{
    "segments": [
        {
            "id": 1,
            "seek": 3000,
            "start": 1.47,
            "end": 10.05,
            "text": " Hello everyone and welcome to our next talk which is mine. To introduce myself, which is wrong,",
            "tokens": [
                50396,
                2425,
                1518,
                293,
                2928,
                281,
                527,
                958,
                751,
                597,
                307,
                3892,
                13,
                1407,
                5366,
                2059,
                11,
                597,
                307,
                2085,
                11,
                50776
            ],
            "temperature": 0.0,
            "avg_logprob": -0.2786165596394057,
            "compression_ratio": 1.5341880341880343,
            "no_speech_prob": 0.09661865234375,
            "words": [
                {
                    "start": 1.47,
                    "end": 1.81,
                    "word": " Hello",
                    "probability": 0.4892578125
                },
                {
                    "start": 1.81,
                    "end": 2.19,
                    "word": " everyone",
                    "probability": 0.74072265625
                },
                {
                    "start": 2.19,
                    "end": 2.69,
                    "word": " and",
                    "probability": 0.60400390625
                },
                {
                    "start": 2.69,
                    "end": 3.01,
                    "word": " welcome",
                    "probability": 0.76611328125
                },
                {
                    "start": 3.01,
                    "end": 3.21,
                    "word": " to",
                    "probability": 0.90576171875
                },
                {
                    "start": 3.21,
                    "end": 3.39,
                    "word": " our",
                    "probability": 0.81982421875
                },
                {
                    "start": 3.39,
                    "end": 3.55,
                    "word": " next",
                    "probability": 0.8466796875
                },
                {
                    "start": 3.55,
                    "end": 3.93,
                    "word": " talk",
                    "probability": 0.80810546875
                },
                {
                    "start": 3.93,
                    "end": 4.33,
                    "word": " which",
                    "probability": 0.552734375
                },
                {
                    "start": 4.33,
                    "end": 4.63,
                    "word": " is",
                    "probability": 0.90478515625
                },
                {
                    "start": 4.63,
                    "end": 5.71,
                    "word": " mine",
                    "probability": 0.68359375
                },
                {
                    "start": 5.71,
                    "end": 6.27,
                    "word": ".",
                    "probability": 0.67041015625
                },
                {
                    "start": 6.51,
                    "end": 7.05,
                    "word": " To",
                    "probability": 0.87744140625
                },
                {
                    "start": 7.05,
                    "end": 7.53,
                    "word": " introduce",
                    "probability": 0.6982421875
                },
                {
                    "start": 7.53,
                    "end": 8.31,
                    "word": " myself",
                    "probability": 0.69775390625
                },
                {
                    "start": 8.31,
                    "end": 8.43,
                    "word": ",",
                    "probability": 0.434326171875
                },
                {
                    "start": 8.43,
                    "end": 8.63,
                    "word": " which",
                    "probability": 0.89404296875
                },
                {
                    "start": 8.63,
                    "end": 8.77,
                    "word": " is",
                    "probability": 0.89990234375
                },
                {
                    "start": 8.77,
                    "end": 9.17,
                    "word": " wrong",
                    "probability": 0.7890625
                },
                {
                    "start": 9.17,
                    "end": 10.05,
                    "word": ",",
                    "probability": 0.779296875
                }
            ]
        },
        {
            "id": 2,
            "seek": 3000,
            "start": 10.05,
            "end": 16.49,
            "text": " I did a postdoc in machine learning after getting my PhD in algebraic topology with some of the",
            "tokens": [
                50828,
                286,
                630,
                257,
                2183,
                39966,
                294,
                3479,
                2539,
                934,
                1242,
                452,
                14476,
                294,
                21989,
                299,
                1192,
                1793,
                365,
                512,
                295,
                264,
                51140
            ],
            "temperature": 0.0,
            "avg_logprob": -0.2786165596394057,
            "compression_ratio": 1.5341880341880343,
            "no_speech_prob": 0.09661865234375,
            "words": [
                {
                    "start": 10.05,
                    "end": 10.49,
                    "word": " I",
                    "probability": 0.97802734375
                },
                {
                    "start": 10.49,
                    "end": 11.05,
                    "word": " did",
                    "probability": 0.89501953125
                },
                {
                    "start": 11.25,
                    "end": 12.03,
                    "word": " a postdoc",
                    "probability": 0.6600341796875
                },
                {
                    "start": 12.03,
                    "end": 12.15,
                    "word": " in",
                    "probability": 0.87744140625
                },
                {
                    "start": 12.15,
                    "end": 12.35,
                    "word": " machine",
                    "probability": 0.55126953125
                },
                {
                    "start": 12.35,
                    "end": 12.61,
                    "word": " learning",
                    "probability": 0.84619140625
                },
                {
                    "start": 12.61,
                    "end": 12.89,
                    "word": " after",
                    "probability": 0.71875
                },
                {
                    "start": 12.89,
                    "end": 13.11,
                    "word": " getting",
                    "probability": 0.89990234375
                },
                {
                    "start": 13.11,
                    "end": 13.31,
                    "word": " my",
                    "probability": 0.89990234375
                },
                {
                    "start": 13.31,
                    "end": 13.59,
                    "word": " PhD",
                    "probability": 0.68701171875
                },
                {
                    "start": 13.59,
                    "end": 13.87,
                    "word": " in",
                    "probability": 0.89501953125
                },
                {
                    "start": 13.87,
                    "end": 14.31,
                    "word": " algebraic",
                    "probability": 0.832275390625
                },
                {
                    "start": 14.31,
                    "end": 14.69,
                    "word": " topology",
                    "probability": 0.90185546875
                },
                {
                    "start": 14.69,
                    "end": 15.67,
                    "word": " with",
                    "probability": 0.748046875
                },
                {
                    "start": 15.67,
                    "end": 16.31,
                    "word": " some",
                    "probability": 0.8115234375
                },
                {
                    "start": 16.31,
                    "end": 16.41,
                    "word": " of",
                    "probability": 0.8955078125
                },
                {
                    "start": 16.41,
                    "end": 16.49,
                    "word": " the",
                    "probability": 0.81396484375
                }
            ]
        },
        {
            "id": 3,
            "seek": 3000,
            "start": 16.49,
            "end": 22.45,
            "text": " people who did some of the math behind the concepts in this talk and I've just extended",
            "tokens": [
                51140,
                561,
                567,
                630,
                512,
                295,
                264,
                5221,
                2261,
                264,
                10392,
                294,
                341,
                751,
                293,
                286,
                600,
                445,
                10913,
                51444
            ],
            "temperature": 0.0,
            "avg_logprob": -0.2786165596394057,
            "compression_ratio": 1.5341880341880343,
            "no_speech_prob": 0.09661865234375,
            "words": [
                {
                    "start": 16.49,
                    "end": 16.71,
                    "word": " people",
                    "probability": 0.91845703125
                },
                {
                    "start": 16.71,
                    "end": 17.01,
                    "word": " who",
                    "probability": 0.8095703125
                },
                {
                    "start": 17.01,
                    "end": 17.29,
                    "word": " did",
                    "probability": 0.88525390625
                },
                {
                    "start": 17.29,
                    "end": 17.59,
                    "word": " some",
                    "probability": 0.80126953125
                },
                {
                    "start": 17.59,
                    "end": 17.67,
                    "word": " of",
                    "probability": 0.9033203125
                },
                {
                    "start": 17.67,
                    "end": 17.73,
                    "word": " the",
                    "probability": 0.818359375
                },
                {
                    "start": 17.73,
                    "end": 17.91,
                    "word": " math",
                    "probability": 0.8798828125
                },
                {
                    "start": 17.91,
                    "end": 18.47,
                    "word": " behind",
                    "probability": 0.77783203125
                },
                {
                    "start": 18.47,
                    "end": 19.49,
                    "word": " the",
                    "probability": 0.73681640625
                },
                {
                    "start": 19.49,
                    "end": 20.05,
                    "word": " concepts",
                    "probability": 0.43408203125
                },
                {
                    "start": 20.05,
                    "end": 20.33,
                    "word": " in",
                    "probability": 0.90478515625
                },
                {
                    "start": 20.33,
                    "end": 20.49,
                    "word": " this",
                    "probability": 0.88330078125
                },
                {
                    "start": 20.49,
                    "end": 20.83,
                    "word": " talk",
                    "probability": 0.8037109375
                },
                {
                    "start": 20.83,
                    "end": 21.69,
                    "word": " and",
                    "probability": 0.59521484375
                },
                {
                    "start": 21.69,
                    "end": 21.93,
                    "word": " I've",
                    "probability": 0.944580078125
                },
                {
                    "start": 21.93,
                    "end": 22.09,
                    "word": " just",
                    "probability": 0.88818359375
                },
                {
                    "start": 22.09,
                    "end": 22.45,
                    "word": " extended",
                    "probability": 0.74560546875
                }
            ]
        },
        {
            "id": 4,
            "seek": 3000,
            "start": 22.45,
            "end": 30.81,
            "text": " it for this thing. So that's enough about me. Let me play my talk back for you.",
            "tokens": [
                51444,
                309,
                337,
                341,
                551,
                13,
                407,
                300,
                311,
                1547,
                466,
                385,
                13,
                961,
                385,
                862,
                452,
                751,
                646,
                337,
                291,
                13,
                51852
            ],
            "temperature": 0.0,
            "avg_logprob": -0.2786165596394057,
            "compression_ratio": 1.5341880341880343,
            "no_speech_prob": 0.09661865234375,
            "words": [
                {
                    "start": 22.45,
                    "end": 22.69,
                    "word": " it",
                    "probability": 0.9033203125
                },
                {
                    "start": 22.69,
                    "end": 22.85,
                    "word": " for",
                    "probability": 0.68310546875
                },
                {
                    "start": 22.85,
                    "end": 23.09,
                    "word": " this",
                    "probability": 0.87060546875
                },
                {
                    "start": 23.09,
                    "end": 23.35,
                    "word": " thing",
                    "probability": 0.80517578125
                },
                {
                    "start": 23.35,
                    "end": 23.67,
                    "word": ".",
                    "probability": 0.456787109375
                },
                {
                    "start": 23.67,
                    "end": 23.79,
                    "word": " So",
                    "probability": 0.787109375
                },
                {
                    "start": 23.79,
                    "end": 26.61,
                    "word": " that's",
                    "probability": 0.80859375
                },
                {
                    "start": 26.61,
                    "end": 26.75,
                    "word": " enough",
                    "probability": 0.767578125
                },
                {
                    "start": 26.75,
                    "end": 26.97,
                    "word": " about",
                    "probability": 0.82080078125
                },
                {
                    "start": 26.97,
                    "end": 27.19,
                    "word": " me",
                    "probability": 0.904296875
                },
                {
                    "start": 27.19,
                    "end": 27.75,
                    "word": ".",
                    "probability": 0.65869140625
                },
                {
                    "start": 28.07,
                    "end": 28.63,
                    "word": " Let",
                    "probability": 0.8896484375
                },
                {
                    "start": 28.63,
                    "end": 28.91,
                    "word": " me",
                    "probability": 0.904296875
                },
                {
                    "start": 28.91,
                    "end": 29.67,
                    "word": " play",
                    "probability": 0.87744140625
                },
                {
                    "start": 29.67,
                    "end": 29.89,
                    "word": " my",
                    "probability": 0.9013671875
                },
                {
