923583984375
                },
                {
                    "start": 1591.72,
                    "end": 1592.26,
                    "word": " probably",
                    "probability": 0.908203125
                },
                {
                    "start": 1592.26,
                    "end": 1592.8,
                    "word": " one",
                    "probability": 0.82080078125
                },
                {
                    "start": 1592.8,
                    "end": 1592.96,
                    "word": " of",
                    "probability": 0.90380859375
                },
                {
                    "start": 1592.96,
                    "end": 1593.24,
                    "word": " these",
                    "probability": 0.69970703125
                },
                {
                    "start": 1593.24,
                    "end": 1593.58,
                    "word": " two",
                    "probability": 0.8095703125
                },
                {
                    "start": 1593.58,
                    "end": 1594.18,
                    "word": " categories",
                    "probability": 0.7822265625
                },
                {
                    "start": 1594.18,
                    "end": 1594.7,
                    "word": ".",
                    "probability": 0.91162109375
                }
            ]
        },
        {
            "id": 224,
            "seek": 161174,
            "start": 1595.46,
            "end": 1599.74,
            "text": " Test your machine learning systems as if they're part of your infrastructure.",
            "tokens": [
                50412,
                9279,
                428,
                3479,
                2539,
                3652,
                382,
                498,
                436,
                434,
                644,
                295,
                428,
                6896,
                13,
                50592
            ],
            "temperature": 0.0,
            "avg_logprob": -0.20558286248967889,
            "compression_ratio": 1.6777251184834123,
            "no_speech_prob": 0.0011157989501953125,
            "words": [
                {
                    "start": 1595.46,
                    "end": 1595.98,
                    "word": " Test",
                    "probability": 0.5732421875
                },
                {
                    "start": 1595.98,
                    "end": 1596.18,
                    "word": " your",
                    "probability": 0.78466796875
                },
                {
                    "start": 1596.18,
                    "end": 1596.44,
                    "word": " machine",
                    "probability": 0.681640625
                },
                {
                    "start": 1596.44,
                    "end": 1596.92,
                    "word": " learning",
                    "probability": 0.80419921875
                },
                {
                    "start": 1596.92,
                    "end": 1597.5,
                    "word": " systems",
                    "probability": 0.7919921875
                },
                {
                    "start": 1597.5,
                    "end": 1597.88,
                    "word": " as",
                    "probability": 0.87060546875
                },
                {
                    "start": 1597.88,
                    "end": 1598.02,
                    "word": " if",
                    "probability": 0.89990234375
                },
                {
                    "start": 1598.02,
                    "end": 1598.14,
                    "word": " they're",
                    "probability": 0.797607421875
                },
                {
                    "start": 1598.14,
                    "end": 1598.4,
                    "word": " part",
                    "probability": 0.80517578125
                },
                {
                    "start": 1598.4,
                    "end": 1598.52,
                    "word": " of",
                    "probability": 0.90625
                },
                {
                    "start": 1598.52,
                    "end": 1598.64,
                    "word": " your",
                    "probability": 0.80810546875
                },
                {
                    "start": 1598.64,
                    "end": 1599.22,
                    "word": " infrastructure",
                    "probability": 0.86865234375
                },
                {
                    "start": 1599.22,
                    "end": 1599.74,
                    "word": ".",
                    "probability": 0.92626953125
                }
            ]
        },
        {
            "id": 225,
            "seek": 161174,
            "start": 1600.46,
            "end": 1607.62,
            "text": " So, what we see a lot is that machine learning engineers and data scientists develop these models,",
            "tokens": [
                50644,
                407,
                11,
                437,
                321,
                536,
                257,
                688,
                307,
                300,
                3479,
                2539,
                11955,
                293,
                1412,
                7708,
                1499,
                613,
                5245,
                11,
                51004
            ],
            "temperature": 0.0,
            "avg_logprob": -0.20558286248967889,
            "compression_ratio": 1.6777251184834123,
            "no_speech_prob": 0.0011157989501953125,
            "words": [
                {
                    "start": 1600.46,
                    "end": 1600.62,
                    "word": " So",
                    "probability": 0.66748046875
                },
                {
                    "start": 1600.62,
                    "end": 1600.8,
                    "word": ",",
                    "probability": 0.389892578125
                },
                {
                    "start": 1600.8,
                    "end": 1601.16,
                    "word": " what",
                    "probability": 0.865234375
                },
                {
                    "start": 1601.16,
                    "end": 1601.34,
                    "word": " we",
                    "probability": 0.90380859375
                },
                {
                    "start": 1601.34,
                    "end": 1601.52,
                    "word": " see",
                    "probability": 0.81591796875
                },
                {
                    "start": 1601.68,
                    "end": 1601.86,
                    "word": " a lot",
                    "probability": 0.8935546875
                },
                {
                    "start": 1601.86,
                    "end": 1602.16,
                    "word": " is",
                    "probability": 0.89697265625
                },
                {
                    "start": 1602.16,
                    "end": 1602.88,
                    "word": " that",
                    "probability": 0.85205078125
                },
                {
                    "start": 1602.88,
                    "end": 1604.1,
                    "word": " machine",
                    "probability": 0.6494140625
                },
                {
                    "start": 1604.1,
                    "end": 1604.42,
                    "word": " learning",
                    "probability": 0.8251953125
                },
                {
                    "start": 1604.42,
                    "end": 1604.98,
                    "word": " engineers",
                    "probability": 0.67041015625
                },
                {
                    "start": 1604.98,
                    "end": 1605.2,
                    "word": " and",
                    "probability": 0.89111328125
                },
                {
                    "start": 1605.2,
                    "end": 1605.42,
                    "word": " data",
                    "probability": 0.82373046875
                },
                {
                    "start": 1605.42,
                    "end": 1605.92,
                    "word": " scientists",
                    "probability": 0.72705078125
                },
                {
                    "start": 1605.92,
                    "end": 1606.34,
                    "word": " develop",
                    "probability": 0.6220703125
                },
                {
                    "start": 1606.34,
                    "end": 1606.64,
                    "word": " these",
                    "probability": 0.69921875
                },
                {
                    "start": 1606.64,
                    "end": 1607.02,
                    "word": " models",
                    "probability": 0.78662109375
                },
                {
                    "start": 1607.02,
                    "end": 1607.62,
                    "word": ",",
                    "probability": 0.345947265625
                }
            ]
        },
        {
            "id": 226,
            "seek": 161174,
            "start": 1607.62,
            "end": 1613.52,
            "text": " and they take these models in, like, a .py file, and they hand it off to engineering,",
            "tokens": [
                51004,
                293,
                436,
                747,
                613,
                5245,
                294,
                11,
                411,
                11,
                257,
                2411,
                8200,
                3991,
                11,
                293,
                436,
                1011,
                309,
                766,
                281,
                7043,
                11,
                51300
            ],
            "temperature": 0.0,
            "avg_logprob": -0.20558286248967889,
            "compression_ratio": 1.6777251184834123,
            "no_speech_prob": 0.0011157989501953125,
            "words": [
                {
                    "start": 1607.62,
                    "end": 1607.78,
                    "word": " and",
                    "probability": 0.88330078125
                },
                {
                    "start": 1607.78,
                    "end": 1607.94,
                    "word": " they",
                    "probability": 0.7333984375
                },
                {
                    "start": 1607.94,
                    "end": 1608.16,
                    "word": " take",
                    "probability": 0.81103515625
                },
                {
                    "start": 1608.16,
                    "end": 1608.38,
                    "word": " these",
                    "probability": 0.68798828125
                },
                {
                    "start": 1608.38,
                    "end": 1608.8,
                    "word": " models",
                    "probability": 0.8076171875
                },
                {
                    "start": 1608.8,
                    "end": 1609.26,
                    "word": " in",
                    "probability": 0.87060546875
                },
                {
                    "start": 1609.26,
                    "end": 1609.38,
                    "word": ",",
                    "probability": 0.1998291015625
                },
                {
                    "start": 1609.38,
                    "end": 1609.44,
                    "word": " like",
                    "probability": 0.876953125
                },
                {
                    "start": 1609.44,
                    "end": 1609.46,
                    "word": ",",
                    "probability": 0.9248046875
                },
                {
                    "start": 1609.56,
                    "end": 1610.04,
                    "word": " a .py",
                    "probability": 0.781494140625
                },
                {
                    "start": 1610.04,
                    "end": 1610.54,
                    "word": " file",
                    "probability": 0.82568359375
                },
                {
                    "start": 1610.54,
                    "end": 1610.98,
                    "word": ",",
                    "probability": 0.65380859375
                },
                {
                    "start": 1610.98,
                    "end": 1611.12,
                    "word": " and",
                    "probability": 0.8876953125
                },
                {
                    "start": 1611.12,
                    "end": 1611.26,
                    "word": " they",
                    "probability": 0.70263671875
                },
                {
                    "start": 1611.26,
                    "end": 1611.48,
                    "word": " hand",
                    "probability": 0.77099609375
                },
                {
                    "start": 1611.48,
                    "end": 1611.6,
                    "word": " it",
                    "probability": 0.8916015625
                },
                {
                    "start": 1611.6,
                    "end": 1611.84,
                    "word": " off",
                    "probability": 0.7998046875
                },
                {
                    "start": 1611.84,
                    "end": 1612.2,
                    "word": " to",
                    "probability": 0.900390625
                },
                {
                    "start": 1612.2,
                    "end": 1612.9,
                    "word": " engineering",
                    "probability": 0.8134765625
                },
                {
                    "start": 1612.9,
                    "end": 1613.52,
                    "word": ",",
                    "probability": 0.73388671875
                }
            ]
        },
        {
            "id": 227,
            "seek": 161174,
            "start": 1613.52,
            "end": 1619.34,
            "text": " and engineering goes, I don't know what this is, and they slap a Django API in front of it,",
            "tokens": [
                51300,
                293,
                7043,
                1709,
                11,
                286,
                500,
                380,
                458,
                437,
                341,
                307,
                11,
                293,
                436,
                21075,
                257,
                33464,
                17150,
                9362,
                294,
                1868,
                295,
                309,
                11,
                51588
            ],
            "temperature": 0.0,
            "avg_logprob": -0.20558286248967889,
            "compression_ratio": 1.6777251184834123,
            "no_speech_prob": 0.0011157989501953125,
            "words": [
                {
                    "start": 1613.52,
                    "end": 1613.84,
                    "word": " and",
                    "probability": 0.892578125
                },
                {
                    "start": 1613.84,
                    "end": 1614.36,
                    "word": " engineering",
                    "probability": 0.822265625
                },
                {
                    "start": 1614.36,
                    "end": 1614.72,
                    "word": " goes",
                    "probability": 0.8759765625
                },
                {
                    "start": 1614.72,
                    "end": 1614.9,
                    "word": ",",
                    "probability": 0.8681640625
                },
                {
                    "start": 1614.9,
                    "end": 1615.46,
                    "word": " I",
                    "probability": 0.196533203125
                },
                {
                    "start": 1615.46,
                    "end": 1615.62,
                    "word": " don't",
                    "probability": 0.94775390625
                },
                {
                    "start": 1615.62,
                    "end": 1615.76,
                    "word": " know",
                    "probability": 0.8154296875
                },
                {
                    "start": 1615.76,
                    "end": 1615.9,
                    "word": " what",
                    "probability": 0.8798828125
                },
                {
                    "start": 1615.9,
                    "end": 1616.16,
                    "word": " this",
                    "probability": 0.88134765625
                },
                {
                    "start": 1616.16,
                    "end": 1616.46,
                    "word": " is",
                    "probability": 0.90380859375
                },
                {
                    "start": 1616.46,
                    "end": 1616.74,
                    "word": ",",
                    "probability": 0.763671875
                },
                {
                    "start": 1616.74,
                    "end": 1616.94,
                    "word": " and",
                    "probability": 0.89453125
                },
                {
                    "start": 1616.94,
                    "end": 1617.14,
                    "word": " they",
                    "probability": 0.7412109375
                },
                {
                    "start": 1617.14,
                    "end": 1617.5,
                    "word": " slap",
                    "probability": 0.8759765625
                },
                {
                    "start": 1617.68,
                    "end": 1618.02,
                    "word": " a Django",
                    "probability": 0.88818359375
                },
                {
                    "start": 1618.02,
                    "end": 1618.5,
                    "word": " API",
                    "probability": 0.77880859375
                },
                {
                    "start": 1618.5,
                    "end": 1618.76,
                    "word": " in",
                    "probability": 0.87841796875
                },
                {
                    "start": 1618.76,
                    "end": 1618.98,
                    "word": " front",
                    "probability": 0.9306640625
                },
                {
                    "start": 1618.98,
                    "end": 1619.1,
                    "word": " of",
                    "probability": 0.9033203125
                },
                {
                    "start": 1619.1,
                    "end": 1619.24,
                    "word": " it",
                    "probability": 0.9013671875
                },
                {
                    "start": 1619.24,
                    "end": 1619.34,
                    "word": ",",
                    "probability": 0.1451416015625
                }
            ]
        },
        {
            "id": 228,
            "seek": 164056,
            "start": 1619.34,
            "end": 1627.5,
            "text": " put it in a Docker container, and deploy it in Kubernetes as an API, and then when InfoSec",
            "tokens": [
                50364,
                829,
                309,
                294,
                257,
                33772,
                10129,
                11,
                293,
                7274,
                309,
                294,
                23145,
                382,
                364,
                9362,
                11,
                293,
                550,
                562,
                11537,
                78,
                29511,
                50788
            ],
            "temperature": 0.0,
            "avg_logprob": -0.23437500113532655,
            "compression_ratio": 1.6531531531531531,
            "no_speech_prob": 0.0838623046875,
            "words": [
                {
                    "start": 1619.34,
                    "end": 1619.56,
                    "word": " put",
                    "probability": 0.12261962890625
                },
                {
                    "start": 1619.56,
                    "end": 1619.68,
                    "word": " it",
                    "probability": 0.83935546875
                },
                {
                    "start": 1619.68,
                    "end": 1619.78,
                    "word": " in",
                    "probability": 0.87109375
                },
                {
                    "start": 1619.92,
                    "end": 1620.18,
                    "word": " a Docker",
                    "probability": 0.3486328125
                },
                {
                    "start": 1620.18,
                    "end": 1620.74,
                    "word": " container",
                    "probability": 0.7861328125
                },
                {
                    "start": 1620.74,
                    "end": 1621.02,
                    "word": ",",
                    "probability": 0.06414794921875
                },
                {
                    "start": 1621.02,
                    "end": 1621.02,
                    "word": " and",
                    "probability": 0.8017578125
                },
                {
                    "start": 1621.02,
                    "end": 1621.8,
                    "word": " deploy",
                    "probability": 0.047698974609375
                },
                {
                    "start": 1621.8,
                    "end": 1622.3,
                    "word": " it",
                    "probability": 0.8837890625
                },
                {
                    "start": 1622.3,
                    "end": 1622.74,
                    "word": " in",
                    "probability": 0.78662109375
                },
                {
                    "start": 1622.74,
                    "end": 1623.2,
                    "word": " Kubernetes",
                    "probability": 0.57421875
                },
                {
                    "start": 1623.2,
                    "end": 1624.28,
                    "word": " as",
                    "probability": 0.77001953125
                },
                {
                    "start": 1624.28,
                    "end": 1624.4,
                    "word": " an",
                    "probability": 0.8916015625
                },
                {
                    "start": 1624.4,
                    "end": 1624.74,
                    "word": " API",
                    "probability": 0.8056640625
                },
                {
                    "start": 1624.74,
                    "end": 1625.62,
                    "word": ",",
                    "probability": 0.348388671875
                },
                {
                    "start": 1625.62,
                    "end": 1625.72,
                    "word": " and",
                    "probability": 0.8720703125
                },
                {
                    "start": 1625.72,
                    "end": 1625.94,
                    "word": " then",
                    "probability": 0.402099609375
                },
                {
                    "start": 1625.94,
                    "end": 1626.38,
                    "word": " when",
                    "probability": 0.78759765625
                },
                {
                    "start": 1626.38,
                    "end": 1627.5,
                    "word": " InfoSec",
                    "probability": 0.80908203125
                }
            ]
        },
        {
            "id": 229,
            "seek": 164056,
            "start": 1627.5,
            "end": 1634.8,
            "text": " looks at it, they go, I don't know what this is. It's magic. You put in JSON, and you get",
            "tokens": [
                50788,
                1542,
                412,
                309,
                11,
                436,
                352,
                11,
                286,
                500,
                380,
                458,
                437,
                341,
                307,
                13,
                467,
                311,
                5585,
                13,
                509,
                829,
                294,
                31828,
                11,
                293,
                291,
                483,
                51132
            ],
            "temperature": 0.0,
            "avg_logprob": -0.23437500113532655,
            "compression_ratio": 1.6531531531531531,
            "no_speech_prob": 0.0838623046875,
            "words": [
                {
                    "start": 1627.5,
                    "end": 1628.12,
                    "word": " looks",
                    "probability": 0.7646484375
                },
                {
                    "start": 1628.12,
                    "end": 1628.28,
                    "word": " at",
                    "probability": 0.8994140625
                },
                {
                    "start": 1628.28,
                    "end": 1628.42,
                    "word": " it",
                    "probability": 0.8935546875
                },
                {
                    "start": 1628.42,
                    "end": 1628.5,
                    "word": ",",
                    "probability": 0.85107421875
                },
                {
                    "start": 1628.5,
                    "end": 1628.6,
                    "word": " they",
                    "probability": 0.7216796875
                },
                {
                    "start": 1628.6,
                    "end": 1628.88,
                    "word": " go",
                    "probability": 0.880859375
                },
                {
                    "start": 1628.88,
                    "end": 1629.12,
                    "word": ",",
                    "probability": 0.66552734375
                },
                {
                    "start": 1629.12,
                    "end": 1630.06,
                    "word": " I",
                    "probability": 0.2314453125
                },
                {
                    "start": 1630.06,
                    "end": 1630.9,
                    "word": " don't",
                    "probability": 0.944091796875
                },
                {
                    "start": 1630.9,
                    "end": 1631.02,
                    "word": " know",
                    "probability": 0.81884765625
                },
                {
                    "start": 1631.02,
                    "end": 1631.12,
                    "word": " what",
                    "probability": 0.87109375
                },
                {
                    "start": 1631.12,
                    "end": 1631.32,
                    "word": " this",
                    "probability": 0.8759765625
                },
                {
                    "start": 1631.32,
                    "end": 1631.6,
                    "word": " is",
                    "probability": 0.90087890625
                },
                {
                    "start": 1631.6,
                    "end": 1631.72,
                    "word": ".",
                    "probability": 0.494140625
                },
                {
                    "start": 1631.72,
                    "end": 1631.86,
                    "word": " It's",
                    "probability": 0.908935546875
                },
                {
                    "start": 1631.86,
                    "end": 1632.2,
                    "word": " magic",
                    "probability": 0.65771484375
                },
                {
                    "start": 1632.2,
                    "end": 1632.66,
                    "word": ".",
                    "probability": 0.9609375
                },
                {
                    "start": 1632.66,
                    "end": 1632.76,
                    "word": " You",
                    "probability": 0.85595703125
                },
                {
                    "start": 1632.76,
                    "end": 1633.06,
                    "word": " put",
                    "probability": 0.86279296875
                },
                {
                    "start": 1633.06,
                    "end": 1633.36,
                    "word": " in",
                    "probability": 0.8896484375
                },
                {
                    "start": 1633.36,
                    "end": 1633.8,
                    "word": " JSON",
                    "probability": 0.779296875
                },
                {
                    "start": 1633.8,
                    "end": 1634.16,
                    "word": ",",
                    "probability": 0.435302734375
                },
                {
                    "start": 1634.16,
                    "end": 1634.28,
                    "word": " and",
                    "probability": 0.86767578125
                },
                {
                    "start": 1634.28,
                    "end": 1634.42,
                    "word": " you",
                    "probability": 0.8251953125
                },
                {
                    "start": 1634.42,
                    "end": 1634.8,
                    "word": " get",
                    "probability": 0.896484375
                }
            ]
        },
        {
            "id": 230,
            "seek": 164056,
            "start": 1635.34,
            "end": 1642.9,
            "text": " an output. It's magic, and we don't mess with it. And that sucks. Don't do that. Don't do that.",
            "tokens": [
                51164,
                364,
                5598,
                13,
                467,
                311,
                5585,
                11,
                293,
                321,
                500,
                380,
                2082,
                365,
                309,
                13,
                400,
                300,
                15846,
                13,
                1468,
                380,
                360,
                300,
                13,
                1468,
                380,
                360,
                300,
                13,
                51540
            ],
            "temperature": 0.0,
            "avg_logprob": -0.23437500113532655,
            "compression_ratio": 1.6531531531531531,
            "no_speech_prob": 0.0838623046875,
            "words": [
                {
                    "start": 1635.34,
                    "end": 1635.6,
                    "word": " an",
                    "probability": 0.740234375
                },
                {
                    "start": 1635.6,
                    "end": 1635.92,
                    "word": " output",
                    "probability": 0.81396484375
                },
                {
                    "start": 1635.92,
                    "end": 1636.4,
                    "word": ".",
                    "probability": 0.9677734375
                },
                {
                    "start": 1636.4,
                    "end": 1636.68,
                    "word": " It's",
                    "probability": 0.9375
                },
                {
                    "start": 1636.68,
                    "end": 1637.04,
                    "word": " magic",
                    "probability": 0.8115234375
                },
                {
                    "start": 1637.04,
                    "end": 1637.36,
                    "word": ",",
                    "probability": 0.736328125
                },
                {
                    "start": 1637.36,
                    "end": 1637.52,
                    "word": " and",
                    "probability": 0.89453125
                },
                {
                    "start": 1637.52,
                    "end": 1637.7,
                    "word": " we",
                    "probability": 0.89990234375
                },
                {
                    "start": 1637.7,
                    "end": 1637.86,
                    "word": " don't",
                    "probability": 0.944580078125
                },
                {
                    "start": 1637.86,
                    "end": 1638.1,
                    "word": " mess",
                    "probability": 0.802734375
                },
                {
                    "start": 1638.1,
                    "end": 1638.3,
                    "word": " with",
                    "probability": 0.81396484375
                },
                {
                    "start": 1638.3,
                    "end": 1638.52,
                    "word": " it",
                    "probability": 0.8984375
                },
                {
                    "start": 1638.52,
                    "end": 1639.04,
                    "word": ".",
                    "probability": 0.93017578125
                },
                {
                    "start": 1639.28,
                    "end": 1639.56,
                    "word": " And",
                    "probability": 0.7578125
                },
                {
                    "start": 1639.56,
                    "end": 1639.74,
                    "word": " that",
                    "probability": 0.8720703125
                },
                {
                    "start": 1639.74,
                    "end": 1640.34,
                    "word": " sucks",
                    "probability": 0.80859375
                },
                {
                    "start": 1640.34,
                    "end": 1640.66,
                    "word": ".",
                    "probability": 0.89599609375
                },
                {
                    "start": 1640.66,
                    "end": 1640.92,
                    "word": " Don't",
                    "probability": 0.946044921875
                },
                {
                    "start": 1640.92,
                    "end": 1641.1,
                    "word": " do",
                    "probability": 0.90478515625
                },
                {
                    "start": 1641.1,
                    "end": 1641.36,
                    "word": " that",
                    "probability": 0.880859375
                },
                {
                    "start": 1641.36,
                    "end": 1641.76,
                    "word": ".",
                    "probability": 0.98095703125
                },
                {
                    "start": 1641.76,
                    "end": 1642.08,
                    "word": " Don't",
                    "probability": 0.91162109375
                },
                {
                    "start": 1642.08,
                    "end": 1642.54,
                    "word": " do",
                    "probability": 0.74365234375
                },
                {
                    "start": 1642.54,
                    "end": 1642.78,
                    "word": " that",
                    "probability": 0.8779296875
                },
                {
                    "start": 1642.78,
                    "end": 1642.9,
                    "word": ".",
                    "probability": 0.99072265625
                }
            ]
        },
        {
            "id": 231,
            "seek": 164056,
            "start": 1642.9,
            "end": 1648.16,
            "text": " Test your systems. Test the machine learning systems. Work with your data scientists. Work",
            "tokens": [
                51540,
                9279,
                428,
                3652,
                13,
                9279,
                264,
                3479,
                2539,
                3652,
                13,
                6603,
                365,
                428,
                1412,
                7708,
                13,
                6603,
                51800
            ],
            "temperature": 0.0,
            "avg_logprob": -0.23437500113532655,
            "compression_ratio": 1.6531531531531531,
            "no_speech_prob": 0.0838623046875,
            "words": [
                {
                    "start": 1642.9,
                    "end": 1643.14,
                    "word": " Test",
                    "probability": 0.8876953125
                },
                {
                    "start": 1643.14,
                    "end": 1643.32,
                    "word": " your",
                    "probability": 0.73193359375
                },
                {
                    "start": 1643.32,
                    "end": 1643.72,
                    "word": " systems",
                    "probability": 0.77001953125
                },
                {
                    "start": 1643.72,
                    "end": 1644.04,
                    "word": ".",
                    "probability": 0.94873046875
                },
                {
                    "start": 1644.04,
                    "end": 1644.26,
                    "word": " Test",
                    "probability": 0.90234375
                },
                {
                    "start": 1644.26,
                    "end": 1644.46,
                    "word": " the",
                    "probability": 0.7802734375
                },
                {
                    "start": 1644.46,
                    "end": 1644.68,
                    "word": " machine",
                    "probability": 0.70703125
                },
                {
                    "start": 1644.68,
                    "end": 1644.94,
                    "word": " learning",
                    "probability": 0.7080078125
                },
                {
                    "start": 1644.94,
                    "end": 1645.34,
                    "word": " systems",
                    "probability": 0.80859375
                },
                {
                    "start": 1645.34,
                    "end": 1645.86,
                    "word": ".",
                    "probability": 0.98046875
                },
                {
                    "start": 1646.38,
                    "end": 1646.9,
                    "word": " Work",
                    "probability": 0.8642578125
                },
                {
                    "start": 1646.9,
                    "end": 1647.08,
                    "word": " with",
                    "probability": 0.81689453125
                },
                {
                    "start": 1647.08,
                    "end": 1647.2,
                    "word": " your",
                    "probability": 0.80419921875
                },
                {
                    "start": 1647.2,
                    "end": 1647.42,
                    "word": " data",
                    "probability": 0.83251953125
                },
                {
                    "start": 1647.42,
                    "end": 1647.86,
                    "word": " scientists",
                    "probability": 0.70556640625
                },
                {
                    "start": 1647.86,
                    "end": 1647.98,
                    "word": ".",
                    "probability": 0.66064453125
                },
                {
                    "start": 1647.98,
                    "end": 1648.16,
                    "word": " Work",
                    "probability": 0.86474609375
                }
            ]
        },
        {
            "id": 232,
            "seek": 166652,
            "start": 1648.16,
            "end": 1653.22,
            "text": " with your ops people, because this is part of your attack surface. This is part of your attack",
            "tokens": [
                50364,
                365,
                428,
                44663,
                561,
                11,
                570,
                341,
                307,
                644,
                295,
                428,
                2690,
                3753,
                13,
                639,
                307,
                644,
                295,
                428,
                2690,
                50616
            ],
            "temperature": 0.0,
            "avg_logprob": -0.1956992515242926,
            "compression_ratio": 1.6266666666666667,
            "no_speech_prob": 0.00757598876953125,
            "words": [
                {
                    "start": 1648.16,
                    "end": 1648.3,
                    "word": " with",
                    "probability": 0.489501953125
                },
                {
                    "start": 1648.3,
                    "end": 1648.42,
                    "word": " your",
                    "probability": 0.7763671875
                },
                {
                    "start": 1648.42,
                    "end": 1648.66,
                    "word": " ops",
                    "probability": 0.71533203125
                },
                {
                    "start": 1648.66,
                    "end": 1648.96,
                    "word": " people",
                    "probability": 0.89892578125
                },
                {
                    "start": 1648.96,
                    "end": 1649.18,
                    "word": ",",
                    "probability": 0.3017578125
                },
                {
                    "start": 1649.18,
                    "end": 1649.28,
                    "word": " because",
                    "probability": 0.78466796875
                },
                {
                    "start": 1649.28,
                    "end": 1649.78,
                    "word": " this",
                    "probability": 0.86376953125
                },
                {
                    "start": 1649.78,
                    "end": 1649.98,
                    "word": " is",
                    "probability": 0.8603515625
                },
                {
                    "start": 1649.98,
                    "end": 1650.22,
                    "word": " part",
                    "probability": 0.8076171875
                },
                {
                    "start": 1650.22,
                    "end": 1650.34,
                    "word": " of",
                    "probability": 0.900390625
                },
                {
                    "start": 1650.34,
                    "end": 1650.48,
                    "word": " your",
                    "probability": 0.80224609375
                },
                {
                    "start": 1650.48,
                    "end": 1650.92,
                    "word": " attack",
                    "probability": 0.8447265625
                },
                {
                    "start": 1650.92,
                    "end": 1652.3,
                    "word": " surface",
                    "probability": 0.7021484375
                },
                {
                    "start": 1652.3,
                    "end": 1652.54,
                    "word": ".",
                    "probability": 0.7314453125
                },
                {
                    "start": 1652.54,
                    "end": 1652.66,
                    "word": " This",
                    "probability": 0.59375
                },
                {
                    "start": 1652.66,
                    "end": 1652.74,
                    "word": " is",
                    "probability": 0.8310546875
                },
                {
                    "start": 1652.74,
                    "end": 1652.86,
                    "word": " part",
                    "probability": 0.82421875
                },
                {
                    "start": 1652.86,
                    "end": 1652.9,
                    "word": " of",
                    "probability": 0.8984375
                },
                {
                    "start": 1652.9,
                    "end": 1652.98,
                    "word": " your",
                    "probability": 0.80224609375
                },
                {
                    "start": 1652.98,
                    "end": 1653.22,
                    "word": " attack",
                    "probability": 0.8408203125
                }
            ]
        },
        {
            "id": 233,
            "seek": 166652,
            "start": 1653.22,
            "end": 1660.18,
            "text": " surface, right? And don't let people hype you up over AI-generated phishing emails.",
            "tokens": [
                50616,
                3753,
                11,
                558,
                30,
                400,
                500,
                380,
                718,
                561,
                24144,
                291,
                493,
                670,
                7318,
                12,
                21848,
                770,
                903,
                3807,
                12524,
                13,
                50944
            ],
            "temperature": 0.0,
            "avg_logprob": -0.1956992515242926,
            "compression_ratio": 1.6266666666666667,
            "no_speech_prob": 0.00757598876953125,
            "words": [
                {
                    "start": 1653.22,
                    "end": 1653.6,
                    "word": " surface",
                    "probability": 0.77880859375
                },
                {
                    "start": 1653.6,
                    "end": 1653.72,
                    "word": ",",
                    "probability": 0.66064453125
                },
                {
                    "start": 1653.72,
                    "end": 1653.94,
                    "word": " right",
                    "probability": 0.85302734375
                },
                {
                    "start": 1653.94,
                    "end": 1654.42,
                    "word": "?",
                    "probability": 0.85107421875
                },
                {
                    "start": 1654.52,
                    "end": 1654.76,
                    "word": " And",
                    "probability": 0.84375
                },
                {
                    "start": 1654.76,
                    "end": 1655.92,
                    "word": " don't",
                    "probability": 0.89892578125
                },
                {
                    "start": 1655.92,
                    "end": 1656.34,
                    "word": " let",
                    "probability": 0.88818359375
                },
                {
                    "start": 1656.34,
                    "end": 1657.02,
                    "word": " people",
                    "probability": 0.92138671875
                },
                {
                    "start": 1657.02,
                    "end": 1657.26,
                    "word": " hype",
                    "probability": 0.9306640625
                },
                {
                    "start": 1657.26,
                    "end": 1657.4,
                    "word": " you",
                    "probability": 0.89306640625
                },
                {
                    "start": 1657.4,
                    "end": 1657.6,
                    "word": " up",
                    "probability": 0.89990234375
                },
                {
                    "start": 1657.6,
                    "end": 1658.1,
                    "word": " over",
                    "probability": 0.89013671875
                },
                {
                    "start": 1658.1,
                    "end": 1658.4,
                    "word": " AI",
                    "probability": 0.9150390625
                },
                {
                    "start": 1658.4,
                    "end": 1658.96,
                    "word": "-generated",
                    "probability": 0.7586263020833334
                },
                {
                    "start": 1658.96,
                    "end": 1659.34,