                    "start": 29.89,
                    "end": 30.09,
                    "word": " talk",
                    "probability": 0.82666015625
                },
                {
                    "start": 30.09,
                    "end": 30.39,
                    "word": " back",
                    "probability": 0.80029296875
                },
                {
                    "start": 30.39,
                    "end": 30.57,
                    "word": " for",
                    "probability": 0.89501953125
                },
                {
                    "start": 30.57,
                    "end": 30.73,
                    "word": " you",
                    "probability": 0.89599609375
                },
                {
                    "start": 30.73,
                    "end": 30.81,
                    "word": ".",
                    "probability": 0.95263671875
                }
            ]
        },
        {
            "id": 5,
            "seek": 6000,
            "start": 32.49,
            "end": 34.53,
            "text": " So, here we go.",
            "tokens": [
                50364,
                407,
                11,
                510,
                321,
                352,
                13,
                50540
            ],
            "temperature": 0.0,
            "avg_logprob": -0.45617815749398594,
            "compression_ratio": 1.4405940594059405,
            "no_speech_prob": 0.201416015625,
            "words": [
                {
                    "start": 32.49,
                    "end": 33.01,
                    "word": " So",
                    "probability": 0.1380615234375
                },
                {
                    "start": 33.01,
                    "end": 33.35,
                    "word": ",",
                    "probability": 0.39892578125
                },
                {
                    "start": 33.35,
                    "end": 33.79,
                    "word": " here",
                    "probability": 0.75390625
                },
                {
                    "start": 33.79,
                    "end": 34.05,
                    "word": " we",
                    "probability": 0.2445068359375
                },
                {
                    "start": 34.05,
                    "end": 34.27,
                    "word": " go",
                    "probability": 0.8291015625
                },
                {
                    "start": 34.27,
                    "end": 34.53,
                    "word": ".",
                    "probability": 0.728515625
                }
            ]
        },
        {
            "id": 6,
            "seek": 6000,
            "start": 34.53,
            "end": 39.83,
            "text": " I welcome everyone to my talk on calculating concepts very quickly with this library called",
            "tokens": [
                50540,
                286,
                2928,
                1518,
                281,
                452,
                751,
                322,
                28258,
                416,
                1336,
                82,
                588,
                2661,
                365,
                341,
                6405,
                1219,
                50808
            ],
            "temperature": 0.0,
            "avg_logprob": -0.45617815749398594,
            "compression_ratio": 1.4405940594059405,
            "no_speech_prob": 0.201416015625,
            "words": [
                {
                    "start": 34.53,
                    "end": 34.63,
                    "word": " I",
                    "probability": 0.416015625
                },
                {
                    "start": 34.63,
                    "end": 34.89,
                    "word": " welcome",
                    "probability": 0.0347900390625
                },
                {
                    "start": 34.89,
                    "end": 35.73,
                    "word": " everyone",
                    "probability": 0.64404296875
                },
                {
                    "start": 35.73,
                    "end": 35.95,
                    "word": " to",
                    "probability": 0.8857421875
                },
                {
                    "start": 35.95,
                    "end": 36.09,
                    "word": " my",
                    "probability": 0.87548828125
                },
                {
                    "start": 36.09,
                    "end": 36.33,
                    "word": " talk",
                    "probability": 0.80615234375
                },
                {
                    "start": 36.33,
                    "end": 36.79,
                    "word": " on",
                    "probability": 0.810546875
                },
                {
                    "start": 36.79,
                    "end": 37.31,
                    "word": " calculating",
                    "probability": 0.7939453125
                },
                {
                    "start": 37.31,
                    "end": 37.79,
                    "word": " concepts",
                    "probability": 0.3097279866536458
                },
                {
                    "start": 37.79,
                    "end": 38.25,
                    "word": " very",
                    "probability": 0.56103515625
                },
                {
                    "start": 38.25,
                    "end": 38.65,
                    "word": " quickly",
                    "probability": 0.873046875
                },
                {
                    "start": 38.65,
                    "end": 39.07,
                    "word": " with",
                    "probability": 0.7578125
                },
                {
                    "start": 39.07,
                    "end": 39.25,
                    "word": " this",
                    "probability": 0.8505859375
                },
                {
                    "start": 39.25,
                    "end": 39.55,
                    "word": " library",
                    "probability": 0.84423828125
                },
                {
                    "start": 39.55,
                    "end": 39.83,
                    "word": " called",
                    "probability": 0.73046875
                }
            ]
        },
        {
            "id": 7,
            "seek": 6000,
            "start": 39.83,
            "end": 61.73,
            "text": " Google which is based on... I hear no audio.",
            "tokens": [
                50808,
                3329,
                597,
                307,
                2361,
                322,
                485,
                286,
                1568,
                572,
                6278,
                13,
                51036
            ],
            "temperature": 0.0,
            "avg_logprob": -0.45617815749398594,
            "compression_ratio": 1.4405940594059405,
            "no_speech_prob": 0.201416015625,
            "words": [
                {
                    "start": 39.83,
                    "end": 40.07,
                    "word": " Google",
                    "probability": 0.4765625
                },
                {
                    "start": 40.07,
                    "end": 40.45,
                    "word": " which",
                    "probability": 0.46923828125
                },
                {
                    "start": 40.45,
                    "end": 40.61,
                    "word": " is",
                    "probability": 0.8994140625
                },
                {
                    "start": 40.61,
                    "end": 40.91,
                    "word": " based",
                    "probability": 0.8134765625
                },
                {
                    "start": 40.91,
                    "end": 41.47,
                    "word": " on...",
                    "probability": 0.46805572509765625
                },
                {
                    "start": 41.47,
                    "end": 42.61,
                    "word": " I",
                    "probability": 0.86962890625
                },
                {
                    "start": 59.53,
                    "end": 60.73,
                    "word": " hear",
                    "probability": 0.8671875
                },
                {
                    "start": 60.73,
                    "end": 60.93,
                    "word": " no",
                    "probability": 0.875
                },
                {
                    "start": 60.93,
                    "end": 61.21,
                    "word": " audio",
                    "probability": 0.90625
                },
                {
                    "start": 61.21,
                    "end": 61.73,
                    "word": ".",
                    "probability": 0.8173828125
                }
            ]
        },
        {
            "id": 8,
            "seek": 6000,
            "start": 65.79,
            "end": 66.93,
            "text": " Oopsie daisy.",
            "tokens": [
                51116,
                21726,
                414,
                1120,
                14169,
                13,
                51152
            ],
            "temperature": 0.0,
            "avg_logprob": -0.45617815749398594,
            "compression_ratio": 1.4405940594059405,
            "no_speech_prob": 0.201416015625,
            "words": [
                {
                    "start": 65.79,
                    "end": 66.21,
                    "word": " Oopsie",
                    "probability": 0.4788818359375
                },
                {
                    "start": 66.21,
                    "end": 66.41,
                    "word": " daisy",
                    "probability": 0.6431884765625
                },
                {
                    "start": 66.41,
                    "end": 66.93,
                    "word": ".",
                    "probability": 0.7861328125
                }
            ]
        },
        {
            "id": 9,
            "seek": 6000,
            "start": 88.69,
            "end": 91.43,
            "text": " Um, there's no... do you hear audio from the talk?",
            "tokens": [
                51256,
                3301,
                11,
                456,
                311,
                572,
                485,
                360,
                291,
                1568,
                6278,
                490,
                264,
                751,
                30,
                51360
            ],
            "temperature": 0.0,
            "avg_logprob": -0.45617815749398594,
            "compression_ratio": 1.4405940594059405,
            "no_speech_prob": 0.201416015625,
            "words": [
                {
                    "start": 88.69,
                    "end": 89.21,
                    "word": " Um",
                    "probability": 0.341064453125
                },
                {
                    "start": 89.21,
                    "end": 89.33,
                    "word": ",",
                    "probability": 0.7978515625
                },
                {
                    "start": 89.33,
                    "end": 89.49,
                    "word": " there's",
                    "probability": 0.817626953125
                },
                {
                    "start": 89.49,
                    "end": 89.73,
                    "word": " no...",
                    "probability": 0.5401611328125
                },
                {
                    "start": 89.73,
                    "end": 89.87,
                    "word": " do",
                    "probability": 0.291748046875
                },
                {
                    "start": 89.87,
                    "end": 90.01,
                    "word": " you",
                    "probability": 0.89111328125
                },
                {
                    "start": 90.01,
                    "end": 90.17,
                    "word": " hear",
                    "probability": 0.85986328125
                },
                {
                    "start": 90.17,
                    "end": 90.35,
                    "word": " audio",
                    "probability": 0.904296875
                },
                {
                    "start": 90.35,
                    "end": 90.57,
                    "word": " from",
                    "probability": 0.8271484375
                },
                {
                    "start": 90.57,
                    "end": 90.71,
                    "word": " the",
                    "probability": 0.7958984375
                },
                {
                    "start": 90.71,
                    "end": 90.91,
                    "word": " talk",
                    "probability": 0.74560546875
                },
                {
                    "start": 90.91,
                    "end": 91.43,
                    "word": "?",
                    "probability": 0.93896484375
                }
            ]
        },
        {
            "id": 10,
            "seek": 6000,
            "start": 98.9,
            "end": 101.26,
            "text": " Okay, well that's the problem I'm trying to...",
            "tokens": [
                51540,
                1033,
                11,
                731,
                300,
                311,
                264,
                1154,
                286,
                478,
                1382,
                281,
                485,
                51640
            ],
            "temperature": 0.0,
            "avg_logprob": -0.45617815749398594,
            "compression_ratio": 1.4405940594059405,
            "no_speech_prob": 0.201416015625,
            "words": [
                {
                    "start": 98.9,
                    "end": 99.38,
                    "word": " Okay",
                    "probability": 0.413330078125
                },
                {
                    "start": 99.38,
                    "end": 99.68,
                    "word": ",",
                    "probability": 0.69677734375
                },
                {
                    "start": 99.68,
                    "end": 99.84,
                    "word": " well",
                    "probability": 0.865234375
                },
                {
                    "start": 99.84,
                    "end": 100.12,
                    "word": " that's",
                    "probability": 0.8740234375
                },
                {
                    "start": 100.12,
                    "end": 100.36,
                    "word": " the",
                    "probability": 0.80322265625
                },
                {
                    "start": 100.36,
                    "end": 100.6,
                    "word": " problem",
                    "probability": 0.78076171875
                },
                {
                    "start": 100.6,
                    "end": 100.84,
                    "word": " I'm",
                    "probability": 0.857421875
                },
                {
                    "start": 100.84,
                    "end": 101.0,
                    "word": " trying",
                    "probability": 0.78125
                },
                {
                    "start": 101.0,
                    "end": 101.26,
                    "word": " to...",
                    "probability": 0.873291015625
                }
            ]
        },
        {
            "id": 11,
            "seek": 6000,
            "start": 122.03,
            "end": 124.05,
            "text": " Uh, sorry about this guys.",
            "tokens": [
                51732,
                4019,
                11,
                2597,
                466,
                341,
                1074,
                13,
                51816
            ],
            "temperature": 0.0,
            "avg_logprob": -0.45617815749398594,
            "compression_ratio": 1.4405940594059405,
            "no_speech_prob": 0.201416015625,
            "words": [
                {
                    "start": 122.03,
                    "end": 122.55,
                    "word": " Uh",
                    "probability": 0.31494140625
                },
                {
                    "start": 122.55,
                    "end": 122.71,
                    "word": ",",
                    "probability": 0.86669921875
                },
                {
                    "start": 122.71,
                    "end": 123.07,
                    "word": " sorry",
                    "probability": 0.76611328125
                },
                {
                    "start": 123.07,
                    "end": 123.31,