                    "word": " phishing",
                    "probability": 0.898193359375
                },
                {
                    "start": 1659.34,
                    "end": 1659.7,
                    "word": " emails",
                    "probability": 0.75927734375
                },
                {
                    "start": 1659.7,
                    "end": 1660.18,
                    "word": ".",
                    "probability": 0.97705078125
                }
            ]
        },
        {
            "id": 234,
            "seek": 166652,
            "start": 1660.74,
            "end": 1669.36,
            "text": " When GPT-2 came out, people went nuts and were like, this is the end of detecting phishing.",
            "tokens": [
                51000,
                1133,
                26039,
                51,
                12,
                17,
                1361,
                484,
                11,
                561,
                1437,
                10483,
                293,
                645,
                411,
                11,
                341,
                307,
                264,
                917,
                295,
                40237,
                903,
                3807,
                13,
                51420
            ],
            "temperature": 0.0,
            "avg_logprob": -0.1956992515242926,
            "compression_ratio": 1.6266666666666667,
            "no_speech_prob": 0.00757598876953125,
            "words": [
                {
                    "start": 1660.74,
                    "end": 1661.22,
                    "word": " When",
                    "probability": 0.787109375
                },
                {
                    "start": 1661.22,
                    "end": 1661.7,
                    "word": " GPT",
                    "probability": 0.94091796875
                },
                {
                    "start": 1661.7,
                    "end": 1662.0,
                    "word": "-2",
                    "probability": 0.856689453125
                },
                {
                    "start": 1662.0,
                    "end": 1662.4,
                    "word": " came",
                    "probability": 0.82568359375
                },
                {
                    "start": 1662.4,
                    "end": 1662.72,
                    "word": " out",
                    "probability": 0.8203125
                },
                {
                    "start": 1662.72,
                    "end": 1663.34,
                    "word": ",",
                    "probability": 0.94580078125
                },
                {
                    "start": 1663.34,
                    "end": 1664.58,
                    "word": " people",
                    "probability": 0.9150390625
                },
                {
                    "start": 1664.58,
                    "end": 1664.78,
                    "word": " went",
                    "probability": 0.82421875
                },
                {
                    "start": 1664.78,
                    "end": 1665.38,
                    "word": " nuts",
                    "probability": 0.7802734375
                },
                {
                    "start": 1665.38,
                    "end": 1666.56,
                    "word": " and",
                    "probability": 0.6484375
                },
                {
                    "start": 1666.56,
                    "end": 1666.76,
                    "word": " were",
                    "probability": 0.87890625
                },
                {
                    "start": 1666.76,
                    "end": 1666.98,
                    "word": " like",
                    "probability": 0.841796875
                },
                {
                    "start": 1666.98,
                    "end": 1667.04,
                    "word": ",",
                    "probability": 0.89501953125
                },
                {
                    "start": 1667.04,
                    "end": 1667.24,
                    "word": " this",
                    "probability": 0.334228515625
                },
                {
                    "start": 1667.24,
                    "end": 1667.36,
                    "word": " is",
                    "probability": 0.90625
                },
                {
                    "start": 1667.36,
                    "end": 1667.54,
                    "word": " the",
                    "probability": 0.81591796875
                },
                {
                    "start": 1667.54,
                    "end": 1667.8,
                    "word": " end",
                    "probability": 0.80908203125
                },
                {
                    "start": 1667.8,
                    "end": 1668.04,
                    "word": " of",
                    "probability": 0.90087890625
                },
                {
                    "start": 1668.04,
                    "end": 1668.68,
                    "word": " detecting",
                    "probability": 0.79150390625
                },
                {
                    "start": 1668.68,
                    "end": 1669.2,
                    "word": " phishing",
                    "probability": 0.88720703125
                },
                {
                    "start": 1669.2,
                    "end": 1669.36,
                    "word": ".",
                    "probability": 0.6083984375
                }
            ]
        },
        {
            "id": 235,
            "seek": 166652,
            "start": 1669.36,
            "end": 1674.12,
            "text": " They're going to be too convincing. And, like, just don't. Okay? It's the same thing that we've",
            "tokens": [
                51420,
                814,
                434,
                516,
                281,
                312,
                886,
                24823,
                13,
                400,
                11,
                411,
                11,
                445,
                500,
                380,
                13,
                1033,
                30,
                467,
                311,
                264,
                912,
                551,
                300,
                321,
                600,
                51660
            ],
            "temperature": 0.0,
            "avg_logprob": -0.1956992515242926,
            "compression_ratio": 1.6266666666666667,
            "no_speech_prob": 0.00757598876953125,
            "words": [
                {
                    "start": 1669.36,
                    "end": 1669.54,
                    "word": " They're",
                    "probability": 0.828857421875
                },
                {
                    "start": 1669.54,
                    "end": 1669.64,
                    "word": " going",
                    "probability": 0.54736328125
                },
                {
                    "start": 1669.64,
                    "end": 1669.68,
                    "word": " to",
                    "probability": 0.90625
                },
                {
                    "start": 1669.68,
                    "end": 1669.82,
                    "word": " be",
                    "probability": 0.90673828125
                },
                {
                    "start": 1669.82,
                    "end": 1670.06,
                    "word": " too",
                    "probability": 0.8330078125
                },
                {
                    "start": 1670.06,
                    "end": 1670.52,
                    "word": " convincing",
                    "probability": 0.8115234375
                },
                {
                    "start": 1670.52,
                    "end": 1671.0,
                    "word": ".",
                    "probability": 0.8623046875
                },
                {
                    "start": 1671.3,
                    "end": 1671.4,
                    "word": " And",
                    "probability": 0.82080078125
                },
                {
                    "start": 1671.4,
                    "end": 1671.44,
                    "word": ",",
                    "probability": 0.1214599609375
                },
                {
                    "start": 1671.44,
                    "end": 1671.58,
                    "word": " like",
                    "probability": 0.822265625
                },
                {
                    "start": 1671.58,
                    "end": 1671.62,
                    "word": ",",
                    "probability": 0.9873046875
                },
                {
                    "start": 1671.62,
                    "end": 1671.82,
                    "word": " just",
                    "probability": 0.87548828125
                },
                {
                    "start": 1671.82,
                    "end": 1672.4,
                    "word": " don't",
                    "probability": 0.938232421875
                },
                {
                    "start": 1672.4,
                    "end": 1672.64,
                    "word": ".",
                    "probability": 0.487060546875
                },
                {
                    "start": 1672.64,
                    "end": 1672.88,
                    "word": " Okay",
                    "probability": 0.448486328125
                },
                {
                    "start": 1672.88,
                    "end": 1673.0,
                    "word": "?",
                    "probability": 0.6611328125
                },
                {
                    "start": 1673.0,
                    "end": 1673.24,
                    "word": " It's",
                    "probability": 0.951904296875
                },
                {
                    "start": 1673.24,
                    "end": 1673.4,
                    "word": " the",
                    "probability": 0.81787109375
                },
                {
                    "start": 1673.4,
                    "end": 1673.62,
                    "word": " same",
                    "probability": 0.81005859375
                },
                {
                    "start": 1673.62,
                    "end": 1673.82,
                    "word": " thing",
                    "probability": 0.81787109375
                },
                {
                    "start": 1673.82,
                    "end": 1673.96,
                    "word": " that",
                    "probability": 0.82958984375
                },
                {
                    "start": 1673.96,
                    "end": 1674.12,
                    "word": " we've",
                    "probability": 0.92529296875
                }
            ]
        },
        {
            "id": 236,
            "seek": 169390,
            "start": 1674.12,
            "end": 1681.44,
            "text": " been saying for 20 years. Patch your systems. Look for bad stuff, and patch your systems.",
            "tokens": [
                50368,
                668,
                1566,
                337,
                945,
                924,
                13,
                44359,
                428,
                3652,
                13,
                2053,
                337,
                1578,
                1507,
                11,
                293,
                9972,
                428,
                3652,
                13,
                50704
            ],
            "temperature": 0.0,
            "avg_logprob": -0.20328444075219484,
            "compression_ratio": 1.7912621359223302,
            "no_speech_prob": 0.10089111328125,
            "words": [
                {
                    "start": 1674.12,
                    "end": 1674.36,
                    "word": " been",
                    "probability": 0.70166015625
                },
                {
                    "start": 1674.36,
                    "end": 1674.9,
                    "word": " saying",
                    "probability": 0.8271484375
                },
                {
                    "start": 1674.9,
                    "end": 1675.14,
                    "word": " for",
                    "probability": 0.89599609375
                },
                {
                    "start": 1675.14,
                    "end": 1675.44,
                    "word": " 20",
                    "probability": 0.84814453125
                },
                {
                    "start": 1675.44,
                    "end": 1675.88,
                    "word": " years",
                    "probability": 0.86865234375
                },
                {
                    "start": 1675.88,
                    "end": 1676.4,
                    "word": ".",
                    "probability": 0.736328125
                },
                {
                    "start": 1677.12,
                    "end": 1677.32,
                    "word": " Patch",
                    "probability": 0.83203125
                },
                {
                    "start": 1677.32,
                    "end": 1677.46,
                    "word": " your",
                    "probability": 0.7890625
                },
                {
                    "start": 1677.46,
                    "end": 1678.24,
                    "word": " systems",
                    "probability": 0.63232421875
                },
                {
                    "start": 1678.24,
                    "end": 1678.76,
                    "word": ".",
                    "probability": 0.75
                },
                {
                    "start": 1678.78,
                    "end": 1678.94,
                    "word": " Look",
                    "probability": 0.83984375
                },
                {
                    "start": 1678.94,
                    "end": 1679.12,
                    "word": " for",
                    "probability": 0.89501953125
                },
                {
                    "start": 1679.12,
                    "end": 1679.4,
                    "word": " bad",
                    "probability": 0.888671875
                },
                {
                    "start": 1679.4,
                    "end": 1679.82,
                    "word": " stuff",
                    "probability": 0.7900390625
                },
                {
                    "start": 1679.82,
                    "end": 1680.0,
                    "word": ",",
                    "probability": 0.1895751953125
                },
                {
                    "start": 1680.0,
                    "end": 1680.22,
                    "word": " and",
                    "probability": 0.888671875
                },
                {
                    "start": 1680.22,
                    "end": 1680.42,
                    "word": " patch",
                    "probability": 0.79150390625
                },
                {
                    "start": 1680.42,
                    "end": 1680.56,
                    "word": " your",
                    "probability": 0.7998046875
                },
                {
                    "start": 1680.56,
                    "end": 1680.92,
                    "word": " systems",
                    "probability": 0.810546875
                },
                {
                    "start": 1680.92,
                    "end": 1681.44,
                    "word": ".",
                    "probability": 0.97998046875
                }
            ]
        },
        {
            "id": 237,
            "seek": 169390,
            "start": 1681.82,
            "end": 1687.24,
            "text": " That's it. There's nothing new under the sun. It's the exact same thing we've been saying for",
            "tokens": [
                50756,
                663,
                311,
                309,
                13,
                821,
                311,
                1825,
                777,
                833,
                264,
                3295,
                13,
                467,
                311,
                264,
                1900,
                912,
                551,
                321,
                600,
                668,
                1566,
                337,
                51020
            ],
            "temperature": 0.0,
            "avg_logprob": -0.20328444075219484,
            "compression_ratio": 1.7912621359223302,
            "no_speech_prob": 0.10089111328125,
            "words": [
                {
                    "start": 1681.82,
                    "end": 1682.34,
                    "word": " That's",
                    "probability": 0.8828125
                },
                {
                    "start": 1682.34,
                    "end": 1682.5,
                    "word": " it",
                    "probability": 0.90380859375
                },
                {
                    "start": 1682.5,
                    "end": 1683.02,
                    "word": ".",
                    "probability": 0.9287109375
                },
                {
                    "start": 1683.32,
                    "end": 1683.84,
                    "word": " There's",
                    "probability": 0.77880859375
                },
                {
                    "start": 1683.84,
                    "end": 1684.26,
                    "word": " nothing",
                    "probability": 0.85986328125
                },
                {
                    "start": 1684.26,
                    "end": 1684.52,
                    "word": " new",
                    "probability": 0.8193359375
                },
                {
                    "start": 1684.52,
                    "end": 1684.72,
                    "word": " under",
                    "probability": 0.82568359375
                },
                {
                    "start": 1684.72,
                    "end": 1684.92,
                    "word": " the",
                    "probability": 0.80859375
                },
                {
                    "start": 1684.92,
                    "end": 1685.1,
                    "word": " sun",
                    "probability": 0.880859375
                },
                {
                    "start": 1685.1,
                    "end": 1685.3,
                    "word": ".",
                    "probability": 0.9482421875
                },
                {
                    "start": 1685.3,
                    "end": 1685.52,
                    "word": " It's",
                    "probability": 0.951171875
                },
                {
                    "start": 1685.52,
                    "end": 1685.64,
                    "word": " the",
                    "probability": 0.8125
                },
                {
                    "start": 1685.64,
                    "end": 1685.98,
                    "word": " exact",
                    "probability": 0.84033203125
                },
                {
                    "start": 1685.98,
                    "end": 1686.26,
                    "word": " same",
                    "probability": 0.80517578125
                },
                {
                    "start": 1686.26,
                    "end": 1686.42,
                    "word": " thing",
                    "probability": 0.82666015625
                },
                {
                    "start": 1686.42,
                    "end": 1686.66,
                    "word": " we've",
                    "probability": 0.925048828125
                },
                {
                    "start": 1686.66,
                    "end": 1686.78,
                    "word": " been",
                    "probability": 0.8271484375
                },
                {
                    "start": 1686.78,
                    "end": 1687.04,
                    "word": " saying",
                    "probability": 0.8154296875
                },
                {
                    "start": 1687.04,
                    "end": 1687.24,
                    "word": " for",
                    "probability": 0.89990234375
                }
            ]
        },
        {
            "id": 238,
            "seek": 169390,
            "start": 1687.24,
            "end": 1695.46,
            "text": " 20 years. Just keep looking for bad stuff, and patch your systems. And if anybody gives you",
            "tokens": [
                51020,
                945,
                924,
                13,
                1449,
                1066,
                1237,
                337,
                1578,
                1507,
                11,
                293,
                9972,
                428,
                3652,
                13,
                400,
                498,
                4472,
                2709,
                291,
                51428
            ],
            "temperature": 0.0,
            "avg_logprob": -0.20328444075219484,
            "compression_ratio": 1.7912621359223302,
            "no_speech_prob": 0.10089111328125,
            "words": [
                {
                    "start": 1687.24,
                    "end": 1687.56,
                    "word": " 20",
                    "probability": 0.892578125
                },
                {
                    "start": 1687.56,
                    "end": 1687.94,
                    "word": " years",
                    "probability": 0.8740234375
                },
                {
                    "start": 1687.94,
                    "end": 1688.46,
                    "word": ".",
                    "probability": 0.94775390625
                },
                {
                    "start": 1688.78,
                    "end": 1689.3,
                    "word": " Just",
                    "probability": 0.822265625
                },
                {
                    "start": 1689.3,
                    "end": 1689.76,
                    "word": " keep",
                    "probability": 0.81494140625
                },
                {
                    "start": 1689.76,
                    "end": 1690.14,
                    "word": " looking",
                    "probability": 0.80419921875
                },
                {
                    "start": 1690.14,
                    "end": 1690.38,
                    "word": " for",
                    "probability": 0.89453125
                },
                {
                    "start": 1690.38,
                    "end": 1690.66,
                    "word": " bad",
                    "probability": 0.896484375
                },
                {
                    "start": 1690.66,
                    "end": 1691.06,
                    "word": " stuff",
                    "probability": 0.79345703125
                },
                {
                    "start": 1691.06,
                    "end": 1691.46,
                    "word": ",",
                    "probability": 0.5068359375
                },
                {
                    "start": 1691.46,
                    "end": 1691.54,
                    "word": " and",
                    "probability": 0.89453125
                },
                {
                    "start": 1691.54,
                    "end": 1692.12,
                    "word": " patch",
                    "probability": 0.8154296875
                },
                {
                    "start": 1692.12,
                    "end": 1692.28,
                    "word": " your",
                    "probability": 0.80908203125
                },
                {
                    "start": 1692.28,
                    "end": 1692.64,
                    "word": " systems",
                    "probability": 0.8056640625
                },
                {
                    "start": 1692.64,
                    "end": 1693.16,
                    "word": ".",
                    "probability": 0.9541015625
                },
                {
                    "start": 1693.34,
                    "end": 1693.72,
                    "word": " And",
                    "probability": 0.7021484375
                },
                {
                    "start": 1693.72,
                    "end": 1694.86,
                    "word": " if",
                    "probability": 0.87109375
                },
                {
                    "start": 1694.86,
                    "end": 1695.12,
                    "word": " anybody",
                    "probability": 0.76171875
                },
                {
                    "start": 1695.12,
                    "end": 1695.36,
                    "word": " gives",
                    "probability": 0.85986328125
                },
                {
                    "start": 1695.36,
                    "end": 1695.46,
                    "word": " you",
                    "probability": 0.8916015625
                }
            ]
        },
        {
            "id": 239,
            "seek": 169390,
            "start": 1695.46,
            "end": 1701.5,
            "text": " pushback on patching your systems, like, have them talk to me, and I'll yell at them for you.",
            "tokens": [
                51428,
                2944,
                3207,
                322,
                9972,
                278,
                428,
                3652,
                11,
                411,
                11,
                362,
                552,
                751,
                281,
                385,
                11,
                293,
                286,
                603,
                20525,
                412,
                552,
                337,
                291,
                13,
                51704
            ],
            "temperature": 0.0,
            "avg_logprob": -0.20328444075219484,
            "compression_ratio": 1.7912621359223302,
            "no_speech_prob": 0.10089111328125,
            "words": [
                {
                    "start": 1695.46,
                    "end": 1695.92,
                    "word": " pushback",
                    "probability": 0.844482421875
                },
                {
                    "start": 1695.92,
                    "end": 1696.16,
                    "word": " on",
                    "probability": 0.908203125
                },
                {
                    "start": 1696.16,
                    "end": 1696.48,
                    "word": " patching",
                    "probability": 0.88037109375
                },
                {
                    "start": 1696.48,
                    "end": 1696.64,
                    "word": " your",
                    "probability": 0.80908203125
                },
                {
                    "start": 1696.64,
                    "end": 1697.02,
                    "word": " systems",
                    "probability": 0.8046875
                },
                {
                    "start": 1697.02,
                    "end": 1697.56,
                    "word": ",",
                    "probability": 0.9755859375
                },
                {
                    "start": 1697.56,
                    "end": 1697.72,
                    "word": " like",
                    "probability": 0.18310546875
                },
                {
                    "start": 1697.72,
                    "end": 1698.12,
                    "word": ",",
                    "probability": 0.68212890625
                },
                {
                    "start": 1698.12,
                    "end": 1698.36,
                    "word": " have",
                    "probability": 0.88232421875
                },
                {
                    "start": 1698.36,
                    "end": 1698.54,
                    "word": " them",
                    "probability": 0.71875
                },
                {
                    "start": 1698.54,
                    "end": 1699.0,
                    "word": " talk",
                    "probability": 0.8056640625
                },
                {
                    "start": 1699.0,
                    "end": 1699.14,
                    "word": " to",
                    "probability": 0.9072265625
                },
                {
                    "start": 1699.14,
                    "end": 1699.34,
                    "word": " me",
                    "probability": 0.9013671875
                },
                {
                    "start": 1699.34,
                    "end": 1699.46,
                    "word": ",",
                    "probability": 0.65673828125
                },
                {
                    "start": 1699.46,
                    "end": 1699.68,
                    "word": " and",
                    "probability": 0.89404296875
                },
                {
                    "start": 1699.68,
                    "end": 1700.0,
                    "word": " I'll",
                    "probability": 0.97412109375
                },
                {
                    "start": 1700.0,
                    "end": 1700.32,
                    "word": " yell",
                    "probability": 0.8125
                },
                {
                    "start": 1700.32,
                    "end": 1700.46,
                    "word": " at",
                    "probability": 0.904296875
                },
                {
                    "start": 1700.46,
                    "end": 1700.6,
                    "word": " them",
                    "probability": 0.728515625
                },
                {
                    "start": 1700.6,
                    "end": 1700.82,
                    "word": " for",
                    "probability": 0.89794921875
                },
                {
                    "start": 1700.82,
                    "end": 1701.06,
                    "word": " you",
                    "probability": 0.88720703125
                },
                {
                    "start": 1701.06,
                    "end": 1701.5,
                    "word": ".",
                    "probability": 0.364013671875
                }
            ]
        },
        {
            "id": 240,
            "seek": 171922,
            "start": 1701.5,
            "end": 1707.3,
            "text": " Because you have to patch your systems. If you're watching this talk and you're a data",
            "tokens": [
                50388,
                1436,
                291,
                362,
                281,
                9972,
                428,
                3652,
                13,
                759,
                291,
                434,
                1976,
                341,
                751,
                293,
                291,
                434,
                257,
                1412,
                50652
            ],
            "temperature": 0.0,
            "avg_logprob": -0.20460525719743025,
            "compression_ratio": 1.5676855895196506,
            "no_speech_prob": 0.0035381317138671875,
            "words": [
                {
                    "start": 1701.5,
                    "end": 1702.24,
                    "word": " Because",
                    "probability": 0.1600341796875
                },
                {
                    "start": 1702.24,
                    "end": 1702.4,
                    "word": " you",
                    "probability": 0.8818359375
                },
                {
                    "start": 1702.4,
                    "end": 1702.56,
                    "word": " have",
                    "probability": 0.84521484375
                },
                {
                    "start": 1702.56,
                    "end": 1702.68,
                    "word": " to",
                    "probability": 0.880859375
                },
                {
                    "start": 1702.68,
                    "end": 1702.88,
                    "word": " patch",
                    "probability": 0.83740234375
                },
                {
                    "start": 1702.88,
                    "end": 1703.02,
                    "word": " your",
                    "probability": 0.77880859375
                },
                {
                    "start": 1703.02,
                    "end": 1703.38,
                    "word": " systems",
                    "probability": 0.7919921875
                },
                {
                    "start": 1703.38,
                    "end": 1703.86,
                    "word": ".",
                    "probability": 0.9794921875
                },
                {
                    "start": 1705.24,
                    "end": 1705.72,
                    "word": " If",
                    "probability": 0.875
                },
                {
                    "start": 1705.72,
                    "end": 1705.82,
                    "word": " you're",
                    "probability": 0.8818359375
                },
                {
                    "start": 1705.82,
                    "end": 1706.08,
                    "word": " watching",
                    "probability": 0.7177734375
                },
                {
                    "start": 1706.08,
                    "end": 1706.32,
                    "word": " this",
                    "probability": 0.88134765625
                },
                {
                    "start": 1706.32,
                    "end": 1706.62,
                    "word": " talk",
                    "probability": 0.814453125
                },
                {
                    "start": 1706.62,
                    "end": 1706.88,
                    "word": " and",
                    "probability": 0.7646484375
                },
                {
                    "start": 1706.88,
                    "end": 1706.98,
                    "word": " you're",
                    "probability": 0.893798828125
                },
                {
                    "start": 1707.12,
                    "end": 1707.3,
                    "word": " a data",
                    "probability": 0.7783203125
                }
            ]
        },
        {
            "id": 241,
            "seek": 171922,
            "start": 1707.3,
            "end": 1711.72,
            "text": " scientist or a machine learning engineer, first of all, thank you for coming to DEF CON.",
            "tokens": [
                50652,
                12662,
                420,
                257,
                3479,
                2539,
                11403,
                11,
                700,
                295,
                439,
                11,
                1309,
                291,
                337,
                1348,
                281,
                10113,
                37,
                16596,
                13,
                50852
            ],
            "temperature": 0.0,
            "avg_logprob": -0.20460525719743025,
            "compression_ratio": 1.5676855895196506,
            "no_speech_prob": 0.0035381317138671875,
            "words": [
                {
                    "start": 1707.3,
                    "end": 1707.82,
                    "word": " scientist",
                    "probability": 0.74560546875
                },
                {
                    "start": 1707.82,
                    "end": 1708.1,
                    "word": " or",
                    "probability": 0.81787109375
                },
                {
                    "start": 1708.22,
                    "end": 1708.44,
                    "word": " a machine",
                    "probability": 0.71044921875
                },
                {
                    "start": 1708.44,
                    "end": 1708.74,
                    "word": " learning",
                    "probability": 0.765625
                },
                {
                    "start": 1708.74,
                    "end": 1709.14,
                    "word": " engineer",
                    "probability": 0.71044921875
                },
                {
                    "start": 1709.14,
                    "end": 1709.48,
                    "word": ",",
                    "probability": 0.8876953125
                },
                {
                    "start": 1709.48,
                    "end": 1709.66,
                    "word": " first",
                    "probability": 0.7099609375
                },
                {
                    "start": 1709.66,
                    "end": 1709.78,
                    "word": " of",
                    "probability": 0.9052734375
                },
                {
                    "start": 1709.78,
                    "end": 1709.92,
                    "word": " all",
                    "probability": 0.9033203125
                },
                {
                    "start": 1709.92,
                    "end": 1710.12,
                    "word": ",",
                    "probability": 0.94140625
                },
                {
                    "start": 1710.12,
                    "end": 1710.28,
                    "word": " thank",
                    "probability": 0.78759765625
                },
                {
                    "start": 1710.28,
                    "end": 1710.36,
                    "word": " you",
                    "probability": 0.896484375
                },
                {
                    "start": 1710.36,
                    "end": 1710.46,
                    "word": " for",
                    "probability": 0.90185546875
                },
                {
                    "start": 1710.46,
                    "end": 1710.64,
                    "word": " coming",
                    "probability": 0.82080078125
                },
                {
                    "start": 1710.64,
                    "end": 1710.82,
                    "word": " to",
                    "probability": 0.90234375
                },
                {
                    "start": 1710.82,
                    "end": 1711.0,
                    "word": " DEF",
                    "probability": 0.69384765625
                },
                {
                    "start": 1711.0,
                    "end": 1711.24,
                    "word": " CON",
                    "probability": 0.89990234375
                },
                {
                    "start": 1711.24,
                    "end": 1711.72,
                    "word": ".",
                    "probability": 0.92138671875
                }
            ]
        },
        {
            "id": 242,
            "seek": 171922,
            "start": 1712.36,
            "end": 1718.44,
            "text": " DEF CON is dope. And, like, even though it's free and remote this year, come back next year,",
            "tokens": [
                50908,
                10113,
                37,
                16596,
                307,
                23383,
                13,
                400,
                11,
                411,
                11,
                754,
                1673,
                309,
                311,
                1737,
                293,
                8607,
                341,
                1064,
                11,
                808,
                646,
                958,
                1064,
                11,
                51208
            ],
            "temperature": 0.0,
            "avg_logprob": -0.20460525719743025,
            "compression_ratio": 1.5676855895196506,
            "no_speech_prob": 0.0035381317138671875,
            "words": [
                {
                    "start": 1712.36,
                    "end": 1712.7,
                    "word": " DEF",
                    "probability": 0.8720703125
                },
                {
                    "start": 1712.7,
                    "end": 1712.88,
                    "word": " CON",
                    "probability": 0.89794921875
                },
                {
                    "start": 1712.88,
                    "end": 1713.34,
                    "word": " is",
                    "probability": 0.900390625
                },
                {
                    "start": 1713.34,
                    "end": 1713.66,
                    "word": " dope",
                    "probability": 0.705078125
                },
                {
                    "start": 1713.66,
                    "end": 1714.14,
                    "word": ".",
                    "probability": 0.5966796875
                },
                {
                    "start": 1714.46,
                    "end": 1714.52,
                    "word": " And",
                    "probability": 0.693359375
                },
                {
                    "start": 1714.52,
                    "end": 1714.52,
                    "word": ",",
                    "probability": 0.172607421875
                },
                {
                    "start": 1714.52,
                    "end": 1714.92,
                    "word": " like",
                    "probability": 0.84619140625
                },
                {
                    "start": 1714.92,
                    "end": 1714.98,
                    "word": ",",
                    "probability": 0.97314453125
                },
                {
                    "start": 1714.98,
                    "end": 1715.14,
                    "word": " even",
                    "probability": 0.81005859375
                },
                {
                    "start": 1715.14,
                    "end": 1715.34,
                    "word": " though",
                    "probability": 0.77294921875
                },
                {
                    "start": 1715.34,
                    "end": 1715.6,
                    "word": " it's",
                    "probability": 0.94677734375
                },
                {
                    "start": 1715.6,
                    "end": 1716.22,
                    "word": " free",
                    "probability": 0.876953125
                },
                {
                    "start": 1716.22,
                    "end": 1716.56,
                    "word": " and",
                    "probability": 0.892578125
                },
                {
                    "start": 1716.56,
                    "end": 1716.9,
                    "word": " remote",
                    "probability": 0.86279296875
                },
                {
                    "start": 1716.9,
                    "end": 1717.14,
                    "word": " this",
                    "probability": 0.8759765625
                },
                {
                    "start": 1717.14,
                    "end": 1717.34,
                    "word": " year",
                    "probability": 0.88232421875
                },
                {
                    "start": 1717.34,
                    "end": 1717.6,
                    "word": ",",
                    "probability": 0.927734375
                },
                {
                    "start": 1717.6,
                    "end": 1717.76,
                    "word": " come",
                    "probability": 0.78515625
                },
                {
                    "start": 1717.76,
                    "end": 1717.94,
                    "word": " back",
                    "probability": 0.826171875
                },
                {
                    "start": 1717.94,
                    "end": 1718.14,
                    "word": " next",
                    "probability": 0.86962890625
                },
                {
                    "start": 1718.14,
                    "end": 1718.32,
                    "word": " year",
                    "probability": 0.880859375
                },
                {
                    "start": 1718.32,
                    "end": 1718.44,
                    "word": ",",
                    "probability": 0.256103515625
                }
            ]
        },
        {
            "id": 243,
            "seek": 171922,
            "start": 1718.44,
            "end": 1726.82,
            "text": " because it's cool. Conduct threat modeling on your models. So, before you put a model into",
            "tokens": [
                51208,
                570,
                309,
                311,
                1627,
                13,
                21793,
                11130,
                4734,
                15983,
                322,
                428,
                5245,
                13,
                407,
                11,
                949,
                291,
                829,
                257,
                2316,
                666,
                51628
            ],
            "temperature": 0.0,
            "avg_logprob": -0.20460525719743025,
            "compression_ratio": 1.5676855895196506,
            "no_speech_prob": 0.0035381317138671875,
            "words": [
                {
                    "start": 1718.44,
                    "end": 1718.66,
                    "word": " because",
                    "probability": 0.77734375
                },
                {
                    "start": 1718.66,
                    "end": 1718.94,
                    "word": " it's",
                    "probability": 0.9501953125
                },
                {
                    "start": 1718.94,
                    "end": 1719.36,
                    "word": " cool",
                    "probability": 0.728515625
                },
                {
                    "start": 1719.36,
                    "end": 1719.84,
                    "word": ".",
                    "probability": 0.98828125
                },
                {
                    "start": 1722.28,
                    "end": 1722.7,
                    "word": " Conduct",
                    "probability": 0.87744140625
                },
                {
                    "start": 1722.7,
                    "end": 1722.98,
                    "word": " threat",
                    "probability": 0.7763671875
                },
                {
                    "start": 1722.98,
                    "end": 1723.4,
                    "word": " modeling",
                    "probability": 0.69140625
                },
                {
                    "start": 1723.4,
                    "end": 1723.88,
                    "word": " on",
                    "probability": 0.89501953125
                },
                {
                    "start": 1723.88,
                    "end": 1724.02,
                    "word": " your",
                    "probability": 0.8056640625
                },
                {
                    "start": 1724.02,
                    "end": 1724.38,
                    "word": " models",
                    "probability": 0.79443359375
                },
                {
                    "start": 1724.38,
                    "end": 1724.86,
                    "word": ".",
                    "probability": 0.96826171875
                },
                {
                    "start": 1725.04,
                    "end": 1725.14,
                    "word": " So",
                    "probability": 0.7724609375
                },
                {
                    "start": 1725.14,
                    "end": 1725.22,
                    "word": ",",
                    "probability": 0.61181640625
                },
                {
                    "start": 1725.22,
                    "end": 1725.42,
                    "word": " before",
                    "probability": 0.80322265625
                },
                {
                    "start": 1725.42,
                    "end": 1725.66,
                    "word": " you",
                    "probability": 0.8935546875
                },
                {
                    "start": 1725.66,