                    "word": " about",
                    "probability": 0.83056640625
                },
                {
                    "start": 123.31,
                    "end": 123.53,
                    "word": " this",
                    "probability": 0.845703125
                },
                {
                    "start": 123.53,
                    "end": 123.77,
                    "word": " guys",
                    "probability": 0.69287109375
                },
                {
                    "start": 123.77,
                    "end": 124.05,
                    "word": ".",
                    "probability": 0.95361328125
                }
            ]
        },
        {
            "id": 12,
            "seek": 8422,
            "start": 142.81,
            "end": 147.95,
            "text": " Uh, sorry, I don't know what's going on.",
            "tokens": [
                50412,
                4019,
                11,
                2597,
                11,
                286,
                500,
                380,
                458,
                437,
                311,
                516,
                322,
                13,
                50632
            ],
            "temperature": 0.0,
            "avg_logprob": -0.2715163836713697,
            "compression_ratio": 1.4339622641509433,
            "no_speech_prob": 0.0036487579345703125,
            "words": [
                {
                    "start": 142.81,
                    "end": 143.41,
                    "word": " Uh",
                    "probability": 0.053802490234375
                },
                {
                    "start": 143.41,
                    "end": 143.77,
                    "word": ",",
                    "probability": 0.383544921875
                },
                {
                    "start": 143.77,
                    "end": 146.27,
                    "word": " sorry",
                    "probability": 0.53857421875
                },
                {
                    "start": 146.27,
                    "end": 146.53,
                    "word": ",",
                    "probability": 0.69482421875
                },
                {
                    "start": 146.53,
                    "end": 146.83,
                    "word": " I",
                    "probability": 0.99365234375
                },
                {
                    "start": 146.83,
                    "end": 146.95,
                    "word": " don't",
                    "probability": 0.943359375
                },
                {
                    "start": 146.95,
                    "end": 147.07,
                    "word": " know",
                    "probability": 0.8154296875
                },
                {
                    "start": 147.07,
                    "end": 147.25,
                    "word": " what's",
                    "probability": 0.940185546875
                },
                {
                    "start": 147.25,
                    "end": 147.43,
                    "word": " going",
                    "probability": 0.84716796875
                },
                {
                    "start": 147.43,
                    "end": 147.71,
                    "word": " on",
                    "probability": 0.9033203125
                },
                {
                    "start": 147.71,
                    "end": 147.95,
                    "word": ".",
                    "probability": 0.89306640625
                }
            ]
        },
        {
            "id": 13,
            "seek": 8422,
            "start": 198.25,
            "end": 203.59,
            "text": " Um, I am... I'll apologize.",
            "tokens": [
                50748,
                3301,
                11,
                286,
                669,
                485,
                286,
                603,
                12328,
                13,
                50988
            ],
            "temperature": 0.0,
            "avg_logprob": -0.2715163836713697,
            "compression_ratio": 1.4339622641509433,
            "no_speech_prob": 0.0036487579345703125,
            "words": [
                {
                    "start": 198.25,
                    "end": 198.85,
                    "word": " Um",
                    "probability": 0.273681640625
                },
                {
                    "start": 198.85,
                    "end": 199.39,
                    "word": ",",
                    "probability": 0.81494140625
                },
                {
                    "start": 199.39,
                    "end": 201.79,
                    "word": " I",
                    "probability": 0.953125
                },
                {
                    "start": 201.79,
                    "end": 202.33,
                    "word": " am...",
                    "probability": 0.4393119812011719
                },
                {
                    "start": 202.33,
                    "end": 202.55,
                    "word": " I'll",
                    "probability": 0.8076171875
                },
                {
                    "start": 202.55,
                    "end": 202.99,
                    "word": " apologize",
                    "probability": 0.7744140625
                },
                {
                    "start": 202.99,
                    "end": 203.59,
                    "word": ".",
                    "probability": 0.8466796875
                }
            ]
        },
        {
            "id": 14,
            "seek": 8422,
            "start": 214.22,
            "end": 219.14,
            "text": " Key takeaway for this talk is this is a novel way of calculating drift.",
            "tokens": [
                51080,
                12759,
                30681,
                337,
                341,
                751,
                307,
                341,
                307,
                257,
                7613,
                636,
                295,
                28258,
                19699,
                13,
                51280
            ],
            "temperature": 0.0,
            "avg_logprob": -0.2715163836713697,
            "compression_ratio": 1.4339622641509433,
            "no_speech_prob": 0.0036487579345703125,
            "words": [
                {
                    "start": 214.22,
                    "end": 214.82,
                    "word": " Key",
                    "probability": 0.480224609375
                },
                {
                    "start": 214.82,
                    "end": 215.12,
                    "word": " takeaway",
                    "probability": 0.63818359375
                },
                {
                    "start": 215.12,
                    "end": 215.28,
                    "word": " for",
                    "probability": 0.59765625
                },
                {
                    "start": 215.28,
                    "end": 215.4,
                    "word": " this",
                    "probability": 0.88134765625
                },
                {
                    "start": 215.4,
                    "end": 215.6,
                    "word": " talk",
                    "probability": 0.8134765625
                },
                {
                    "start": 215.6,
                    "end": 215.8,
                    "word": " is",
                    "probability": 0.87646484375
                },
                {
                    "start": 215.8,
                    "end": 216.0,
                    "word": " this",
                    "probability": 0.81591796875
                },
                {
                    "start": 216.0,
                    "end": 216.16,
                    "word": " is",
                    "probability": 0.89501953125
                },
                {
                    "start": 216.38,
                    "end": 216.76,
                    "word": " a novel",
                    "probability": 0.84521484375
                },
                {
                    "start": 216.76,
                    "end": 217.68,
                    "word": " way",
                    "probability": 0.6728515625
                },
                {
                    "start": 217.68,
                    "end": 217.82,
                    "word": " of",
                    "probability": 0.908203125
                },
                {
                    "start": 217.82,
                    "end": 218.2,
                    "word": " calculating",
                    "probability": 0.8876953125
                },
                {
                    "start": 218.2,
                    "end": 218.54,
                    "word": " drift",
                    "probability": 0.8662109375
                },
                {
                    "start": 218.54,
                    "end": 219.14,
                    "word": ".",
                    "probability": 0.9375
                }
            ]
        },
        {
            "id": 15,
            "seek": 8422,
            "start": 219.98,
            "end": 224.44,
            "text": " Welcome everyone to my talk on calculating concept drift very quickly with this library",
            "tokens": [
                51364,
                4027,
                1518,
                281,
                452,
                751,
                322,
                28258,
                3410,
                19699,
                588,
                2661,
                365,
                341,
                6405,
                51572
            ],
            "temperature": 0.0,
            "avg_logprob": -0.2715163836713697,
            "compression_ratio": 1.4339622641509433,
            "no_speech_prob": 0.0036487579345703125,
            "words": [
                {
                    "start": 219.98,
                    "end": 220.58,
                    "word": " Welcome",
                    "probability": 0.444580078125
                },
                {
                    "start": 220.58,
                    "end": 220.9,
                    "word": " everyone",
                    "probability": 0.85400390625
                },
                {
                    "start": 220.9,
                    "end": 221.08,
                    "word": " to",
                    "probability": 0.8662109375
                },
                {
                    "start": 221.08,
                    "end": 221.24,
                    "word": " my",
                    "probability": 0.90625
                },
                {
                    "start": 221.24,
                    "end": 221.52,
                    "word": " talk",
                    "probability": 0.81201171875
                },
                {
                    "start": 221.52,
                    "end": 221.86,
                    "word": " on",
                    "probability": 0.89599609375
                },
                {
                    "start": 221.86,
                    "end": 222.3,
                    "word": " calculating",
                    "probability": 0.81396484375
                },
                {
                    "start": 222.3,
                    "end": 222.72,
                    "word": " concept",
                    "probability": 0.77978515625
                },
                {
                    "start": 222.72,
                    "end": 222.92,
                    "word": " drift",
                    "probability": 0.75830078125
                },
                {
                    "start": 222.92,
                    "end": 223.18,
                    "word": " very",
                    "probability": 0.64404296875
                },
                {
                    "start": 223.18,
                    "end": 223.5,
                    "word": " quickly",
                    "probability": 0.8798828125
                },
                {
                    "start": 223.5,
                    "end": 223.94,
                    "word": " with",
                    "probability": 0.7724609375
                },
                {
                    "start": 223.94,
                    "end": 224.16,
                    "word": " this",
                    "probability": 0.87841796875
                },
                {
                    "start": 224.16,
                    "end": 224.44,
                    "word": " library",
                    "probability": 0.81787109375
                }
            ]
        },
        {
            "id": 16,
            "seek": 11338,
            "start": 224.44,
            "end": 229.4,
            "text": " called Google which is based on cover trees. My name is Sven Kittel. I am a senior data",
            "tokens": [
                50364,
                1219,
                3329,
                597,
                307,
                2361,
                322,
                2060,
                5852,
                13,
                1222,
                1315,
                307,
                49787,
                591,
                47265,
                13,
                286,
                669,
                257,
                7965,
                1412,
                50612
            ],
            "temperature": 0.0,
            "avg_logprob": -0.29942907794163776,
            "compression_ratio": 1.5819672131147542,
            "no_speech_prob": 0.28125,
            "words": [
                {
                    "start": 224.44,
                    "end": 224.66,
                    "word": " called",
                    "probability": 0.1273193359375
                },
                {
                    "start": 224.66,
                    "end": 224.94,
                    "word": " Google",
                    "probability": 0.0026531219482421875
                },
                {
                    "start": 224.94,
                    "end": 225.34,
                    "word": " which",
                    "probability": 0.29052734375
                },
                {
                    "start": 225.34,
                    "end": 225.5,
                    "word": " is",
                    "probability": 0.9072265625
                },
                {
                    "start": 225.5,
                    "end": 225.82,
                    "word": " based",
                    "probability": 0.8154296875
                },
                {
                    "start": 225.82,
                    "end": 226.16,
                    "word": " on",
                    "probability": 0.9111328125
                },
                {
                    "start": 226.16,
                    "end": 226.36,
                    "word": " cover",
                    "probability": 0.6015625
                },
                {
                    "start": 226.36,
                    "end": 226.58,
                    "word": " trees",
                    "probability": 0.72607421875
                },
                {
                    "start": 226.58,
                    "end": 227.16,
                    "word": ".",
                    "probability": 0.9189453125
                },
                {
                    "start": 227.26,
                    "end": 227.64,
                    "word": " My",
                    "probability": 0.8193359375
                },
                {
                    "start": 227.64,
                    "end": 227.78,
                    "word": " name",
                    "probability": 0.82177734375
                },
                {
                    "start": 227.78,
                    "end": 227.88,
                    "word": " is",
                    "probability": 0.79833984375
                },
                {
                    "start": 227.88,
                    "end": 228.08,
                    "word": " Sven",
                    "probability": 0.919921875
                },
                {
                    "start": 228.08,
                    "end": 228.3,
                    "word": " Kittel",
                    "probability": 0.32452392578125
                },
                {
                    "start": 228.3,
                    "end": 228.56,
                    "word": ".",
                    "probability": 0.463623046875
                },
                {
                    "start": 228.56,
                    "end": 228.64,
                    "word": " I",
                    "probability": 0.9892578125
                },
                {
                    "start": 228.64,
                    "end": 228.86,
                    "word": " am",
                    "probability": 0.87939453125
                },
                {
                    "start": 228.98,