                    "end": 1726.0,
                    "word": " put",
                    "probability": 0.91064453125
                },
                {
                    "start": 1726.38,
                    "end": 1726.58,
                    "word": " a model",
                    "probability": 0.79833984375
                },
                {
                    "start": 1726.58,
                    "end": 1726.82,
                    "word": " into",
                    "probability": 0.708984375
                }
            ]
        },
        {
            "id": 244,
            "seek": 174658,
            "start": 1726.82,
            "end": 1732.42,
            "text": " deployment, work with your infosec team, work with ops, and sort of ask, like, what could go",
            "tokens": [
                50364,
                19317,
                11,
                589,
                365,
                428,
                1536,
                541,
                66,
                1469,
                11,
                589,
                365,
                44663,
                11,
                293,
                1333,
                295,
                1029,
                11,
                411,
                11,
                437,
                727,
                352,
                50652
            ],
            "temperature": 0.0,
            "avg_logprob": -0.21153845629849277,
            "compression_ratio": 1.6985645933014355,
            "no_speech_prob": 0.007232666015625,
            "words": [
                {
                    "start": 1726.82,
                    "end": 1727.28,
                    "word": " deployment",
                    "probability": 0.6181640625
                },
                {
                    "start": 1727.28,
                    "end": 1727.66,
                    "word": ",",
                    "probability": 0.59716796875
                },
                {
                    "start": 1727.66,
                    "end": 1728.18,
                    "word": " work",
                    "probability": 0.79541015625
                },
                {
                    "start": 1728.18,
                    "end": 1728.34,
                    "word": " with",
                    "probability": 0.8095703125
                },
                {
                    "start": 1728.34,
                    "end": 1728.48,
                    "word": " your",
                    "probability": 0.81298828125
                },
                {
                    "start": 1728.48,
                    "end": 1729.04,
                    "word": " infosec",
                    "probability": 0.7139485677083334
                },
                {
                    "start": 1729.04,
                    "end": 1729.52,
                    "word": " team",
                    "probability": 0.90478515625
                },
                {
                    "start": 1729.52,
                    "end": 1729.76,
                    "word": ",",
                    "probability": 0.85009765625
                },
                {
                    "start": 1729.76,
                    "end": 1729.96,
                    "word": " work",
                    "probability": 0.75
                },
                {
                    "start": 1729.96,
                    "end": 1730.2,
                    "word": " with",
                    "probability": 0.81396484375
                },
                {
                    "start": 1730.2,
                    "end": 1730.62,
                    "word": " ops",
                    "probability": 0.77392578125
                },
                {
                    "start": 1730.62,
                    "end": 1730.84,
                    "word": ",",
                    "probability": 0.325927734375
                },
                {
                    "start": 1730.84,
                    "end": 1730.9,
                    "word": " and",
                    "probability": 0.87255859375
                },
                {
                    "start": 1730.9,
                    "end": 1731.14,
                    "word": " sort",
                    "probability": 0.58740234375
                },
                {
                    "start": 1731.14,
                    "end": 1731.24,
                    "word": " of",
                    "probability": 0.908203125
                },
                {
                    "start": 1731.24,
                    "end": 1731.46,
                    "word": " ask",
                    "probability": 0.7939453125
                },
                {
                    "start": 1731.46,
                    "end": 1731.66,
                    "word": ",",
                    "probability": 0.431396484375
                },
                {
                    "start": 1731.66,
                    "end": 1731.8,
                    "word": " like",
                    "probability": 0.58251953125
                },
                {
                    "start": 1731.8,
                    "end": 1731.86,
                    "word": ",",
                    "probability": 0.8818359375
                },
                {
                    "start": 1731.86,
                    "end": 1732.06,
                    "word": " what",
                    "probability": 0.810546875
                },
                {
                    "start": 1732.06,
                    "end": 1732.24,
                    "word": " could",
                    "probability": 0.81591796875
                },
                {
                    "start": 1732.24,
                    "end": 1732.42,
                    "word": " go",
                    "probability": 0.90380859375
                }
            ]
        },
        {
            "id": 245,
            "seek": 174658,
            "start": 1732.42,
            "end": 1741.18,
            "text": " wrong? Right? What are the risks of an adversarial attack on this? What are the risks associated with",
            "tokens": [
                50652,
                2085,
                30,
                1779,
                30,
                708,
                366,
                264,
                10888,
                295,
                364,
                17641,
                44745,
                2690,
                322,
                341,
                30,
                708,
                366,
                264,
                10888,
                6615,
                365,
                51076
            ],
            "temperature": 0.0,
            "avg_logprob": -0.21153845629849277,
            "compression_ratio": 1.6985645933014355,
            "no_speech_prob": 0.007232666015625,
            "words": [
                {
                    "start": 1732.42,
                    "end": 1733.06,
                    "word": " wrong",
                    "probability": 0.8330078125
                },
                {
                    "start": 1733.06,
                    "end": 1733.66,
                    "word": "?",
                    "probability": 0.49072265625
                },
                {
                    "start": 1733.84,
                    "end": 1734.02,
                    "word": " Right",
                    "probability": 0.71337890625
                },
                {
                    "start": 1734.02,
                    "end": 1734.34,
                    "word": "?",
                    "probability": 0.77685546875
                },
                {
                    "start": 1734.34,
                    "end": 1734.72,
                    "word": " What",
                    "probability": 0.8056640625
                },
                {
                    "start": 1734.72,
                    "end": 1735.44,
                    "word": " are",
                    "probability": 0.52392578125
                },
                {
                    "start": 1735.44,
                    "end": 1735.64,
                    "word": " the",
                    "probability": 0.81298828125
                },
                {
                    "start": 1735.64,
                    "end": 1736.16,
                    "word": " risks",
                    "probability": 0.916015625
                },
                {
                    "start": 1736.16,
                    "end": 1736.44,
                    "word": " of",
                    "probability": 0.90185546875
                },
                {
                    "start": 1736.44,
                    "end": 1736.56,
                    "word": " an",
                    "probability": 0.9013671875
                },
                {
                    "start": 1736.56,
                    "end": 1737.18,
                    "word": " adversarial",
                    "probability": 0.91748046875
                },
                {
                    "start": 1737.18,
                    "end": 1737.58,
                    "word": " attack",
                    "probability": 0.8603515625
                },
                {
                    "start": 1737.58,
                    "end": 1737.82,
                    "word": " on",
                    "probability": 0.90869140625
                },
                {
                    "start": 1737.82,
                    "end": 1738.18,
                    "word": " this",
                    "probability": 0.88427734375
                },
                {
                    "start": 1738.18,
                    "end": 1738.66,
                    "word": "?",
                    "probability": 0.97607421875
                },
                {
                    "start": 1738.66,
                    "end": 1739.1,
                    "word": " What",
                    "probability": 0.775390625
                },
                {
                    "start": 1739.1,
                    "end": 1739.22,
                    "word": " are",
                    "probability": 0.89599609375
                },
                {
                    "start": 1739.22,
                    "end": 1739.44,
                    "word": " the",
                    "probability": 0.8134765625
                },
                {
                    "start": 1739.44,
                    "end": 1739.94,
                    "word": " risks",
                    "probability": 0.91943359375
                },
                {
                    "start": 1739.94,
                    "end": 1740.76,
                    "word": " associated",
                    "probability": 0.853515625
                },
                {
                    "start": 1740.76,
                    "end": 1741.18,
                    "word": " with",
                    "probability": 0.8173828125
                }
            ]
        },
        {
            "id": 246,
            "seek": 174658,
            "start": 1741.7,
            "end": 1746.56,
            "text": " this particular model? And what is the attack surface of this model?",
            "tokens": [
                51108,
                341,
                1729,
                2316,
                30,
                400,
                437,
                307,
                264,
                2690,
                3753,
                295,
                341,
                2316,
                30,
                51320
            ],
            "temperature": 0.0,
            "avg_logprob": -0.21153845629849277,
            "compression_ratio": 1.6985645933014355,
            "no_speech_prob": 0.007232666015625,
            "words": [
                {
                    "start": 1741.7,
                    "end": 1742.0,
                    "word": " this",
                    "probability": 0.88720703125
                },
                {
                    "start": 1742.0,
                    "end": 1742.64,
                    "word": " particular",
                    "probability": 0.8076171875
                },
                {
                    "start": 1742.64,
                    "end": 1743.22,
                    "word": " model",
                    "probability": 0.8134765625
                },
                {
                    "start": 1743.22,
                    "end": 1743.44,
                    "word": "?",
                    "probability": 0.6533203125
                },
                {
                    "start": 1743.44,
                    "end": 1743.6,
                    "word": " And",
                    "probability": 0.88720703125
                },
                {
                    "start": 1743.6,
                    "end": 1743.78,
                    "word": " what",
                    "probability": 0.87255859375
                },
                {
                    "start": 1743.78,
                    "end": 1743.92,
                    "word": " is",
                    "probability": 0.908203125
                },
                {
                    "start": 1743.92,
                    "end": 1744.04,
                    "word": " the",
                    "probability": 0.81787109375
                },
                {
                    "start": 1744.04,
                    "end": 1744.46,
                    "word": " attack",
                    "probability": 0.8359375
                },
                {
                    "start": 1744.46,
                    "end": 1744.88,
                    "word": " surface",
                    "probability": 0.79736328125
                },
                {
                    "start": 1744.88,
                    "end": 1745.3,
                    "word": " of",
                    "probability": 0.90234375
                },
                {
                    "start": 1745.3,
                    "end": 1745.58,
                    "word": " this",
                    "probability": 0.8828125
                },
                {
                    "start": 1745.58,
                    "end": 1745.96,
                    "word": " model",
                    "probability": 0.8232421875
                },
                {
                    "start": 1745.96,
                    "end": 1746.56,
                    "word": "?",
                    "probability": 0.97216796875
                }
            ]
        },
        {
            "id": 247,
            "seek": 174658,
            "start": 1748.08,
            "end": 1754.18,
            "text": " You know, and also when you're deploying those models, work with infosec and ops to balance",
            "tokens": [
                51424,
                509,
                458,
                11,
                293,
                611,
                562,
                291,
                434,
                34198,
                729,
                5245,
                11,
                589,
                365,
                1536,
                541,
                66,
                293,
                44663,
                281,
                4772,
                51740
            ],
            "temperature": 0.0,
            "avg_logprob": -0.21153845629849277,
            "compression_ratio": 1.6985645933014355,
            "no_speech_prob": 0.007232666015625,
            "words": [
                {
                    "start": 1748.08,
                    "end": 1748.24,
                    "word": " You",
                    "probability": 0.623046875
                },
                {
                    "start": 1748.24,
                    "end": 1748.4,
                    "word": " know",
                    "probability": 0.8173828125
                },
                {
                    "start": 1748.4,
                    "end": 1748.76,
                    "word": ",",
                    "probability": 0.7041015625
                },
                {
                    "start": 1748.76,
                    "end": 1749.06,
                    "word": " and",
                    "probability": 0.86376953125
                },
                {
                    "start": 1749.06,
                    "end": 1750.0,
                    "word": " also",
                    "probability": 0.78271484375
                },
                {
                    "start": 1750.0,
                    "end": 1750.18,
                    "word": " when",
                    "probability": 0.62939453125
                },
                {
                    "start": 1750.18,
                    "end": 1750.32,
                    "word": " you're",
                    "probability": 0.91455078125
                },
                {
                    "start": 1750.32,
                    "end": 1750.68,
                    "word": " deploying",
                    "probability": 0.7001953125
                },
                {
                    "start": 1750.68,
                    "end": 1750.98,
                    "word": " those",
                    "probability": 0.86279296875
                },
                {
                    "start": 1750.98,
                    "end": 1751.34,
                    "word": " models",
                    "probability": 0.810546875
                },
                {
                    "start": 1751.34,
                    "end": 1752.12,
                    "word": ",",
                    "probability": 0.9072265625
                },
                {
                    "start": 1752.12,
                    "end": 1752.42,
                    "word": " work",
                    "probability": 0.80712890625
                },
                {
                    "start": 1752.42,
                    "end": 1752.66,
                    "word": " with",
                    "probability": 0.81201171875
                },
                {
                    "start": 1752.66,
                    "end": 1753.18,
                    "word": " infosec",
                    "probability": 0.9347330729166666
                },
                {
                    "start": 1753.18,
                    "end": 1753.34,
                    "word": " and",
                    "probability": 0.87890625
                },
                {
                    "start": 1753.34,
                    "end": 1753.64,
                    "word": " ops",
                    "probability": 0.775390625
                },
                {
                    "start": 1753.64,
                    "end": 1753.84,
                    "word": " to",
                    "probability": 0.88720703125
                },
                {
                    "start": 1753.84,
                    "end": 1754.18,
                    "word": " balance",
                    "probability": 0.9111328125
                }
            ]
        },
        {
            "id": 248,
            "seek": 177432,
            "start": 1754.18,
            "end": 1761.3,
            "text": " uptime with the risk of model theft. These two kind of go hand in hand, is, like, you're going",
            "tokens": [
                50368,
                493,
                3766,
                365,
                264,
                3148,
                295,
                2316,
                28508,
                13,
                1981,
                732,
                733,
                295,
                352,
                1011,
                294,
                1011,
                11,
                307,
                11,
                411,
                11,
                291,
                434,
                516,
                50712
            ],
            "temperature": 0.0,
            "avg_logprob": -0.2117883598450387,
            "compression_ratio": 1.668122270742358,
            "no_speech_prob": 0.005062103271484375,
            "words": [
                {
                    "start": 1754.18,
                    "end": 1754.98,
                    "word": " uptime",
                    "probability": 0.406005859375
                },
                {
                    "start": 1754.98,
                    "end": 1755.36,
                    "word": " with",
                    "probability": 0.7822265625
                },
                {
                    "start": 1755.36,
                    "end": 1755.5,
                    "word": " the",
                    "probability": 0.81298828125
                },
                {
                    "start": 1755.5,
                    "end": 1755.72,
                    "word": " risk",
                    "probability": 0.806640625
                },
                {
                    "start": 1755.72,
                    "end": 1755.86,
                    "word": " of",
                    "probability": 0.9033203125
                },
                {
                    "start": 1755.86,
                    "end": 1756.14,
                    "word": " model",
                    "probability": 0.81298828125
                },
                {
                    "start": 1756.14,
                    "end": 1756.62,
                    "word": " theft",
                    "probability": 0.78515625
                },
                {
                    "start": 1756.62,
                    "end": 1757.14,
                    "word": ".",
                    "probability": 0.5322265625
                },
                {
                    "start": 1757.7,
                    "end": 1758.22,
                    "word": " These",
                    "probability": 0.83203125
                },
                {
                    "start": 1758.22,
                    "end": 1758.42,
                    "word": " two",
                    "probability": 0.798828125
                },
                {
                    "start": 1758.42,
                    "end": 1758.62,
                    "word": " kind",
                    "probability": 0.67578125
                },
                {
                    "start": 1758.62,
                    "end": 1758.7,
                    "word": " of",
                    "probability": 0.90576171875
                },
                {
                    "start": 1758.7,
                    "end": 1758.84,
                    "word": " go",
                    "probability": 0.908203125
                },
                {
                    "start": 1758.84,
                    "end": 1759.14,
                    "word": " hand",
                    "probability": 0.80224609375
                },
                {
                    "start": 1759.14,
                    "end": 1759.28,
                    "word": " in",
                    "probability": 0.6787109375
                },
                {
                    "start": 1759.28,
                    "end": 1759.66,
                    "word": " hand",
                    "probability": 0.7978515625
                },
                {
                    "start": 1759.66,
                    "end": 1759.88,
                    "word": ",",
                    "probability": 0.255859375
                },
                {
                    "start": 1759.88,
                    "end": 1759.98,
                    "word": " is",
                    "probability": 0.4140625
                },
                {
                    "start": 1759.98,
                    "end": 1760.22,
                    "word": ",",
                    "probability": 0.101806640625
                },
                {
                    "start": 1760.22,
                    "end": 1760.26,
                    "word": " like",
                    "probability": 0.84423828125
                },
                {
                    "start": 1760.26,
                    "end": 1760.68,
                    "word": ",",
                    "probability": 0.96728515625
                },
                {
                    "start": 1760.68,
                    "end": 1761.18,
                    "word": " you're",
                    "probability": 0.858154296875
                },
                {
                    "start": 1761.18,
                    "end": 1761.3,
                    "word": " going",
                    "probability": 0.666015625
                }
            ]
        },
        {
            "id": 249,
            "seek": 177432,
            "start": 1761.3,
            "end": 1766.88,
            "text": " to deploy these models. Somebody may try to steal them. Is there a risk if somebody steals your",
            "tokens": [
                50712,
                281,
                7274,
                613,
                5245,
                13,
                13463,
                815,
                853,
                281,
                11009,
                552,
                13,
                1119,
                456,
                257,
                3148,
                498,
                2618,
                46962,
                428,
                50996
            ],
            "temperature": 0.0,
            "avg_logprob": -0.2117883598450387,
            "compression_ratio": 1.668122270742358,
            "no_speech_prob": 0.005062103271484375,
            "words": [
                {
                    "start": 1761.3,
                    "end": 1761.4,
                    "word": " to",
                    "probability": 0.89990234375
                },
                {
                    "start": 1761.4,
                    "end": 1761.72,
                    "word": " deploy",
                    "probability": 0.74462890625
                },
                {
                    "start": 1761.72,
                    "end": 1761.96,
                    "word": " these",
                    "probability": 0.6865234375
                },
                {
                    "start": 1761.96,
                    "end": 1762.36,
                    "word": " models",
                    "probability": 0.79736328125
                },
                {
                    "start": 1762.36,
                    "end": 1762.88,
                    "word": ".",
                    "probability": 0.595703125
                },
                {
                    "start": 1763.38,
                    "end": 1763.64,
                    "word": " Somebody",
                    "probability": 0.845703125
                },
                {
                    "start": 1763.64,
                    "end": 1763.96,
                    "word": " may",
                    "probability": 0.89794921875
                },
                {
                    "start": 1763.96,
                    "end": 1764.2,
                    "word": " try",
                    "probability": 0.80908203125
                },
                {
                    "start": 1764.2,
                    "end": 1764.36,
                    "word": " to",
                    "probability": 0.90576171875
                },
                {
                    "start": 1764.36,
                    "end": 1764.58,
                    "word": " steal",
                    "probability": 0.83447265625
                },
                {
                    "start": 1764.58,
                    "end": 1764.82,
                    "word": " them",
                    "probability": 0.73974609375
                },
                {
                    "start": 1764.82,
                    "end": 1765.04,
                    "word": ".",
                    "probability": 0.9638671875
                },
                {
                    "start": 1765.04,
                    "end": 1765.16,
                    "word": " Is",
                    "probability": 0.90234375
                },
                {
                    "start": 1765.16,
                    "end": 1765.3,
                    "word": " there",
                    "probability": 0.8583984375
                },
                {
                    "start": 1765.54,
                    "end": 1765.84,
                    "word": " a risk",
                    "probability": 0.81005859375
                },
                {
                    "start": 1765.84,
                    "end": 1766.02,
                    "word": " if",
                    "probability": 0.86474609375
                },
                {
                    "start": 1766.02,
                    "end": 1766.32,
                    "word": " somebody",
                    "probability": 0.765625
                },
                {
                    "start": 1766.32,
                    "end": 1766.66,
                    "word": " steals",
                    "probability": 0.85791015625
                },
                {
                    "start": 1766.66,
                    "end": 1766.88,
                    "word": " your",
                    "probability": 0.7998046875
                }
            ]
        },
        {
            "id": 250,
            "seek": 177432,
            "start": 1766.88,
            "end": 1775.12,
            "text": " model? Like, does it matter? And if it doesn't matter, like, think about would it be okay to",
            "tokens": [
                50996,
                2316,
                30,
                1743,
                11,
                775,
                309,
                1871,
                30,
                400,
                498,
                309,
                1177,
                380,
                1871,
                11,
                411,
                11,
                519,
                466,
                576,
                309,
                312,
                1392,
                281,
                51412
            ],
            "temperature": 0.0,
            "avg_logprob": -0.2117883598450387,
            "compression_ratio": 1.668122270742358,
            "no_speech_prob": 0.005062103271484375,
            "words": [
                {
                    "start": 1766.88,
                    "end": 1767.16,
                    "word": " model",
                    "probability": 0.80224609375
                },
                {
                    "start": 1767.16,
                    "end": 1767.52,
                    "word": "?",
                    "probability": 0.93896484375
                },
                {
                    "start": 1767.52,
                    "end": 1767.68,
                    "word": " Like",
                    "probability": 0.5693359375
                },
                {
                    "start": 1767.68,
                    "end": 1767.68,
                    "word": ",",
                    "probability": 0.8994140625
                },
                {
                    "start": 1767.68,
                    "end": 1767.82,
                    "word": " does",
                    "probability": 0.91650390625
                },
                {
                    "start": 1767.82,
                    "end": 1767.98,
                    "word": " it",
                    "probability": 0.9033203125
                },
                {
                    "start": 1767.98,
                    "end": 1768.3,
                    "word": " matter",
                    "probability": 0.7958984375
                },
                {
                    "start": 1768.3,
                    "end": 1768.82,
                    "word": "?",
                    "probability": 0.984375
                },
                {
                    "start": 1769.64,
                    "end": 1770.16,
                    "word": " And",
                    "probability": 0.86767578125
                },
                {
                    "start": 1770.16,
                    "end": 1770.24,
                    "word": " if",
                    "probability": 0.88818359375
                },
                {
                    "start": 1770.24,
                    "end": 1770.36,
                    "word": " it",
                    "probability": 0.91015625
                },
                {
                    "start": 1770.36,
                    "end": 1770.6,
                    "word": " doesn't",
                    "probability": 0.93359375
                },
                {
                    "start": 1770.6,
                    "end": 1770.92,
                    "word": " matter",
                    "probability": 0.8046875
                },
                {
                    "start": 1770.92,
                    "end": 1771.36,
                    "word": ",",
                    "probability": 0.984375
                },
                {
                    "start": 1771.36,
                    "end": 1771.72,
                    "word": " like",
                    "probability": 0.814453125
                },
                {
                    "start": 1771.72,
                    "end": 1772.74,
                    "word": ",",
                    "probability": 0.92626953125
                },
                {
                    "start": 1772.74,
                    "end": 1773.08,
                    "word": " think",
                    "probability": 0.8115234375
                },
                {
                    "start": 1773.08,
                    "end": 1773.4,
                    "word": " about",
                    "probability": 0.822265625
                },
                {
                    "start": 1773.4,
                    "end": 1773.96,
                    "word": " would",
                    "probability": 0.580078125
                },
                {
                    "start": 1773.96,
                    "end": 1774.1,
                    "word": " it",
                    "probability": 0.90283203125
                },
                {
                    "start": 1774.1,
                    "end": 1774.32,
                    "word": " be",
                    "probability": 0.9013671875
                },
                {
                    "start": 1774.32,
                    "end": 1774.72,
                    "word": " okay",
                    "probability": 0.72802734375
                },
                {
                    "start": 1774.72,
                    "end": 1775.12,
                    "word": " to",
                    "probability": 0.90625
                }
            ]
        },
        {
            "id": 251,
            "seek": 177432,
            "start": 1775.12,
            "end": 1781.92,
            "text": " publish the source code and data for this model? And if your answer to that is no, then you really",
            "tokens": [
                51412,
                11374,
                264,
                4009,
                3089,
                293,
                1412,
                337,
                341,
                2316,
                30,
                400,
                498,
                428,
                1867,
                281,
                300,
                307,
                572,
                11,
                550,
                291,
                534,
                51752
            ],
            "temperature": 0.0,
            "avg_logprob": -0.2117883598450387,
            "compression_ratio": 1.668122270742358,
            "no_speech_prob": 0.005062103271484375,
            "words": [
                {
                    "start": 1775.12,
                    "end": 1775.7,
                    "word": " publish",
                    "probability": 0.85107421875
                },
                {
                    "start": 1775.7,
                    "end": 1775.96,
                    "word": " the",
                    "probability": 0.7880859375
                },
                {
                    "start": 1775.96,
                    "end": 1776.36,
                    "word": " source",
                    "probability": 0.71435546875
                },
                {
                    "start": 1776.36,
                    "end": 1776.74,
                    "word": " code",
                    "probability": 0.8349609375
                },
                {
                    "start": 1776.74,
                    "end": 1777.06,
                    "word": " and",
                    "probability": 0.8935546875
                },
                {
                    "start": 1777.06,
                    "end": 1777.36,
                    "word": " data",
                    "probability": 0.814453125
                },
                {
                    "start": 1777.36,
                    "end": 1777.58,
                    "word": " for",
                    "probability": 0.8837890625
                },
                {
                    "start": 1777.58,
                    "end": 1777.8,
                    "word": " this",
                    "probability": 0.8818359375
                },
                {
                    "start": 1777.8,
                    "end": 1778.14,
                    "word": " model",
                    "probability": 0.81005859375
                },
                {
                    "start": 1778.14,
                    "end": 1778.64,
                    "word": "?",
                    "probability": 0.7607421875
                },
                {
                    "start": 1778.64,
                    "end": 1778.92,
                    "word": " And",
                    "probability": 0.89013671875
                },
                {
                    "start": 1778.92,
                    "end": 1779.04,
                    "word": " if",
                    "probability": 0.9052734375
                },
                {
                    "start": 1779.04,
                    "end": 1779.18,
                    "word": " your",
                    "probability": 0.7998046875
                },
                {
                    "start": 1779.18,
                    "end": 1779.42,
                    "word": " answer",
                    "probability": 0.89111328125
                },
                {
                    "start": 1779.42,
                    "end": 1779.64,
                    "word": " to",
                    "probability": 0.89990234375
                },
                {
                    "start": 1779.64,
                    "end": 1779.82,
                    "word": " that",
                    "probability": 0.8818359375
                },
                {
                    "start": 1779.82,
                    "end": 1780.04,
                    "word": " is",
                    "probability": 0.90625
                },
                {
                    "start": 1780.04,
                    "end": 1780.58,
                    "word": " no",
                    "probability": 0.80322265625
                },
                {
                    "start": 1780.58,
                    "end": 1781.14,
                    "word": ",",
                    "probability": 0.9853515625
                },
                {
                    "start": 1781.14,
                    "end": 1781.36,
                    "word": " then",
                    "probability": 0.728515625
                },
                {
                    "start": 1781.36,
                    "end": 1781.52,
                    "word": " you",
                    "probability": 0.8955078125
                },
                {
                    "start": 1781.52,
                    "end": 1781.92,
                    "word": " really",
                    "probability": 0.76953125
                }
            ]
        },
        {
            "id": 252,
            "seek": 180120,
            "start": 1781.92,
            "end": 1789.28,
            "text": " need to make sure that infosec and operations are aware at deployment time. And finally, don't",
            "tokens": [
                50364,
                643,
                281,
                652,
                988,
                300,
                1536,
                541,
                66,
                293,
                7705,
                366,
                3650,
                412,
                19317,
                565,
                13,
                400,
                2721,
                11,
                500,
                380,
                50732
            ],
            "temperature": 0.0,
            "avg_logprob": -0.17420212369649968,
            "compression_ratio": 1.5720338983050848,
            "no_speech_prob": 0.0017004013061523438,
            "words": [
                {
                    "start": 1781.92,
                    "end": 1782.26,
                    "word": " need",
                    "probability": 0.73486328125
                },
                {
                    "start": 1782.26,
                    "end": 1782.38,
                    "word": " to",
                    "probability": 0.900390625
                },
                {
                    "start": 1782.38,
                    "end": 1782.62,
                    "word": " make",
                    "probability": 0.86865234375
                },
                {
                    "start": 1782.62,
                    "end": 1782.8,
                    "word": " sure",
                    "probability": 0.74365234375
                },
                {
                    "start": 1782.8,
                    "end": 1783.0,
                    "word": " that",
                    "probability": 0.7041015625
                },
                {
                    "start": 1783.0,
                    "end": 1783.64,
                    "word": " infosec",
                    "probability": 0.7496744791666666
                },
                {
                    "start": 1783.64,
                    "end": 1783.84,
                    "word": " and",
                    "probability": 0.86572265625
                },
                {
                    "start": 1783.84,
                    "end": 1784.38,
                    "word": " operations",
                    "probability": 0.87255859375
                },
                {
                    "start": 1784.38,
                    "end": 1784.74,
                    "word": " are",
                    "probability": 0.88818359375
                },
                {
                    "start": 1784.74,
                    "end": 1785.18,
                    "word": " aware",
                    "probability": 0.79736328125
                },
                {
                    "start": 1785.18,
                    "end": 1786.02,
                    "word": " at",
                    "probability": 0.85595703125
                },
                {
                    "start": 1786.02,
                    "end": 1786.46,
                    "word": " deployment",
                    "probability": 0.73193359375
                },
                {
                    "start": 1786.46,
                    "end": 1786.92,
                    "word": " time",
                    "probability": 0.8037109375
                },
                {
                    "start": 1786.92,
                    "end": 1787.58,
                    "word": ".",
                    "probability": 0.93798828125
                },
                {
                    "start": 1787.94,
                    "end": 1788.1,
                    "word": " And",
                    "probability": 0.7890625
                },
                {
                    "start": 1788.1,
                    "end": 1788.52,
                    "word": " finally",
                    "probability": 0.70458984375
                },
                {
                    "start": 1788.52,
                    "end": 1789.0,
                    "word": ",",
                    "probability": 0.9423828125
                },
                {
                    "start": 1789.0,
                    "end": 1789.28,
                    "word": " don't",
                    "probability": 0.951171875
                }
            ]
        },
        {
            "id": 253,
            "seek": 180120,
            "start": 1789.28,
            "end": 1796.18,
            "text": " hype people up over text generation models. Like, GPT-2 and GPT-3 are super cool from an NLP",
            "tokens": [
                50732,
                24144,
                561,
                493,
                670,
                2487,
                5125,
                5245,
                13,
                1743,
                11,
                26039,
                51,
                12,
                17,
                293,
                26039,
                51,
                12,
                18,
                366,
                1687,
                1627,
                490,
                364,
                426,
                45196,
                51080
            ],
            "temperature": 0.0,
            "avg_logprob": -0.17420212369649968,
            "compression_ratio": 1.5720338983050848,
            "no_speech_prob": 0.0017004013061523438,
            "words": [
                {
                    "start": 1789.28,
                    "end": 1789.58,
                    "word": " hype",
                    "probability": 0.86279296875
                },
                {
                    "start": 1789.58,
                    "end": 1789.88,
                    "word": " people",
                    "probability": 0.91259765625
                },
                {
                    "start": 1789.88,
                    "end": 1790.06,
                    "word": " up",
                    "probability": 0.88671875
                },
                {
                    "start": 1790.06,
                    "end": 1790.28,
                    "word": " over",
                    "probability": 0.8603515625
                },
                {
                    "start": 1790.28,
                    "end": 1790.52,
                    "word": " text",
                    "probability": 0.85400390625
                },
                {
                    "start": 1790.52,
                    "end": 1790.9,
                    "word": " generation",
                    "probability": 0.77490234375
                },
                {
                    "start": 1790.9,
                    "end": 1791.38,
                    "word": " models",
                    "probability": 0.779296875
                },
                {
                    "start": 1791.38,
                    "end": 1791.78,
                    "word": ".",
                    "probability": 0.92041015625
                },
                {
                    "start": 1791.78,
                    "end": 1791.86,
                    "word": " Like",
                    "probability": 0.62158203125
                },
                {
                    "start": 1791.86,
                    "end": 1791.94,
                    "word": ",",
                    "probability": 0.4619140625
                },
                {
                    "start": 1791.94,
                    "end": 1792.26,
                    "word": " GPT",
                    "probability": 0.908203125
                },
                {
                    "start": 1792.26,
                    "end": 1792.58,
                    "word": "-2",
                    "probability": 0.7216796875
                },
                {
                    "start": 1792.58,
                    "end": 1792.92,
                    "word": " and",
                    "probability": 0.88037109375
                },
                {
                    "start": 1792.92,
                    "end": 1793.28,
                    "word": " GPT",
                    "probability": 0.942626953125
                },
                {
                    "start": 1793.28,
                    "end": 1793.54,
                    "word": "-3",
                    "probability": 0.993896484375
                },
                {
                    "start": 1793.54,
                    "end": 1793.8,
                    "word": " are",
                    "probability": 0.88623046875
                },
                {