                    "end": 229.14,
                    "word": " a senior",
                    "probability": 0.68701171875
                },
                {
                    "start": 229.14,
                    "end": 229.4,
                    "word": " data",
                    "probability": 0.8310546875
                }
            ]
        },
        {
            "id": 17,
            "seek": 11338,
            "start": 229.4,
            "end": 233.94,
            "text": " scientist at Elastic and my Twitter handle is co-mathematician.",
            "tokens": [
                50612,
                12662,
                412,
                2699,
                2750,
                293,
                452,
                5794,
                4813,
                307,
                598,
                12,
                24761,
                14911,
                952,
                13,
                50808
            ],
            "temperature": 0.0,
            "avg_logprob": -0.29942907794163776,
            "compression_ratio": 1.5819672131147542,
            "no_speech_prob": 0.28125,
            "words": [
                {
                    "start": 229.4,
                    "end": 229.8,
                    "word": " scientist",
                    "probability": 0.7431640625
                },
                {
                    "start": 229.8,
                    "end": 230.18,
                    "word": " at",
                    "probability": 0.8916015625
                },
                {
                    "start": 230.18,
                    "end": 230.6,
                    "word": " Elastic",
                    "probability": 0.864013671875
                },
                {
                    "start": 230.6,
                    "end": 231.64,
                    "word": " and",
                    "probability": 0.6083984375
                },
                {
                    "start": 231.64,
                    "end": 231.76,
                    "word": " my",
                    "probability": 0.8837890625
                },
                {
                    "start": 231.76,
                    "end": 231.92,
                    "word": " Twitter",
                    "probability": 0.662109375
                },
                {
                    "start": 231.92,
                    "end": 232.22,
                    "word": " handle",
                    "probability": 0.85400390625
                },
                {
                    "start": 232.22,
                    "end": 232.54,
                    "word": " is",
                    "probability": 0.82666015625
                },
                {
                    "start": 232.54,
                    "end": 232.72,
                    "word": " co",
                    "probability": 0.1689453125
                },
                {
                    "start": 232.72,
                    "end": 233.36,
                    "word": "-mathematician",
                    "probability": 0.8057861328125
                },
                {
                    "start": 233.36,
                    "end": 233.94,
                    "word": ".",
                    "probability": 0.927734375
                }
            ]
        },
        {
            "id": 18,
            "seek": 11338,
            "start": 234.48,
            "end": 241.66,
            "text": " So, the key outline... key takeaway for this talk is this is a novel way of calculating drift.",
            "tokens": [
                50864,
                407,
                11,
                264,
                2141,
                277,
                325,
                1889,
                485,
                2141,
                30681,
                337,
                341,
                751,
                307,
                341,
                307,
                257,
                7613,
                636,
                295,
                28258,
                19699,
                13,
                51228
            ],
            "temperature": 0.0,
            "avg_logprob": -0.29942907794163776,
            "compression_ratio": 1.5819672131147542,
            "no_speech_prob": 0.28125,
            "words": [
                {
                    "start": 234.48,
                    "end": 234.8,
                    "word": " So",
                    "probability": 0.74462890625
                },
                {
                    "start": 234.8,
                    "end": 235.78,
                    "word": ",",
                    "probability": 0.2330322265625
                },
                {
                    "start": 235.78,
                    "end": 236.26,
                    "word": " the",
                    "probability": 0.75830078125
                },
                {
                    "start": 236.26,
                    "end": 236.58,
                    "word": " key",
                    "probability": 0.76806640625
                },
                {
                    "start": 236.58,
                    "end": 237.5,
                    "word": " outline...",
                    "probability": 0.49997711181640625
                },
                {
                    "start": 237.5,
                    "end": 237.74,
                    "word": " key",
                    "probability": 0.51806640625
                },
                {
                    "start": 237.74,
                    "end": 238.02,
                    "word": " takeaway",
                    "probability": 0.71875
                },
                {
                    "start": 238.02,
                    "end": 238.18,
                    "word": " for",
                    "probability": 0.71630859375
                },
                {
                    "start": 238.18,
                    "end": 238.32,
                    "word": " this",
                    "probability": 0.88623046875
                },
                {
                    "start": 238.32,
                    "end": 238.5,
                    "word": " talk",
                    "probability": 0.81494140625
                },
                {
                    "start": 238.5,
                    "end": 238.7,
                    "word": " is",
                    "probability": 0.8828125
                },
                {
                    "start": 238.7,
                    "end": 238.9,
                    "word": " this",
                    "probability": 0.81640625
                },
                {
                    "start": 238.9,
                    "end": 239.08,
                    "word": " is",
                    "probability": 0.9033203125
                },
                {
                    "start": 239.3,
                    "end": 239.66,
                    "word": " a novel",
                    "probability": 0.86279296875
                },
                {
                    "start": 239.66,
                    "end": 240.6,
                    "word": " way",
                    "probability": 0.87646484375
                },
                {
                    "start": 240.6,
                    "end": 240.72,
                    "word": " of",
                    "probability": 0.90966796875
                },
                {
                    "start": 240.72,
                    "end": 241.08,
                    "word": " calculating",
                    "probability": 0.8779296875
                },
                {
                    "start": 241.08,
                    "end": 241.5,
                    "word": " drift",
                    "probability": 0.85595703125
                },
                {
                    "start": 241.5,
                    "end": 241.66,
                    "word": ".",
                    "probability": 0.88525390625
                }
            ]
        },
        {
            "id": 19,
            "seek": 11338,
            "start": 241.66,
            "end": 245.97,
            "text": " It's done using a new technique using cover trees.",
            "tokens": [
                51228,
                467,
                311,
                1096,
                1228,
                257,
                777,
                6532,
                1228,
                2060,
                5852,
                13,
                51428
            ],
            "temperature": 0.0,
            "avg_logprob": -0.29942907794163776,
            "compression_ratio": 1.5819672131147542,
            "no_speech_prob": 0.28125,
            "words": [
                {
                    "start": 241.66,
                    "end": 242.0,
                    "word": " It's",
                    "probability": 0.94091796875
                },
                {
                    "start": 242.0,
                    "end": 242.34,
                    "word": " done",
                    "probability": 0.8076171875
                },
                {
                    "start": 242.34,
                    "end": 243.24,
                    "word": " using",
                    "probability": 0.81640625
                },
                {
                    "start": 243.8,
                    "end": 244.62,
                    "word": " a new",
                    "probability": 0.8125
                },
                {
                    "start": 244.62,
                    "end": 244.92,
                    "word": " technique",
                    "probability": 0.888671875
                },
                {
                    "start": 244.92,
                    "end": 245.18,
                    "word": " using",
                    "probability": 0.728515625
                },
                {
                    "start": 245.18,
                    "end": 245.4,
                    "word": " cover",
                    "probability": 0.85302734375
                },
                {
                    "start": 245.4,
                    "end": 245.68,
                    "word": " trees",
                    "probability": 0.81591796875
                },
                {
                    "start": 245.68,
                    "end": 245.97,
                    "word": ".",
                    "probability": 0.33837890625
                }
            ]
        },
        {
            "id": 20,
            "seek": 11338,
            "start": 246.8,
            "end": 253.6,
            "text": " And because it's so fast, it's log-in time to actually calculate drift for a new sample.",
            "tokens": [
                51504,
                400,
                570,
                309,
                311,
                370,
                2370,
                11,
                309,
                311,
                3565,
                12,
                259,
                565,
                281,
                767,
                8873,
                19699,
                337,
                257,
                777,
                6889,
                13,
                51800
            ],
            "temperature": 0.0,
            "avg_logprob": -0.29942907794163776,
            "compression_ratio": 1.5819672131147542,
            "no_speech_prob": 0.28125,
            "words": [
                {
                    "start": 246.8,
                    "end": 246.86,
                    "word": " And",
                    "probability": 0.833984375
                },
                {
                    "start": 246.86,
                    "end": 247.76,
                    "word": " because",
                    "probability": 0.71240234375
                },
                {
                    "start": 247.76,
                    "end": 248.12,
                    "word": " it's",
                    "probability": 0.947998046875
                },
                {
                    "start": 248.12,
                    "end": 248.42,
                    "word": " so",
                    "probability": 0.90625
                },
                {
                    "start": 248.42,
                    "end": 248.92,
                    "word": " fast",
                    "probability": 0.80615234375
                },
                {
                    "start": 248.92,
                    "end": 249.1,
                    "word": ",",
                    "probability": 0.7158203125
                },
                {
                    "start": 249.1,
                    "end": 249.46,
                    "word": " it's",
                    "probability": 0.932373046875
                },
                {
                    "start": 249.46,
                    "end": 249.94,
                    "word": " log",
                    "probability": 0.5439453125
                },
                {
                    "start": 249.94,
                    "end": 250.18,
                    "word": "-in",
                    "probability": 0.6622314453125
                },
                {
                    "start": 250.18,
                    "end": 251.0,
                    "word": " time",
                    "probability": 0.81982421875
                },
                {
                    "start": 251.0,
                    "end": 251.2,
                    "word": " to",
                    "probability": 0.90283203125
                },
                {
                    "start": 251.2,
                    "end": 251.44,
                    "word": " actually",
                    "probability": 0.78857421875
                },
                {
                    "start": 251.44,
                    "end": 251.86,
                    "word": " calculate",
                    "probability": 0.8408203125
                },
                {
                    "start": 251.86,
                    "end": 252.38,
                    "word": " drift",
                    "probability": 0.71923828125
                },
                {
                    "start": 252.38,
                    "end": 252.6,
                    "word": " for",
                    "probability": 0.90234375
                },
                {
                    "start": 252.68,
                    "end": 252.8,
                    "word": " a new",
                    "probability": 0.818359375
                },
                {
                    "start": 252.8,
                    "end": 253.12,
                    "word": " sample",
                    "probability": 0.77490234375
                },
                {
                    "start": 253.12,
                    "end": 253.6,
                    "word": ".",
                    "probability": 0.65185546875
                }
            ]
        },
        {
            "id": 21,
            "seek": 14082,
            "start": 253.6,
            "end": 259.46,
            "text": " We can use it for on... in new ways that concept drift couldn't have been applied before now.",
            "tokens": [
                50376,
                492,
                393,
                764,
                309,
                337,
                322,
                485,
                294,
                777,
                2098,
                300,
                3410,
                19699,
                2809,
                380,
                362,
                668,
                6456,
                949,
                586,
                13,
                50628
            ],
            "temperature": 0.0,
            "avg_logprob": -0.26120282569021547,
            "compression_ratio": 1.6721311475409837,
            "no_speech_prob": 0.0002065896987915039,
            "words": [
                {
                    "start": 253.6,
                    "end": 254.08,
                    "word": " We",
                    "probability": 0.556640625
                },
                {
                    "start": 254.08,
                    "end": 254.26,
                    "word": " can",
                    "probability": 0.814453125
                },
                {
                    "start": 254.26,
                    "end": 254.48,
                    "word": " use",
                    "probability": 0.82958984375
                },
                {
                    "start": 254.48,
                    "end": 254.6,
                    "word": " it",
                    "probability": 0.890625
                },
                {
                    "start": 254.6,
                    "end": 254.74,
                    "word": " for",
                    "probability": 0.330810546875
                },
                {
                    "start": 254.74,
                    "end": 255.78,
                    "word": " on...",
                    "probability": 0.296722412109375