                    "start": 1793.8,
                    "end": 1794.24,
                    "word": " super",
                    "probability": 0.8623046875
                },
                {
                    "start": 1794.24,
                    "end": 1794.72,
                    "word": " cool",
                    "probability": 0.7958984375
                },
                {
                    "start": 1794.72,
                    "end": 1795.68,
                    "word": " from",
                    "probability": 0.72314453125
                },
                {
                    "start": 1795.68,
                    "end": 1795.84,
                    "word": " an",
                    "probability": 0.83544921875
                },
                {
                    "start": 1795.84,
                    "end": 1796.18,
                    "word": " NLP",
                    "probability": 0.955810546875
                }
            ]
        },
        {
            "id": 254,
            "seek": 180120,
            "start": 1796.18,
            "end": 1802.1,
            "text": " standpoint, natural language processing, but they're not going to change the security landscape.",
            "tokens": [
                51080,
                15827,
                11,
                3303,
                2856,
                9007,
                11,
                457,
                436,
                434,
                406,
                516,
                281,
                1319,
                264,
                3825,
                9661,
                13,
                51368
            ],
            "temperature": 0.0,
            "avg_logprob": -0.17420212369649968,
            "compression_ratio": 1.5720338983050848,
            "no_speech_prob": 0.0017004013061523438,
            "words": [
                {
                    "start": 1796.18,
                    "end": 1796.78,
                    "word": " standpoint",
                    "probability": 0.73974609375
                },
                {
                    "start": 1796.78,
                    "end": 1797.08,
                    "word": ",",
                    "probability": 0.86572265625
                },
                {
                    "start": 1797.08,
                    "end": 1797.36,
                    "word": " natural",
                    "probability": 0.79150390625
                },
                {
                    "start": 1797.36,
                    "end": 1797.68,
                    "word": " language",
                    "probability": 0.63671875
                },
                {
                    "start": 1797.68,
                    "end": 1798.16,
                    "word": " processing",
                    "probability": 0.76953125
                },
                {
                    "start": 1798.16,
                    "end": 1798.86,
                    "word": ",",
                    "probability": 0.47998046875
                },
                {
                    "start": 1798.86,
                    "end": 1799.72,
                    "word": " but",
                    "probability": 0.89697265625
                },
                {
                    "start": 1799.72,
                    "end": 1800.12,
                    "word": " they're",
                    "probability": 0.6409912109375
                },
                {
                    "start": 1800.12,
                    "end": 1800.3,
                    "word": " not",
                    "probability": 0.89501953125
                },
                {
                    "start": 1800.3,
                    "end": 1800.44,
                    "word": " going",
                    "probability": 0.80224609375
                },
                {
                    "start": 1800.44,
                    "end": 1800.52,
                    "word": " to",
                    "probability": 0.90771484375
                },
                {
                    "start": 1800.52,
                    "end": 1800.72,
                    "word": " change",
                    "probability": 0.82080078125
                },
                {
                    "start": 1800.72,
                    "end": 1800.82,
                    "word": " the",
                    "probability": 0.80126953125
                },
                {
                    "start": 1800.82,
                    "end": 1801.14,
                    "word": " security",
                    "probability": 0.94287109375
                },
                {
                    "start": 1801.14,
                    "end": 1801.56,
                    "word": " landscape",
                    "probability": 0.75830078125
                },
                {
                    "start": 1801.56,
                    "end": 1802.1,
                    "word": ".",
                    "probability": 0.953125
                }
            ]
        },
        {
            "id": 255,
            "seek": 180120,
            "start": 1802.1,
            "end": 1808.8,
            "text": " Stop scaring people, please. I'm literally begging you, don't hype people up over text",
            "tokens": [
                51368,
                5535,
                795,
                1921,
                561,
                11,
                1767,
                13,
                286,
                478,
                3736,
                26600,
                291,
                11,
                500,
                380,
                24144,
                561,
                493,
                670,
                2487,
                51704
            ],
            "temperature": 0.0,
            "avg_logprob": -0.17420212369649968,
            "compression_ratio": 1.5720338983050848,
            "no_speech_prob": 0.0017004013061523438,
            "words": [
                {
                    "start": 1802.1,
                    "end": 1802.38,
                    "word": " Stop",
                    "probability": 0.8466796875
                },
                {
                    "start": 1802.38,
                    "end": 1802.94,
                    "word": " scaring",
                    "probability": 0.921142578125
                },
                {
                    "start": 1802.94,
                    "end": 1803.48,
                    "word": " people",
                    "probability": 0.92724609375
                },
                {
                    "start": 1803.48,
                    "end": 1803.7,
                    "word": ",",
                    "probability": 0.82177734375
                },
                {
                    "start": 1803.7,
                    "end": 1804.02,
                    "word": " please",
                    "probability": 0.7353515625
                },
                {
                    "start": 1804.02,
                    "end": 1804.38,
                    "word": ".",
                    "probability": 0.94189453125
                },
                {
                    "start": 1804.42,
                    "end": 1805.08,
                    "word": " I'm",
                    "probability": 0.98681640625
                },
                {
                    "start": 1805.08,
                    "end": 1805.66,
                    "word": " literally",
                    "probability": 0.708984375
                },
                {
                    "start": 1805.66,
                    "end": 1806.02,
                    "word": " begging",
                    "probability": 0.69970703125
                },
                {
                    "start": 1806.02,
                    "end": 1806.44,
                    "word": " you",
                    "probability": 0.9013671875
                },
                {
                    "start": 1806.44,
                    "end": 1807.02,
                    "word": ",",
                    "probability": 0.5517578125
                },
                {
                    "start": 1807.02,
                    "end": 1807.76,
                    "word": " don't",
                    "probability": 0.948486328125
                },
                {
                    "start": 1807.76,
                    "end": 1807.96,
                    "word": " hype",
                    "probability": 0.88134765625
                },
                {
                    "start": 1807.96,
                    "end": 1808.22,
                    "word": " people",
                    "probability": 0.91943359375
                },
                {
                    "start": 1808.22,
                    "end": 1808.38,
                    "word": " up",
                    "probability": 0.90234375
                },
                {
                    "start": 1808.38,
                    "end": 1808.58,
                    "word": " over",
                    "probability": 0.87158203125
                },
                {
                    "start": 1808.58,
                    "end": 1808.8,
                    "word": " text",
                    "probability": 0.88720703125
                }
            ]
        },
        {
            "id": 256,
            "seek": 181939,
            "start": 1808.8,
            "end": 1815.68,
            "text": " generation models. Here are some references. These are some of the things I talked about.",
            "tokens": [
                50364,
                5125,
                5245,
                13,
                1692,
                366,
                512,
                15400,
                13,
                1981,
                366,
                512,
                295,
                264,
                721,
                286,
                2825,
                466,
                13,
                50716
            ],
            "temperature": 0.0,
            "avg_logprob": -0.21814123995892412,
            "compression_ratio": 1.53125,
            "no_speech_prob": 0.00490570068359375,
            "words": [
                {
                    "start": 1808.8,
                    "end": 1809.2,
                    "word": " generation",
                    "probability": 0.7138671875
                },
                {
                    "start": 1809.2,
                    "end": 1809.76,
                    "word": " models",
                    "probability": 0.7939453125
                },
                {
                    "start": 1809.76,
                    "end": 1810.12,
                    "word": ".",
                    "probability": 0.9716796875
                },
                {
                    "start": 1810.86,
                    "end": 1811.1,
                    "word": " Here",
                    "probability": 0.78369140625
                },
                {
                    "start": 1811.1,
                    "end": 1811.22,
                    "word": " are",
                    "probability": 0.87255859375
                },
                {
                    "start": 1811.22,
                    "end": 1811.38,
                    "word": " some",
                    "probability": 0.8203125
                },
                {
                    "start": 1811.38,
                    "end": 1811.86,
                    "word": " references",
                    "probability": 0.80712890625
                },
                {
                    "start": 1811.86,
                    "end": 1812.22,
                    "word": ".",
                    "probability": 0.7861328125
                },
                {
                    "start": 1812.86,
                    "end": 1813.08,
                    "word": " These",
                    "probability": 0.87060546875
                },
                {
                    "start": 1813.08,
                    "end": 1813.26,
                    "word": " are",
                    "probability": 0.8984375
                },
                {
                    "start": 1813.26,
                    "end": 1813.64,
                    "word": " some",
                    "probability": 0.8125
                },
                {
                    "start": 1813.64,
                    "end": 1814.0,
                    "word": " of",
                    "probability": 0.9033203125
                },
                {
                    "start": 1814.0,
                    "end": 1814.3,
                    "word": " the",
                    "probability": 0.81640625
                },
                {
                    "start": 1814.3,
                    "end": 1814.68,
                    "word": " things",
                    "probability": 0.74755859375
                },
                {
                    "start": 1814.68,
                    "end": 1814.82,
                    "word": " I",
                    "probability": 0.98193359375
                },
                {
                    "start": 1814.82,
                    "end": 1815.08,
                    "word": " talked",
                    "probability": 0.748046875
                },
                {
                    "start": 1815.08,
                    "end": 1815.32,
                    "word": " about",
                    "probability": 0.82421875
                },
                {
                    "start": 1815.32,
                    "end": 1815.68,
                    "word": ".",
                    "probability": 0.6376953125
                }
            ]
        },
        {
            "id": 257,
            "seek": 181939,
            "start": 1816.0,
            "end": 1820.38,
            "text": " And thank you so much for attending my talk. I'll be in the Discord most of the weekend,",
            "tokens": [
                50716,
                400,
                1309,
                291,
                370,
                709,
                337,
                15862,
                452,
                751,
                13,
                286,
                603,
                312,
                294,
                264,
                32623,
                881,
                295,
                264,
                6711,
                11,
                50936
            ],
            "temperature": 0.0,
            "avg_logprob": -0.21814123995892412,
            "compression_ratio": 1.53125,
            "no_speech_prob": 0.00490570068359375,
            "words": [
                {
                    "start": 1816.0,
                    "end": 1816.14,
                    "word": " And",
                    "probability": 0.7509765625
                },
                {
                    "start": 1816.14,
                    "end": 1816.36,
                    "word": " thank",
                    "probability": 0.80224609375
                },
                {
                    "start": 1816.36,
                    "end": 1816.46,
                    "word": " you",
                    "probability": 0.89306640625
                },
                {
                    "start": 1816.46,
                    "end": 1816.58,
                    "word": " so",
                    "probability": 0.9072265625
                },
                {
                    "start": 1816.58,
                    "end": 1816.72,
                    "word": " much",
                    "probability": 0.8212890625
                },
                {
                    "start": 1816.72,
                    "end": 1816.88,
                    "word": " for",
                    "probability": 0.89599609375
                },
                {
                    "start": 1816.88,
                    "end": 1817.12,
                    "word": " attending",
                    "probability": 0.80859375
                },
                {
                    "start": 1817.12,
                    "end": 1817.38,
                    "word": " my",
                    "probability": 0.9052734375
                },
                {
                    "start": 1817.38,
                    "end": 1817.68,
                    "word": " talk",
                    "probability": 0.77783203125
                },
                {
                    "start": 1817.68,
                    "end": 1818.04,
                    "word": ".",
                    "probability": 0.9404296875
                },
                {
                    "start": 1818.08,
                    "end": 1818.44,
                    "word": " I'll",
                    "probability": 0.95556640625
                },
                {
                    "start": 1818.44,
                    "end": 1818.6,
                    "word": " be",
                    "probability": 0.904296875
                },
                {
                    "start": 1818.6,
                    "end": 1818.72,
                    "word": " in",
                    "probability": 0.90478515625
                },
                {
                    "start": 1818.72,
                    "end": 1818.84,
                    "word": " the",
                    "probability": 0.810546875
                },
                {
                    "start": 1818.84,
                    "end": 1819.22,
                    "word": " Discord",
                    "probability": 0.591796875
                },
                {
                    "start": 1819.22,
                    "end": 1819.66,
                    "word": " most",
                    "probability": 0.78759765625
                },
                {
                    "start": 1819.66,
                    "end": 1819.8,
                    "word": " of",
                    "probability": 0.90380859375
                },
                {
                    "start": 1819.8,
                    "end": 1819.9,
                    "word": " the",
                    "probability": 0.8115234375
                },
                {
                    "start": 1819.9,
                    "end": 1820.2,
                    "word": " weekend",
                    "probability": 0.7822265625
                },
                {
                    "start": 1820.2,
                    "end": 1820.38,
                    "word": ",",
                    "probability": 0.364990234375
                }
            ]
        },
        {
            "id": 258,
            "seek": 181939,
            "start": 1820.38,
            "end": 1825.0,
            "text": " so if you have any questions, feel free to drop me a line. Thank you so much. Have a great weekend.",
            "tokens": [
                50936,
                370,
                498,
                291,
                362,
                604,
                1651,
                11,
                841,
                1737,
                281,
                3270,
                385,
                257,
                1622,
                13,
                1044,
                291,
                370,
                709,
                13,
                3560,
                257,
                869,
                6711,
                13,
                51168
            ],
            "temperature": 0.0,
            "avg_logprob": -0.21814123995892412,
            "compression_ratio": 1.53125,
            "no_speech_prob": 0.00490570068359375,
            "words": [
                {
                    "start": 1820.38,
                    "end": 1820.52,
                    "word": " so",
                    "probability": 0.89404296875
                },
                {
                    "start": 1820.52,
                    "end": 1820.62,
                    "word": " if",
                    "probability": 0.89208984375
                },
                {
                    "start": 1820.62,
                    "end": 1820.68,
                    "word": " you",
                    "probability": 0.89599609375
                },
                {
                    "start": 1820.68,
                    "end": 1820.78,
                    "word": " have",
                    "probability": 0.884765625
                },
                {
                    "start": 1820.78,
                    "end": 1820.92,
                    "word": " any",
                    "probability": 0.8193359375
                },
                {
                    "start": 1820.92,
                    "end": 1821.36,
                    "word": " questions",
                    "probability": 0.90576171875
                },
                {
                    "start": 1821.36,
                    "end": 1821.68,
                    "word": ",",
                    "probability": 0.80859375
                },
                {
                    "start": 1821.68,
                    "end": 1821.84,
                    "word": " feel",
                    "probability": 0.91357421875
                },
                {
                    "start": 1821.84,
                    "end": 1822.0,
                    "word": " free",
                    "probability": 0.8642578125
                },
                {
                    "start": 1822.0,
                    "end": 1822.18,
                    "word": " to",
                    "probability": 0.9091796875
                },
                {
                    "start": 1822.18,
                    "end": 1822.38,
                    "word": " drop",
                    "probability": 0.81640625
                },
                {
                    "start": 1822.38,
                    "end": 1822.54,
                    "word": " me",
                    "probability": 0.90283203125
                },
                {
                    "start": 1822.66,
                    "end": 1822.88,
                    "word": " a line",
                    "probability": 0.8798828125
                },
                {
                    "start": 1822.88,
                    "end": 1823.24,
                    "word": ".",
                    "probability": 0.9951171875
                },
                {
                    "start": 1823.3,
                    "end": 1823.48,
                    "word": " Thank",
                    "probability": 0.8798828125
                },
                {
                    "start": 1823.48,
                    "end": 1823.58,
                    "word": " you",
                    "probability": 0.896484375
                },
                {
                    "start": 1823.58,
                    "end": 1823.72,
                    "word": " so",
                    "probability": 0.9072265625
                },
                {
                    "start": 1823.72,
                    "end": 1823.98,
                    "word": " much",
                    "probability": 0.81201171875
                },
                {
                    "start": 1823.98,
                    "end": 1824.16,
                    "word": ".",
                    "probability": 0.87451171875
                },
                {
                    "start": 1824.16,
                    "end": 1824.3,
                    "word": " Have",
                    "probability": 0.88134765625
                },
                {
                    "start": 1824.4,
                    "end": 1824.54,
                    "word": " a great",
                    "probability": 0.87744140625
                },
                {
                    "start": 1824.54,
                    "end": 1824.84,
                    "word": " weekend",
                    "probability": 0.7958984375
                },
                {
                    "start": 1824.84,
                    "end": 1825.0,
                    "word": ".",
                    "probability": 0.7744140625
                }
            ]
        },
        {
            "id": 259,
            "seek": 181939,
            "start": 1825.0,
            "end": 1825.84,
            "text": " Enjoy your con.",
            "tokens": [
                51168,
                15411,
                428,
                416,
                13,