                },
                {
                    "start": 255.78,
                    "end": 255.94,
                    "word": " in",
                    "probability": 0.59619140625
                },
                {
                    "start": 255.94,
                    "end": 256.2,
                    "word": " new",
                    "probability": 0.81591796875
                },
                {
                    "start": 256.2,
                    "end": 256.5,
                    "word": " ways",
                    "probability": 0.8076171875
                },
                {
                    "start": 256.5,
                    "end": 256.66,
                    "word": " that",
                    "probability": 0.8359375
                },
                {
                    "start": 256.66,
                    "end": 257.02,
                    "word": " concept",
                    "probability": 0.336181640625
                },
                {
                    "start": 257.02,
                    "end": 257.18,
                    "word": " drift",
                    "probability": 0.615234375
                },
                {
                    "start": 257.18,
                    "end": 257.5,
                    "word": " couldn't",
                    "probability": 0.830078125
                },
                {
                    "start": 257.5,
                    "end": 257.62,
                    "word": " have",
                    "probability": 0.86572265625
                },
                {
                    "start": 257.62,
                    "end": 257.74,
                    "word": " been",
                    "probability": 0.80908203125
                },
                {
                    "start": 257.74,
                    "end": 258.04,
                    "word": " applied",
                    "probability": 0.6923828125
                },
                {
                    "start": 258.04,
                    "end": 258.58,
                    "word": " before",
                    "probability": 0.7939453125
                },
                {
                    "start": 258.58,
                    "end": 259.0,
                    "word": " now",
                    "probability": 0.85693359375
                },
                {
                    "start": 259.0,
                    "end": 259.46,
                    "word": ".",
                    "probability": 0.78369140625
                }
            ]
        },
        {
            "id": 22,
            "seek": 14082,
            "start": 259.66,
            "end": 262.56,
            "text": " So, we're going to go over the objective, why we care about concept drift,",
            "tokens": [
                50656,
                407,
                11,
                321,
                434,
                516,
                281,
                352,
                670,
                264,
                10024,
                11,
                983,
                321,
                1127,
                466,
                3410,
                19699,
                11,
                50816
            ],
            "temperature": 0.0,
            "avg_logprob": -0.26120282569021547,
            "compression_ratio": 1.6721311475409837,
            "no_speech_prob": 0.0002065896987915039,
            "words": [
                {
                    "start": 259.66,
                    "end": 259.86,
                    "word": " So",
                    "probability": 0.1890869140625
                },
                {
                    "start": 259.86,
                    "end": 259.96,
                    "word": ",",
                    "probability": 0.22509765625
                },
                {
                    "start": 259.96,
                    "end": 260.1,
                    "word": " we're",
                    "probability": 0.8896484375
                },
                {
                    "start": 260.1,
                    "end": 260.14,
                    "word": " going",
                    "probability": 0.6474609375
                },
                {
                    "start": 260.14,
                    "end": 260.18,
                    "word": " to",
                    "probability": 0.904296875
                },
                {
                    "start": 260.18,
                    "end": 260.3,
                    "word": " go",
                    "probability": 0.8955078125
                },
                {
                    "start": 260.3,
                    "end": 260.4,
                    "word": " over",
                    "probability": 0.256591796875
                },
                {
                    "start": 260.4,
                    "end": 260.48,
                    "word": " the",
                    "probability": 0.80322265625
                },
                {
                    "start": 260.48,
                    "end": 260.8,
                    "word": " objective",
                    "probability": 0.78125
                },
                {
                    "start": 260.8,
                    "end": 260.98,
                    "word": ",",
                    "probability": 0.59130859375
                },
                {
                    "start": 260.98,
                    "end": 261.12,
                    "word": " why",
                    "probability": 0.7890625
                },
                {
                    "start": 261.12,
                    "end": 261.3,
                    "word": " we",
                    "probability": 0.88232421875
                },
                {
                    "start": 261.3,
                    "end": 261.5,
                    "word": " care",
                    "probability": 0.8232421875
                },
                {
                    "start": 261.5,
                    "end": 261.66,
                    "word": " about",
                    "probability": 0.81005859375
                },
                {
                    "start": 261.66,
                    "end": 261.98,
                    "word": " concept",
                    "probability": 0.66552734375
                },
                {
                    "start": 261.98,
                    "end": 262.26,
                    "word": " drift",
                    "probability": 0.87939453125
                },
                {
                    "start": 262.26,
                    "end": 262.56,
                    "word": ",",
                    "probability": 0.81005859375
                }
            ]
        },
        {
            "id": 23,
            "seek": 14082,
            "start": 262.56,
            "end": 268.36,
            "text": " mathematically how it's done, then a couple results, and then finishing up with next steps.",
            "tokens": [
                50816,
                44003,
                577,
                309,
                311,
                1096,
                11,
                550,
                257,
                1916,
                3542,
                11,
                293,
                550,
                12693,
                493,
                365,
                958,
                4439,
                13,
                51076
            ],
            "temperature": 0.0,
            "avg_logprob": -0.26120282569021547,
            "compression_ratio": 1.6721311475409837,
            "no_speech_prob": 0.0002065896987915039,
            "words": [
                {
                    "start": 262.56,
                    "end": 263.1,
                    "word": " mathematically",
                    "probability": 0.46240234375
                },
                {
                    "start": 263.1,
                    "end": 263.5,
                    "word": " how",
                    "probability": 0.5234375
                },
                {
                    "start": 263.5,
                    "end": 263.72,
                    "word": " it's",
                    "probability": 0.9482421875
                },
                {
                    "start": 263.72,
                    "end": 264.0,
                    "word": " done",
                    "probability": 0.798828125
                },
                {
                    "start": 264.0,
                    "end": 264.3,
                    "word": ",",
                    "probability": 0.849609375
                },
                {
                    "start": 264.3,
                    "end": 264.48,
                    "word": " then",
                    "probability": 0.6767578125
                },
                {
                    "start": 264.6,
                    "end": 264.74,
                    "word": " a couple",
                    "probability": 0.798828125
                },
                {
                    "start": 264.74,
                    "end": 265.08,
                    "word": " results",
                    "probability": 0.2320556640625
                },
                {
                    "start": 265.08,
                    "end": 265.44,
                    "word": ",",
                    "probability": 0.56005859375
                },
                {
                    "start": 265.44,
                    "end": 265.58,
                    "word": " and",
                    "probability": 0.88671875
                },
                {
                    "start": 265.58,
                    "end": 265.86,
                    "word": " then",
                    "probability": 0.70947265625
                },
                {
                    "start": 265.86,
                    "end": 266.66,
                    "word": " finishing",
                    "probability": 0.779296875
                },
                {
                    "start": 266.66,
                    "end": 266.94,
                    "word": " up",
                    "probability": 0.91357421875
                },
                {
                    "start": 266.94,
                    "end": 267.1,
                    "word": " with",
                    "probability": 0.81298828125
                },
                {
                    "start": 267.1,
                    "end": 267.42,
                    "word": " next",
                    "probability": 0.802734375
                },
                {
                    "start": 267.42,
                    "end": 267.9,
                    "word": " steps",
                    "probability": 0.7412109375
                },
                {
                    "start": 267.9,
                    "end": 268.36,
                    "word": ".",
                    "probability": 0.98291015625
                }
            ]
        },
        {
            "id": 24,
            "seek": 14082,
            "start": 268.86,
            "end": 273.78,
            "text": " So, why do we care about data set drifts? Well, our data drifts.",
            "tokens": [
                51132,
                407,
                11,
                983,
                360,
                321,
                1127,
                466,
                1412,
                992,
                1224,
                8065,
                30,
                1042,
                11,
                527,
                1412,
                1224,
                8065,
                13,
                51352
            ],
            "temperature": 0.0,
            "avg_logprob": -0.26120282569021547,
            "compression_ratio": 1.6721311475409837,
            "no_speech_prob": 0.0002065896987915039,
            "words": [
                {
                    "start": 268.86,
                    "end": 269.3,
                    "word": " So",
                    "probability": 0.89453125
                },
                {
                    "start": 269.3,
                    "end": 269.52,
                    "word": ",",
                    "probability": 0.76513671875
                },
                {
                    "start": 269.52,
                    "end": 269.68,
                    "word": " why",
                    "probability": 0.80908203125
                },
                {
                    "start": 269.68,
                    "end": 269.82,
                    "word": " do",
                    "probability": 0.91650390625
                },
                {
                    "start": 269.82,
                    "end": 269.94,
                    "word": " we",
                    "probability": 0.9072265625
                },
                {
                    "start": 269.94,
                    "end": 270.14,
                    "word": " care",
                    "probability": 0.80810546875
                },
                {
                    "start": 270.14,
                    "end": 270.36,
                    "word": " about",
                    "probability": 0.818359375
                },
                {
                    "start": 270.36,
                    "end": 270.54,
                    "word": " data",
                    "probability": 0.32177734375
                },
                {
                    "start": 270.54,
                    "end": 270.74,
                    "word": " set",
                    "probability": 0.79150390625
                },
                {
                    "start": 270.74,
                    "end": 270.94,
                    "word": " drifts",
                    "probability": 0.807861328125
                },
                {
                    "start": 270.94,
                    "end": 271.18,
                    "word": "?",
                    "probability": 0.97998046875
                },
                {
                    "start": 271.18,
                    "end": 271.44,
                    "word": " Well",
                    "probability": 0.73681640625
                },
                {
                    "start": 271.44,
                    "end": 271.6,
                    "word": ",",
                    "probability": 0.89501953125
                },
                {
                    "start": 271.6,
                    "end": 271.8,
                    "word": " our",
                    "probability": 0.8125
                },
                {
                    "start": 271.8,
                    "end": 272.14,
                    "word": " data",
                    "probability": 0.81396484375
                },
                {
                    "start": 272.14,
                    "end": 273.32,
                    "word": " drifts",
                    "probability": 0.829833984375
                },
                {
                    "start": 273.32,
                    "end": 273.78,
                    "word": ".",
                    "probability": 0.86767578125
                }
            ]
        },
        {
            "id": 25,
            "seek": 14082,
            "start": 275.3,
            "end": 281.04,
            "text": " So, I work with malware and there are new malware families coming out every month.",
            "tokens": [
                51448,
                407,
                11,
                286,
                589,
                365,
                40747,
                293,
                456,
                366,
                777,
                40747,
                4466,
                1348,
                484,
                633,
                1618,
                13,
                51732
            ],
            "temperature": 0.0,
            "avg_logprob": -0.26120282569021547,
            "compression_ratio": 1.6721311475409837,
            "no_speech_prob": 0.0002065896987915039,
            "words": [
                {
                    "start": 275.3,
                    "end": 275.76,
                    "word": " So",
                    "probability": 0.66357421875
                },
                {
                    "start": 275.76,
                    "end": 276.22,
                    "word": ",",
                    "probability": 0.787109375
                },
                {
                    "start": 276.22,
                    "end": 276.66,
                    "word": " I",
                    "probability": 0.6826171875
                },
                {
                    "start": 276.66,
                    "end": 277.46,
                    "word": " work",
                    "probability": 0.78564453125
                },
                {
                    "start": 277.46,
                    "end": 277.6,
                    "word": " with",
                    "probability": 0.8115234375
                },
                {
                    "start": 277.6,
                    "end": 277.84,
                    "word": " malware",
                    "probability": 0.81298828125
                },
                {
                    "start": 277.84,
                    "end": 278.32,
                    "word": " and",
                    "probability": 0.40283203125
                },
                {
                    "start": 278.32,
                    "end": 278.44,
                    "word": " there",
                    "probability": 0.85693359375
                },
                {
                    "start": 278.44,
                    "end": 278.52,
                    "word": " are",
                    "probability": 0.8955078125
                },
                {
                    "start": 278.52,
                    "end": 278.68,
                    "word": " new",
                    "probability": 0.74658203125
                },
                {
                    "start": 278.68,
                    "end": 279.1,
                    "word": " malware",
                    "probability": 0.841796875
                },
                {
                    "start": 279.1,
                    "end": 279.42,
                    "word": " families",
                    "probability": 0.65771484375
                },
                {
                    "start": 279.42,
                    "end": 279.72,
                    "word": " coming",
                    "probability": 0.8193359375
                },
                {
                    "start": 279.72,
                    "end": 279.9,
                    "word": " out",
                    "probability": 0.68701171875
                },
                {
                    "start": 279.9,
                    "end": 280.28,
                    "word": " every",
                    "probability": 0.7412109375
                },
                {
                    "start": 280.28,
                    "end": 280.62,
                    "word": " month",
                    "probability": 0.83447265625
                },
                {
                    "start": 280.62,
                    "end": 281.04,
                    "word": ".",
                    "probability": 0.9697265625
                }
            ]
        },
        {
            "id": 26,
            "seek": 16782,
            "start": 281.04,
            "end": 285.3,
            "text": " There are new benign pieces, examples of software coming out every month.",
            "tokens": [
                50364,
                821,
                366,
                777,
                3271,
                788,
                3755,
                11,
                5110,
                295,
                4722,
                1348,
                484,
                633,
                1618,
                13,
                50580
            ],
            "temperature": 0.0,
            "avg_logprob": -0.24403994845360824,
            "compression_ratio": 1.728395061728395,
            "no_speech_prob": 0.00070953369140625,
            "words": [
                {
                    "start": 281.04,
                    "end": 281.22,
                    "word": " There",
                    "probability": 0.62451171875
                },
                {
                    "start": 281.22,
                    "end": 281.32,
                    "word": " are",
                    "probability": 0.89453125
                },
                {
                    "start": 281.32,
                    "end": 281.54,
                    "word": " new",
                    "probability": 0.7685546875
                },
                {
                    "start": 281.54,
                    "end": 282.24,
                    "word": " benign",
                    "probability": 0.83837890625
                },
                {
                    "start": 282.24,
                    "end": 282.64,
                    "word": " pieces",
                    "probability": 0.50830078125
                },
                {
                    "start": 282.64,
                    "end": 282.88,
                    "word": ",",
                    "probability": 0.239501953125
                },
                {
                    "start": 282.88,
                    "end": 283.3,
                    "word": " examples",
                    "probability": 0.763671875
                },
                {
                    "start": 283.3,
                    "end": 283.5,
                    "word": " of",
                    "probability": 0.8349609375
                },
                {
                    "start": 283.5,
                    "end": 283.78,
                    "word": " software",
                    "probability": 0.8330078125
                },
                {
                    "start": 283.78,
                    "end": 284.06,
                    "word": " coming",
                    "probability": 0.80126953125
                },
                {
                    "start": 284.06,
                    "end": 284.24,
                    "word": " out",
                    "probability": 0.79833984375
                },
                {
                    "start": 284.24,
                    "end": 284.68,
                    "word": " every",
                    "probability": 0.7197265625
                },
                {
                    "start": 284.68,
                    "end": 285.04,
                    "word": " month",
                    "probability": 0.83740234375
                },
                {
                    "start": 285.04,
                    "end": 285.3,
                    "word": ".",
                    "probability": 0.9423828125
                }
            ]
        },
        {
            "id": 27,
            "seek": 16782,
            "start": 285.3,
            "end": 289.58,
            "text": " Every time that Adobe patches Photoshop, it changes the pattern slightly and it changes",
            "tokens": [
                50580,
                2048,
                565,
                300,
                24862,
                26531,
                20821,
                11,
                309,
                2962,
                264,
                5102,
                4748,
                293,
                309,
                2962,
                50792
            ],
            "temperature": 0.0,
            "avg_logprob": -0.24403994845360824,
            "compression_ratio": 1.728395061728395,
            "no_speech_prob": 0.00070953369140625,
            "words": [
                {
                    "start": 285.3,
                    "end": 285.62,
                    "word": " Every",
                    "probability": 0.80810546875
                },
                {
                    "start": 285.62,
                    "end": 285.9,
                    "word": " time",
                    "probability": 0.80029296875
                },
                {
                    "start": 285.9,
                    "end": 286.06,
                    "word": " that",
                    "probability": 0.86767578125
                },
                {
                    "start": 286.06,
                    "end": 286.34,
                    "word": " Adobe",
                    "probability": 0.7333984375
                },
                {
                    "start": 286.34,
                    "end": 286.82,
                    "word": " patches",
                    "probability": 0.7568359375
                },
                {
                    "start": 286.82,
                    "end": 287.3,
                    "word": " Photoshop",
                    "probability": 0.5732421875
                },
                {
                    "start": 287.3,
                    "end": 287.62,
                    "word": ",",
                    "probability": 0.8505859375
                },
                {
                    "start": 287.62,
                    "end": 287.84,
                    "word": " it",
                    "probability": 0.896484375
                },
                {
                    "start": 287.84,
                    "end": 288.18,
                    "word": " changes",
                    "probability": 0.7763671875
                },
                {
                    "start": 288.18,
                    "end": 288.36,
                    "word": " the",
                    "probability": 0.81298828125
                },
                {
                    "start": 288.36,
                    "end": 288.64,
                    "word": " pattern",
                    "probability": 0.693359375
                },
                {
                    "start": 288.64,
                    "end": 289.04,
                    "word": " slightly",
                    "probability": 0.76806640625
                },
                {
                    "start": 289.04,
                    "end": 289.2,
                    "word": " and",
                    "probability": 0.63623046875
                },
                {
                    "start": 289.2,
                    "end": 289.3,
                    "word": " it",
                    "probability": 0.830078125
                },
                {
                    "start": 289.3,
                    "end": 289.58,
                    "word": " changes",
                    "probability": 0.79052734375
                }
            ]
        },
        {
            "id": 28,
            "seek": 16782,
            "start": 289.58,
            "end": 294.98,
            "text": " its location slightly in our data set. Every single time a new piece of malware comes up,",
            "tokens": [
                50792,
                1080,
                4914,
                4748,
                294,
                527,
                1412,
                992,
                13,
                2048,
                2167,
                565,
                257,
                777,
                2522,
                295,
                40747,
                1487,
                493,
                11,
                51052
            ],
            "temperature": 0.0,
            "avg_logprob": -0.24403994845360824,
            "compression_ratio": 1.728395061728395,
            "no_speech_prob": 0.00070953369140625,
            "words": [
                {
                    "start": 289.58,
                    "end": 289.84,
                    "word": " its",
                    "probability": 0.79150390625
                },
                {
                    "start": 289.84,
                    "end": 290.3,
                    "word": " location",
                    "probability": 0.80615234375
                },
                {
                    "start": 290.3,
                    "end": 290.76,
                    "word": " slightly",
                    "probability": 0.80029296875
                },
                {
                    "start": 290.76,
                    "end": 290.98,
                    "word": " in",
                    "probability": 0.900390625
                },
                {
                    "start": 290.98,
                    "end": 291.1,
                    "word": " our",
                    "probability": 0.82275390625
                },
                {
                    "start": 291.1,
                    "end": 291.3,
                    "word": " data",
                    "probability": 0.365478515625
                },
                {
                    "start": 291.3,
                    "end": 291.58,
                    "word": " set",
                    "probability": 0.88232421875
                },
                {
                    "start": 291.58,
                    "end": 292.14,
                    "word": ".",
                    "probability": 0.98876953125
                },
                {
                    "start": 292.3,
                    "end": 292.64,
                    "word": " Every",
                    "probability": 0.77734375
                },
                {
                    "start": 292.64,
                    "end": 292.88,
                    "word": " single",
                    "probability": 0.884765625
                },
                {
                    "start": 292.88,
                    "end": 293.18,
                    "word": " time",
                    "probability": 0.79931640625
                },
                {
                    "start": 293.32,
                    "end": 293.5,
                    "word": " a new",
                    "probability": 0.81787109375
                },
                {
                    "start": 293.5,
                    "end": 293.9,
                    "word": " piece",
                    "probability": 0.818359375
                },
                {
                    "start": 293.9,
                    "end": 294.04,
                    "word": " of",
                    "probability": 0.90380859375
                },
                {
                    "start": 294.04,
                    "end": 294.22,
                    "word": " malware",
                    "probability": 0.88232421875
                },
                {
                    "start": 294.22,
                    "end": 294.48,
                    "word": " comes",
                    "probability": 0.79833984375
                },
                {
                    "start": 294.48,
                    "end": 294.74,
                    "word": " up",
                    "probability": 0.9052734375
                },
                {
                    "start": 294.74,
                    "end": 294.98,
                    "word": ",",
                    "probability": 0.86572265625
                }
            ]
        },
        {
            "id": 29,
            "seek": 16782,
            "start": 294.98,
            "end": 300.88,
            "text": " it's a new... it changes the way our... the malware data is distributed slightly.",
            "tokens": [
                51052,
                309,
                311,
                257,
                777,
                485,
                309,
                2962,
                264,
                636,
                527,
                485,
                264,
                40747,
                1412,
                307,
                12631,
                4748,
                13,
                51332
            ],
            "temperature": 0.0,
            "avg_logprob": -0.24403994845360824,
            "compression_ratio": 1.728395061728395,
            "no_speech_prob": 0.00070953369140625,
            "words": [
                {
                    "start": 294.98,
                    "end": 295.16,
                    "word": " it's",
                    "probability": 0.69091796875
                },
                {
                    "start": 295.32,
                    "end": 295.7,
                    "word": " a new...",
                    "probability": 0.457183837890625
                },
                {
                    "start": 295.7,
                    "end": 296.02,
                    "word": " it",
                    "probability": 0.458740234375
                },
                {
                    "start": 296.02,
                    "end": 296.56,
                    "word": " changes",
                    "probability": 0.80224609375
                },
                {
                    "start": 296.56,
                    "end": 297.38,
                    "word": " the",
                    "probability": 0.56640625
                },
                {
                    "start": 297.38,
                    "end": 297.6,
                    "word": " way",
                    "probability": 0.904296875
                },
                {
                    "start": 297.6,
                    "end": 298.0,
                    "word": " our...",
                    "probability": 0.393310546875
                },
                {
                    "start": 298.0,
                    "end": 298.48,
                    "word": " the",
                    "probability": 0.64697265625
                },
                {
                    "start": 298.48,
                    "end": 298.8,
                    "word": " malware",
                    "probability": 0.857421875
                },
                {
                    "start": 298.8,
                    "end": 299.06,
                    "word": " data",
                    "probability": 0.81689453125
                },
                {
                    "start": 299.06,
                    "end": 299.24,
                    "word": " is",
                    "probability": 0.89892578125
                },