                51224
            ],
            "temperature": 0.0,
            "avg_logprob": -0.21814123995892412,
            "compression_ratio": 1.53125,
            "no_speech_prob": 0.00490570068359375,
            "words": [
                {
                    "start": 1825.0,
                    "end": 1825.18,
                    "word": " Enjoy",
                    "probability": 0.90625
                },
                {
                    "start": 1825.18,
                    "end": 1825.34,
                    "word": " your",
                    "probability": 0.77978515625
                },
                {
                    "start": 1825.34,
                    "end": 1825.56,
                    "word": " con",
                    "probability": 0.36572265625
                },
                {
                    "start": 1825.56,
                    "end": 1825.84,
                    "word": ".",
                    "probability": 0.912109375
                }
            ]
        }
    ],
    "text": " Here's our second talk of the day. It's Eric Gallican. He's a... well, you see his Who Am I page, but he's also a PhD student at Brandeis and an all-around awesome guy. He does way too much stuff and is very enthusiastic and involved in the community. So I'm just going to let him talk about... on his intro talk about... for machine learning people. Hi, everybody. Welcome to my talk. This is Baby's First 100 MLSat Words. My name is Eric Gallican. I'm an artificial intelligence researcher at Rapid7. I specialize in breaking machine learning stuff and applying machine learning to detecting bad things. I also do research at the Montreal AI Ethics Institute on applying DevOps principle to machine learning ethics. And last year, I was here at DEF CON in person, not here in my house at DEF CON, presenting at the Cloud Village on a piece of malware that I wrote. Malware, not actually malware, just in case the legal team is listening. And I presented that piece of proof of concept malware called Sassy Boy. But mostly I do boring math stuff. And if you can't see my speaker video for some reason, I put a picture of my face on the right so that you can see it. Or you can see it twice. You can see that my hair hasn't changed basically at all in years. So before we get started, let's just answer the question, what is MLSat? Because I get asked to that a lot by practitioners in security and in machine learning, honestly. And mostly it's an argument. And it's an argument between three groups of people, as it usually is. And the first group of people is people who think it's only applications of machine learning and artificial intelligence to security problems. So that's automated malware detection. That's next generation SIMs. That's automated red teaming and that sort of stuff. And then there are people who think that it's just securing artificial intelligence in machine learning systems. And then there are those of us who are correct who think it's both. It's kind of any intersection of artificial intelligence, right? That broad category of machine learning slash AI and security. And since this is my talk, and since I am the Mao Zedong of MLSat, I get to decide that it's both. And if anybody argues with you or you hear anybody having this totally asinine argument, you can just tell them that I said it's both. And then the argument can be over and everybody can do more productive things with their lives. So we'll start with deep fakes, right? So D is for deep fakes. And this man on the right, if I didn't have a speaker video, I probably could have just put that picture up and said it was me. And everybody would have believed it because he's got kind of he looks like the kind of guy, right? And I'm allowed to say that because he's not real. This is not a real person. This is completely an AI-generated image. And so, you know, to give deep fakes a real definition, they are a convincing synthetic image, video, or audio recording which purports to be real. So, you know, I'm pretty convinced by this picture. If I saw it in the wild, right, if somebody put it as like their Twitter avatar and was like, yeah, I am a software engineer at Google, I could buy that. I could buy that based on this image, right? You know? We have our stereotypes, right? The thin-rimmed glasses, the scruff. And it's super convincing. It's super convincing. And I think that's really, when we think about impact, that's what it is. It's that they're convincing to users. So, we also see deep fakes in the text generation space where it purports to be a real email. Or we see deep fake writing which mimics the style of famous authors. We see deep fake audio recordings which sound like the person who they are trained to sound like. Now, there's a piece of commercial software out there called Liarbird which does this really well. And so, you know, they're very convincing to users. And as a result, you know, they get used widely and they're easy to make, right? That's the other thing, is they don't take as much time as finding and crafting an image or carefully learning a style or mimicking a voice, right? You just kind of train the model and push them out. So, they're easy to make and they're convincing to users. Which means that they can have real impact. But that impact is mostly societal, right? So, there was a famous, you know, semi-viral video of Jordan Peele doing a deep fake where it was an image of President Obama speaking. And you can look it up. It's really easy to find. And so, like, the risk there is really that broad swaths of people will believe that a doctored or deep faked video is real. And so, there's something like President Donald J. Trump, you know, declaring war on North Korea on a hot mic or suggesting that he might invade North Korea on a hot mic. That could be really convincing to people. And so, that's the real risk of deep fakes, is that they are a powerful disinformation tool which is powered by artificial intelligence. And so, you know, our mitigation against deep fakes is kind of, eh. Like, there's very little we can do about deep fakes. They're difficult to detect. They're very convincing to users. They're easy to make. And so, really, it's a broader social problem that we need to tackle of having people be inherently less trustful of information sources and really vet where they get their information from, which is a hard problem. But this isn't an enterprise security threat. This is a social threat. So, moving back to thinking about deployed models, we'll look at adversarial examples. And this gif over here on the right is from a pretty well-known example that was really, really hyped up. And it was the turtle rifle paper. Because the Google computer vision API is what they attacked. And they 3D printed a turtle. And this is pretty obviously a turtle to anybody watching who isn't powered by the Google Vision API. And if you are an artificial intelligence powered by the Google Vision API, I'm so sorry for the rest of this talk. Because you're not gonna like it. That, you know, it's an input to the classifier that's specifically crafted to force misclassification. And so, this object as a turtle, just a little, like, toy turtle, right? Being misclassified as a rifle, the risk there is if you have a CCTV system that's powered by Google Vision API, which seems reasonable to license, that's looking for potential security threats. And, you know, you don't input all of that, like, racist physiognomy that we already see in a lot of those systems. But instead just say, you know what? Let's just look for weapons. Well, this could falsely trigger those systems. And so, there's a risk there, too. Correspondingly, if you 3D printed a gun, right? That got misclassified as a turtle, that would also be a threat, right? So, the impact, really, is that it causes misclassification. And so, that real world security threat and 3D printed examples are really rare, and they're really difficult to do. And they don't tend to work in all situations, and any researcher in computer vision can explain to you why it's not a real part of the threat model. And that's fine. But our threat is that it causes misclassification. As security practitioners, we try to look for ways to use technologies to detect bad things. And so, if you integrate artificial intelligence into your malware detector, for example, an attacker can bypass detection using adversarial examples. And the thing with adversarial examples is when we think about how to mitigate them, one of the only truly effective ways to do it is through what's called adversarial training, where you basically show it a bunch of adversarial examples and tell it, like, no, classify these correctly. And that's really time consuming, and it's expensive, and it's hard. And so, not a lot of people do it. And it requires you to take your model out of production and put it back into training, which, again, has a business impact. So, one of the other ways to sort of mitigate it is through what's called ensembling, where you don't just take one canonical model. You take several models and sort of average their output or take a weighted average of their output or whatever. And so, that's another way to do it. And that's kind of because it's a lot harder to come up with an adversarial example that works across a variety of classifiers, especially if they're not trained on the same data, if they don't have the same architecture. And it's still possible to do, but it sort of raises that barrier to entry. And if we think about it in terms of, like, cryptography, that's really what we're trying to do, right? Is we know that there is hypothetically always an attack, right? You can brute force the key space. That's a hypothetically feasible attack against AES, right? But it's good enough, and it takes long enough to brute force that key space that, you know, it's fine. It does the job, right? And that's sort of the same thing here, is you just have to raise that barrier to entry to make it not worth it for an attacker. So, the next thing we want to talk about in terms of threats to deployed models are backdoors. And so, backdoors in a machine learning context are manipulations of trained model weights that result in a specific outcome each time. So, a common way to do this in neural networks is bias poisoning, where you take one class and you make the bias for that one class in just the output layer really, really, really big. So, you're only changing one weight in the whole network. So, it doesn't change a lot, right? But by virtue of making that weight really big, you'll always get the same classification. And if it's a binary classifier, right, say it's, again, our malware detector, because people love to make neural network powered malware detectors for whatever reason. And I've made one. So, I'm allowed to criticize. People love to do it. And that's a really easy way to just be like, nope, everything's benign all the time. Always. 100% of the time. So, you know, the impact is that it causes misclassification. And I say but not really, because really what it does is remove a class or remove all other classes from the classifier. So, it's not really misclassification. It's just making it a one class classifier. And our mitigation is just don't let attackers get access to your model. And by your model, I mean the trained model weights. And I say this kind of flippantly, because if an attacker has the ability to manipulate your model weights, where you're hosting your model, they can do way worse things than backdoor your model. So, it's not... it's like a local privilege escalation that requires you to have admin to begin with. It's not really a threat. So, again, attacks against deployed models. There's model theft. And I want to talk about model theft and give a lot of credit to Will Pierce, formerly of Silent Break Security, who I stole this graphic from his fantastic DerbyCon presentation last year. Was given the first CVE for machine learning, which is super dope. And so, model theft is the process of creating a copycat model by querying a trained model. Right? So, essentially, what we're doing is we're hitting a model over and over and over again. And using the outputs of that trained model, we train our own model to give us an approximation of that model. So, you know, if you have a box and you put in a 1 and you get out a 2, and then you have a second box, you want to put in the same input and get the same output. And it doesn't really matter what happens inside of that box as much as it matters that whatever input you give it, you get the same output. And that's what happens with model theft. So, our impact here is kind of twofold. And so, the potential for adversarial attacks in model theft goes way up. Because somebody has a model they can specifically query against to validate whether or not it gets misclassified without interacting with your model. So, it gives them sort of a private development environment for these adversarial attacks. And that's what Will did in the proof putting. Is he created this copycat model so that he could test phishing emails to bypass the proof point email security appliance. And it was incredibly clever. Super dope. I recommend watching his talk. But after you watch all the other talks at the AI Village. Maybe, like, after DEF CON, watch it. It's a really good talk. The other impact, though, is the loss of intellectual property. So, one of the things that I think kind of got overlooked in how cool the ability to bypass security appliances is, is that if you hire a bunch of data scientists and collect a bunch of data and clean a bunch of data and spend years creating and building and deploying a model that really differentiates you from the competition, and some unscrupulous company in not Canada, you know, steals your model, approximates your model, and just shoves it into their product and says, yeah, we do the same thing. Well, it costs them a lot less money to just copy your model. It costs them almost nothing in the grand scheme of things to copy your model. And so, there's no moat, so to speak. So, it really hurts your ability to create a differentiated product. So, this is a real vulnerability from that standpoint. And we'll get one layer deeper in our next slide. So, when we think about mitigations for model theft, right, there's two. And the first one is limiting queries to the model. And this one kind of feels bad. Because you don't want to limit queries to your model too much, since most queries to your model are going to be legitimate. For the most part, you're expecting these inputs and giving outputs, and you created this model, and you're hosting this model, because it's supposed to be useful to someone. And for most people, because of just the corporatization of information security and the corporatization of artificial intelligence, it's probably the people who pay your paycheck who want this. And so, inherently, you want to have really, but you want to balance that really high uptime and the fact that there may be a single endpoint that's making a lot of legitimate queries to your model with the fact that there may be an attacker who's trying to use your model for evil. Trying to steal your model. And so, the other thing we can do is limiting information returned from the model. And we saw the efficacy of this with SQL injection, where, like, back in the olden days, you know, a million years ago, because it's 2020, and so time doesn't matter anymore, we would return these detailed error messages for SQL queries, and it made it really, really easy for attackers to find the information they were looking for in SQL databases, especially SQL databases that didn't do a great job of input validation. And so, you know, once we started returning basically there was an error as the error message, it became much harder, right? And it's still feasible to do, but it became much harder. And it's similar here. If all that was returned was a blocked or not blocked, that binary signal versus a detailed header, then it would be much more difficult to create a copycat model. And so, you know, that is a real mitigation. And when we think about model theft, kind of the next step is inversion. Model inversion. And what model inversion is, is recovering training data from a trained model. So, there was a paper by Nick Carlini that came out two years ago, I think, in 2018, about how neural networks unintentionally memorize specific training examples. There have been papers on model inversion, both traditional and using generative adversarial networks that have shown that it's pretty effective. And so, the impact here is the loss of data, right? You're losing training data. And maybe that doesn't matter a ton. So, like at Rapid7, we have our open dataset, right? And that's free and open for anyone to use for noncommercial purposes, and we make it available for, you know, other stuff, right? And so, like, if our model was trained on open data, and somebody inverted our model, we probably wouldn't care that much, because, like, that data's out there, right? But if you are training a model for, like, a large medical device manufacturer, and it's trained on sensitive medical information, and that information gets recovered, well, that's a lot worse. If you're a data scientist at Equifax or some other credit rating bureau, and you have a bunch of sensitive financial data that you train a model on, that's pretty terrifying, right? The penalties for losing everybody's data at Equifax could be huge. You could have to pay a bunch of people's credit monitoring service for a while, and that would be bad for your business. So, when we think about mitigations, really, the first one is to protect access to the model. So, if they have direct access to the model, it's much easier to do inversion. It's much, much easier to do inversion. There are different techniques you can use if you have direct access to the model. And then the other mitigations for model inversion are essentially don't let people steal your model. Because then it's easier to use it as sort of the discriminator in the GAN. The architecture is a little more complicated than a traditional generative adversarial network, but the idea is pretty similar. You basically show it an example and say, does this look familiar? And it goes yes or no. And if it says yes, then you may have inverted the information. So, our next threat is poisoning. Data poisoning. And data poisoning is when malicious users inject bad training data into a model to corrupt it. So, usually, this happens in online learning where the model is continuously trained on the input to the model. And that is definitely a threat that needs to be considered. The other option is that they can, you know, get access to wherever you store your data and just inject bad examples. So, our impact is misclassification. It sort of shifts our decision boundary to deliberately misclassify samples. So, you know, letting spam through because now your threshold for what is spam is so high because you've just been doing online learning, right? And so, when we mitigate it, right, we want to go back and think about protecting access to our data. So, we want data integrity, right? So, having data versioning is really important. Making sure that your training data has a version, has an associated hash, has a label. That's super important. And then our next is we don't want to just allow unvalidated, untrusted input to our model. That's something we see a ton is people who just accept whatever input with no validation on the front end. It's just whatever gets input to the API gets pushed into the model. And that can really cause problems when you're retraining if you just take, you know, those inputs that you get and sort of store them off for training later. So, you know, we're coming up on the end of this talk. And so, we want to talk about, like, what can I do, right? And this is sort of a choose your own adventure story. It depends on your role. So, if you're a hacker, and by hacker, I mean this in the colloquial sense, a black hat, or a red teamer, right? Black hats are already doing some of this stuff. So, I really mean, you know, if you're a red teamer, if you're catching up to the elite APT black hat hackers, right? Take advantage of the lack of defenses on machine learning systems. Microsoft put out a report earlier this year saying that in their survey of large companies and government organizations out of, I think it was 28, only three had any meaningful defenses on their machine learning systems. So, if you're a red teamer, like, check it out. If you're a vulnerability discovery person, like, check it out. You can definitely find stuff. And that's because nobody knows what you're doing, and nobody's looking for you, right? Nobody is looking for black hat hackers, right? APTs, whoever. And if you are, let me know. Yell at me in the discord, and I'll be like, oh damn, cool. If you're a defender, if you're a blue teamer, right? If you're a sock monkey, if you're a researcher, if you're at DEF CON, you're probably one of these two categories. Test your machine learning systems as if they're part of your infrastructure. So, what we see a lot is that machine learning engineers and data scientists develop these models, and they take these models in, like, a .py file, and they hand it off to engineering, and engineering goes, I don't know what this is, and they slap a Django API in front of it, put it in a Docker container, and deploy it in Kubernetes as an API, and then when InfoSec looks at it, they go, I don't know what this is. It's magic. You put in JSON, and you get an output. It's magic, and we don't mess with it. And that sucks. Don't do that. Don't do that. Test your systems. Test the machine learning systems. Work with your data scientists. Work with your ops people, because this is part of your attack surface. This is part of your attack surface, right? And don't let people hype you up over AI-generated phishing emails. When GPT-2 came out, people went nuts and were like, this is the end of detecting phishing. They're going to be too convincing. And, like, just don't. Okay? It's the same thing that we've been saying for 20 years. Patch your systems. Look for bad stuff, and patch your systems. That's it. There's nothing new under the sun. It's the exact same thing we've been saying for 20 years. Just keep looking for bad stuff, and patch your systems. And if anybody gives you pushback on patching your systems, like, have them talk to me, and I'll yell at them for you. Because you have to patch your systems. If you're watching this talk and you're a data scientist or a machine learning engineer, first of all, thank you for coming to DEF CON. DEF CON is dope. And, like, even though it's free and remote this year, come back next year, because it's cool. Conduct threat modeling on your models. So, before you put a model into deployment, work with your infosec team, work with ops, and sort of ask, like, what could go wrong? Right? What are the risks of an adversarial attack on this? What are the risks associated with this particular model? And what is the attack surface of this model? You know, and also when you're deploying those models, work with infosec and ops to balance uptime with the risk of model theft. These two kind of go hand in hand, is, like, you're going to deploy these models. Somebody may try to steal them. Is there a risk if somebody steals your model? Like, does it matter? And if it doesn't matter, like, think about would it be okay to publish the source code and data for this model? And if your answer to that is no, then you really need to make sure that infosec and operations are aware at deployment time. And finally, don't hype people up over text generation models. Like, GPT-2 and GPT-3 are super cool from an NLP standpoint, natural language processing, but they're not going to change the security landscape. Stop scaring people, please. I'm literally begging you, don't hype people up over text generation models. Here are some references. These are some of the things I talked about. And thank you so much for attending my talk. I'll be in the Discord most of the weekend, so if you have any questions, feel free to drop me a line. Thank you so much. Have a great weekend. Enjoy your con."
}