                {
                    "start": 299.24,
                    "end": 299.68,
                    "word": " distributed",
                    "probability": 0.859375
                },
                {
                    "start": 299.68,
                    "end": 300.32,
                    "word": " slightly",
                    "probability": 0.77392578125
                },
                {
                    "start": 300.32,
                    "end": 300.88,
                    "word": ".",
                    "probability": 0.99365234375
                }
            ]
        },
        {
            "id": 30,
            "seek": 16782,
            "start": 301.5,
            "end": 308.04,
            "text": " Another problem... well, what those do, basically, which gets into the second problem,",
            "tokens": [
                51384,
                3996,
                1154,
                485,
                731,
                11,
                437,
                729,
                360,
                11,
                1936,
                11,
                597,
                2170,
                666,
                264,
                1150,
                1154,
                11,
                51712
            ],
            "temperature": 0.0,
            "avg_logprob": -0.24403994845360824,
            "compression_ratio": 1.728395061728395,
            "no_speech_prob": 0.00070953369140625,
            "words": [
                {
                    "start": 301.5,
                    "end": 301.8,
                    "word": " Another",
                    "probability": 0.78515625
                },
                {
                    "start": 301.8,
                    "end": 303.44,
                    "word": " problem...",
                    "probability": 0.700927734375
                },
                {
                    "start": 303.44,
                    "end": 304.64,
                    "word": " well",
                    "probability": 0.254638671875
                },
                {
                    "start": 304.64,
                    "end": 304.86,
                    "word": ",",
                    "probability": 0.79833984375
                },
                {
                    "start": 304.86,
                    "end": 305.06,
                    "word": " what",
                    "probability": 0.88330078125
                },
                {
                    "start": 305.06,
                    "end": 305.28,
                    "word": " those",
                    "probability": 0.78466796875
                },
                {
                    "start": 305.28,
                    "end": 305.84,
                    "word": " do",
                    "probability": 0.9052734375
                },
                {
                    "start": 305.84,
                    "end": 306.02,
                    "word": ",",
                    "probability": 0.432861328125
                },
                {
                    "start": 306.02,
                    "end": 306.34,
                    "word": " basically",
                    "probability": 0.88720703125
                },
                {
                    "start": 306.34,
                    "end": 306.58,
                    "word": ",",
                    "probability": 0.90869140625
                },
                {
                    "start": 306.58,
                    "end": 307.0,
                    "word": " which",
                    "probability": 0.9111328125
                },
                {
                    "start": 307.0,
                    "end": 307.24,
                    "word": " gets",
                    "probability": 0.78076171875
                },
                {
                    "start": 307.24,
                    "end": 307.38,
                    "word": " into",
                    "probability": 0.72119140625
                },
                {
                    "start": 307.38,
                    "end": 307.52,
                    "word": " the",
                    "probability": 0.818359375
                },
                {
                    "start": 307.52,
                    "end": 307.66,
                    "word": " second",
                    "probability": 0.79638671875
                },
                {
                    "start": 307.66,
                    "end": 307.94,
                    "word": " problem",
                    "probability": 0.77001953125
                },
                {
                    "start": 307.94,
                    "end": 308.04,
                    "word": ",",
                    "probability": 0.84228515625
                }
            ]
        },
        {
            "id": 31,
            "seek": 19458,
            "start": 308.04,
            "end": 314.72,
            "text": " they lower the efficacy of our models that we deploy. So, our models expect data in a sort of",
            "tokens": [
                50364,
                436,
                3126,
                264,
                33492,
                295,
                527,
                5245,
                300,
                321,
                7274,
                13,
                407,
                11,
                527,
                5245,
                2066,
                1412,
                294,
                257,
                1333,
                295,
                50696
            ],
            "temperature": 0.0,
            "avg_logprob": -0.26594066619873047,
            "compression_ratio": 1.7043478260869565,
            "no_speech_prob": 0.0029354095458984375,
            "words": [
                {
                    "start": 308.04,
                    "end": 308.22,
                    "word": " they",
                    "probability": 0.03472900390625
                },
                {
                    "start": 308.22,
                    "end": 308.62,
                    "word": " lower",
                    "probability": 0.802734375
                },
                {
                    "start": 308.62,
                    "end": 308.9,
                    "word": " the",
                    "probability": 0.81591796875
                },
                {
                    "start": 308.9,
                    "end": 309.24,
                    "word": " efficacy",
                    "probability": 0.84033203125
                },
                {
                    "start": 309.24,
                    "end": 309.52,
                    "word": " of",
                    "probability": 0.9033203125
                },
                {
                    "start": 309.52,
                    "end": 309.66,
                    "word": " our",
                    "probability": 0.8173828125
                },
                {
                    "start": 309.66,
                    "end": 309.94,
                    "word": " models",
                    "probability": 0.7880859375
                },
                {
                    "start": 309.94,
                    "end": 310.08,
                    "word": " that",
                    "probability": 0.87353515625
                },
                {
                    "start": 310.08,
                    "end": 310.18,
                    "word": " we",
                    "probability": 0.9013671875
                },
                {
                    "start": 310.18,
                    "end": 310.54,
                    "word": " deploy",
                    "probability": 0.74951171875
                },
                {
                    "start": 310.54,
                    "end": 311.06,
                    "word": ".",
                    "probability": 0.92578125
                },
                {
                    "start": 311.7,
                    "end": 312.22,
                    "word": " So",
                    "probability": 0.791015625
                },
                {
                    "start": 312.22,
                    "end": 312.34,
                    "word": ",",
                    "probability": 0.143310546875
                },
                {
                    "start": 312.34,
                    "end": 312.44,
                    "word": " our",
                    "probability": 0.8193359375
                },
                {
                    "start": 312.44,
                    "end": 312.7,
                    "word": " models",
                    "probability": 0.796875
                },
                {
                    "start": 312.7,
                    "end": 313.3,
                    "word": " expect",
                    "probability": 0.85791015625
                },
                {
                    "start": 313.3,
                    "end": 313.76,
                    "word": " data",
                    "probability": 0.8369140625
                },
                {
                    "start": 313.76,
                    "end": 314.06,
                    "word": " in",
                    "probability": 0.82470703125
                },
                {
                    "start": 314.2,
                    "end": 314.36,
                    "word": " a sort",
                    "probability": 0.759765625
                },
                {
                    "start": 314.36,
                    "end": 314.72,
                    "word": " of",
                    "probability": 0.890625
                }
            ]
        },
        {
            "id": 32,
            "seek": 19458,
            "start": 314.72,
            "end": 320.34,
            "text": " the same sort of general location as our training data. And every time we deploy, we are kind of",
            "tokens": [
                50696,
                264,
                912,
                1333,
                295,
                2674,
                4914,
                382,
                527,
                3097,
                1412,
                13,
                400,
                633,
                565,
                321,
                7274,
                11,
                321,
                366,
                733,
                295,
                50976
            ],
            "temperature": 0.0,
            "avg_logprob": -0.26594066619873047,
            "compression_ratio": 1.7043478260869565,
            "no_speech_prob": 0.0029354095458984375,
            "words": [
                {
                    "start": 314.72,
                    "end": 314.86,
                    "word": " the",
                    "probability": 0.62451171875
                },
                {
                    "start": 314.86,
                    "end": 315.28,
                    "word": " same",
                    "probability": 0.8095703125
                },
                {
                    "start": 315.28,
                    "end": 315.52,
                    "word": " sort",
                    "probability": 0.6865234375
                },
                {
                    "start": 315.52,
                    "end": 315.66,
                    "word": " of",
                    "probability": 0.90576171875
                },
                {
                    "start": 315.66,
                    "end": 316.16,
                    "word": " general",
                    "probability": 0.7412109375
                },
                {
                    "start": 316.16,
                    "end": 316.68,
                    "word": " location",
                    "probability": 0.71240234375
                },
                {
                    "start": 316.68,
                    "end": 317.56,
                    "word": " as",
                    "probability": 0.875
                },
                {
                    "start": 317.56,
                    "end": 317.86,
                    "word": " our",
                    "probability": 0.810546875
                },
                {
                    "start": 317.86,
                    "end": 318.12,
                    "word": " training",
                    "probability": 0.798828125
                },
                {
                    "start": 318.12,
                    "end": 318.42,
                    "word": " data",
                    "probability": 0.8310546875
                },
                {
                    "start": 318.42,
                    "end": 318.64,
                    "word": ".",
                    "probability": 0.52880859375
                },
                {
                    "start": 318.64,
                    "end": 318.74,
                    "word": " And",
                    "probability": 0.87451171875
                },
                {
                    "start": 318.74,
                    "end": 318.92,
                    "word": " every",
                    "probability": 0.736328125
                },
                {
                    "start": 318.92,
                    "end": 319.14,
                    "word": " time",
                    "probability": 0.80419921875
                },
                {
                    "start": 319.14,
                    "end": 319.26,
                    "word": " we",
                    "probability": 0.9033203125
                },
                {
                    "start": 319.26,
                    "end": 319.64,
                    "word": " deploy",
                    "probability": 0.77197265625
                },
                {
                    "start": 319.64,
                    "end": 319.84,
                    "word": ",",
                    "probability": 0.83544921875
                },
                {
                    "start": 319.84,
                    "end": 319.94,
                    "word": " we",
                    "probability": 0.89794921875
                },
                {
                    "start": 319.94,
                    "end": 320.04,
                    "word": " are",
                    "probability": 0.67431640625
                },
                {
                    "start": 320.04,
                    "end": 320.22,
                    "word": " kind",
                    "probability": 0.77001953125
                },
                {
                    "start": 320.22,
                    "end": 320.34,
                    "word": " of",
                    "probability": 0.90673828125
                }
            ]
        },
        {
            "id": 33,
            "seek": 19458,
            "start": 320.34,
            "end": 327.64,
            "text": " fixing our model's view of the world at a certain day, and then we're training it and we're testing",
            "tokens": [
                50976,
                19442,
                527,
                2316,
                311,
                1910,
                295,
                264,
                1002,
                412,
                257,
                1629,
                786,
                11,
                293,
                550,
                321,
                434,
                3097,
                309,
                293,
                321,
                434,
                4997,
                51340
            ],
            "temperature": 0.0,
            "avg_logprob": -0.26594066619873047,
            "compression_ratio": 1.7043478260869565,
            "no_speech_prob": 0.0029354095458984375,
            "words": [
                {
                    "start": 320.34,
                    "end": 320.72,
                    "word": " fixing",
                    "probability": 0.869140625
                },
                {
                    "start": 320.72,
                    "end": 321.26,
                    "word": " our",
                    "probability": 0.81640625
                },
                {
                    "start": 321.26,
                    "end": 322.0,
                    "word": " model's",
                    "probability": 0.5882568359375
                },
                {
                    "start": 322.0,
                    "end": 322.12,
                    "word": " view",
                    "probability": 0.828125
                },
                {
                    "start": 322.12,
                    "end": 322.22,
                    "word": " of",
                    "probability": 0.90380859375
                },
                {
                    "start": 322.22,
                    "end": 322.34,
                    "word": " the",
                    "probability": 0.818359375
                },
                {
                    "start": 322.34,
                    "end": 322.58,
                    "word": " world",
                    "probability": 0.88232421875
                },
                {
                    "start": 322.58,
                    "end": 322.88,
                    "word": " at",
                    "probability": 0.89013671875
                },
                {
                    "start": 323.0,
                    "end": 323.14,
                    "word": " a certain",
                    "probability": 0.796875
                },
                {
                    "start": 323.14,
                    "end": 323.58,
                    "word": " day",
                    "probability": 0.89306640625
                },
                {
                    "start": 323.58,
                    "end": 323.76,
                    "word": ",",
                    "probability": 0.35498046875
                },
                {
                    "start": 323.76,
                    "end": 324.42,
                    "word": " and",
                    "probability": 0.890625
                },
                {
                    "start": 324.42,
                    "end": 324.92,
                    "word": " then",
                    "probability": 0.73046875
                },
                {
                    "start": 324.92,
                    "end": 325.46,
                    "word": " we're",
                    "probability": 0.894287109375
                },
                {
                    "start": 325.46,
                    "end": 326.26,
                    "word": " training",
                    "probability": 0.1002197265625
                },
                {
                    "start": 326.26,
                    "end": 326.52,
                    "word": " it",
                    "probability": 0.7705078125
                },
                {
                    "start": 326.52,
                    "end": 326.78,
                    "word": " and",
                    "probability": 0.1602783203125
                },
                {
                    "start": 326.78,
                    "end": 327.0,
                    "word": " we're",
                    "probability": 0.90771484375
                },
                {
                    "start": 327.0,
                    "end": 327.64,
                    "word": " testing",
                    "probability": 0.73095703125
                }
            ]
        },
        {
            "id": 34,
            "seek": 19458,
            "start": 327.64,
            "end": 334.8,
            "text": " it on essentially future data. So, it doesn't have a perspective for the... if it's trained in March,",
            "tokens": [
                51340,
                309,
                322,
                4476,
                2027,
                1412,
                13,
                407,
                11,
                309,
                1177,
                380,
                362,
                257,
                4585,
                337,
                264,
                485,
                498,
                309,
                311,
                8895,
                294,
                6129,
                11,
                51696
            ],
            "temperature": 0.0,
            "avg_logprob": -0.26594066619873047,
            "compression_ratio": 1.7043478260869565,
            "no_speech_prob": 0.0029354095458984375,
            "words": [
                {
                    "start": 327.64,
                    "end": 327.8,
                    "word": " it",
                    "probability": 0.7978515625
                },
                {
                    "start": 327.8,
                    "end": 327.96,
                    "word": " on",
                    "probability": 0.8603515625
                },
                {
                    "start": 327.96,
                    "end": 328.34,
                    "word": " essentially",
                    "probability": 0.58203125
                },
                {
                    "start": 328.34,
                    "end": 328.74,
                    "word": " future",
                    "probability": 0.91552734375
                },
                {
                    "start": 328.74,
                    "end": 329.12,
                    "word": " data",
                    "probability": 0.81982421875
                },
                {
                    "start": 329.12,
                    "end": 329.64,
                    "word": ".",
                    "probability": 0.94970703125
                },
                {
                    "start": 329.76,
                    "end": 329.88,
                    "word": " So",
                    "probability": 0.859375
                },
                {
                    "start": 329.88,
                    "end": 329.92,
                    "word": ",",
                    "probability": 0.783203125
                },
                {
                    "start": 329.92,
                    "end": 330.02,
                    "word": " it",
                    "probability": 0.70458984375
                },
                {
                    "start": 330.02,
                    "end": 330.4,
                    "word": " doesn't",
                    "probability": 0.941650390625
                },
                {
                    "start": 330.4,
                    "end": 330.72,
                    "word": " have",
                    "probability": 0.8876953125
                },
                {
                    "start": 330.86,
                    "end": 331.28,
                    "word": " a perspective",
                    "probability": 0.78076171875
                },
                {
                    "start": 331.28,
                    "end": 333.08,
                    "word": " for",
                    "probability": 0.57958984375
                },
                {
                    "start": 333.08,
                    "end": 333.9,
                    "word": " the...",
                    "probability": 0.4259033203125
                },
                {
                    "start": 333.9,
                    "end": 334.08,
                    "word": " if",
                    "probability": 0.1495361328125
                },
                {
                    "start": 334.08,
                    "end": 334.18,
                    "word": " it's",
                    "probability": 0.927734375
                },
                {
                    "start": 334.18,
                    "end": 334.4,
                    "word": " trained",
                    "probability": 0.806640625
                },
                {
                    "start": 334.4,
                    "end": 334.58,
                    "word": " in",
                    "probability": 0.88720703125
                },
                {
                    "start": 334.58,
                    "end": 334.76,
                    "word": " March",
                    "probability": 0.8994140625
                },
                {
                    "start": 334.76,
                    "end": 334.8,
                    "word": ",",
                    "probability": 0.87939453125
                }
            ]
        },
        {
            "id": 35,
            "seek": 22278,
            "start": 334.8,
            "end": 340.8,
            "text": " it doesn't have... it might not work very well at the end... data coming in at the end of April.",
            "tokens": [
                50364,
                309,
                1177,
                380,
                362,
                485,
                309,
                1062,
                406,
                589,
                588,
                731,
                412,
                264,
                917,
                485,
                1412,
                1348,
                294,
                412,
                264,
                917,
                295,
                6929,
                13,
                50644
            ],
            "temperature": 0.0,
            "avg_logprob": -0.24474897443271074,
            "compression_ratio": 1.7084870848708487,
            "no_speech_prob": 0.009857177734375,
            "words": [
                {
                    "start": 334.8,
                    "end": 334.9,
                    "word": " it",
                    "probability": 0.034027099609375
                },
                {
                    "start": 334.9,
                    "end": 335.1,
                    "word": " doesn't",
                    "probability": 0.6961669921875
                },
                {
                    "start": 335.1,
                    "end": 335.78,
                    "word": " have...",
                    "probability": 0.4656524658203125
                },
                {
                    "start": 335.78,
                    "end": 336.68,
                    "word": " it",
                    "probability": 0.470458984375
                },
                {
                    "start": 336.68,
                    "end": 336.88,
                    "word": " might",
                    "probability": 0.72900390625
                },
                {
                    "start": 336.88,
                    "end": 337.1,
                    "word": " not",
                    "probability": 0.8994140625
                },
                {
                    "start": 337.1,
                    "end": 337.64,
                    "word": " work",
                    "probability": 0.78759765625
                },
                {
                    "start": 337.64,
                    "end": 337.86,
                    "word": " very",
                    "probability": 0.73486328125
                },
                {
                    "start": 337.86,
                    "end": 338.12,
                    "word": " well",
                    "probability": 0.8759765625
                },
                {
                    "start": 338.12,
                    "end": 338.62,
                    "word": " at",
                    "probability": 0.41015625
                },
                {
                    "start": 338.62,
                    "end": 338.72,
                    "word": " the",
                    "probability": 0.80908203125
                },
                {
                    "start": 338.72,
                    "end": 338.96,
                    "word": " end...",
                    "probability": 0.46234130859375
                },
                {
                    "start": 338.96,
                    "end": 339.16,
                    "word": " data",
                    "probability": 0.6171875
                },
                {
                    "start": 339.16,
                    "end": 339.48,
                    "word": " coming",
                    "probability": 0.81396484375
                },
                {
                    "start": 339.48,
                    "end": 339.64,
                    "word": " in",
                    "probability": 0.9013671875
                },
                {
                    "start": 339.64,
                    "end": 339.74,
                    "word": " at",
                    "probability": 0.90478515625
                },
                {
                    "start": 339.74,
                    "end": 339.86,
                    "word": " the",
                    "probability": 0.8173828125
                },
                {
                    "start": 339.86,
                    "end": 340.0,
                    "word": " end",
                    "probability": 0.791015625
                },
                {
                    "start": 340.0,
                    "end": 340.1,
                    "word": " of",
                    "probability": 0.90234375
                },
                {
                    "start": 340.1,
                    "end": 340.38,
                    "word": " April",
                    "probability": 0.77490234375
                },
                {
                    "start": 340.38,
                    "end": 340.8,
                    "word": ".",
                    "probability": 0.900390625
                }
            ]
        },
        {
            "id": 36,
            "seek": 22278,
            "start": 341.8,
            "end": 346.62,
            "text": " So, one of the methods for dealing with this is you build a dashboard that basically tracks the",
            "tokens": [
                50720,
                407,
                11,
                472,
                295,
                264,
                7150,
                337,
                6260,
                365,
                341,
                307,
                291,
                1322,
                257,
                18342,
                300,
                1936,
                10218,
                264,
                50952
            ],
            "temperature": 0.0,
            "avg_logprob": -0.24474897443271074,
            "compression_ratio": 1.7084870848708487,
            "no_speech_prob": 0.009857177734375,
            "words": [
                {
                    "start": 341.8,
                    "end": 342.22,
                    "word": " So",
                    "probability": 0.81689453125
                },
                {
                    "start": 342.22,
                    "end": 342.44,
                    "word": ",",
                    "probability": 0.335693359375
                },
                {
                    "start": 342.44,
                    "end": 342.88,
                    "word": " one",
                    "probability": 0.81591796875
                },
                {
                    "start": 342.88,
                    "end": 343.02,
                    "word": " of",
                    "probability": 0.90087890625
                },
                {
                    "start": 343.02,
                    "end": 343.22,
                    "word": " the",
                    "probability": 0.81591796875
                },
                {
                    "start": 343.22,
                    "end": 343.58,
                    "word": " methods",
                    "probability": 0.83447265625
                },
                {
                    "start": 343.58,
                    "end": 343.78,
                    "word": " for",
                    "probability": 0.86669921875
                },
                {
                    "start": 343.78,
                    "end": 344.02,
                    "word": " dealing",
                    "probability": 0.7802734375
                },
                {
                    "start": 344.02,
                    "end": 344.22,
                    "word": " with",
                    "probability": 0.8115234375
                },
                {
                    "start": 344.22,
                    "end": 344.38,
                    "word": " this",
                    "probability": 0.87646484375
                },
                {
                    "start": 344.38,
                    "end": 344.52,
                    "word": " is",
                    "probability": 0.8662109375
                },
                {
                    "start": 344.52,
                    "end": 344.7,
                    "word": " you",
                    "probability": 0.7392578125
                },
                {
                    "start": 344.7,
                    "end": 344.9,
                    "word": " build",
                    "probability": 0.86474609375
                },
                {
                    "start": 345.04,
                    "end": 345.4,
                    "word": " a dashboard",
                    "probability": 0.822265625
                },
                {
                    "start": 345.4,
                    "end": 345.7,
                    "word": " that",
                    "probability": 0.8671875
                },
                {
                    "start": 345.7,
                    "end": 346.08,
                    "word": " basically",
                    "probability": 0.8984375
                },
                {
                    "start": 346.08,
                    "end": 346.44,
                    "word": " tracks",
                    "probability": 0.708984375
                },
                {
                    "start": 346.44,
                    "end": 346.62,
                    "word": " the",
                    "probability": 0.81591796875
                }
            ]
        },
        {
            "id": 37,
            "seek": 22278,
            "start": 346.62,
            "end": 353.34,
            "text": " error. And when the error gets too high, you discard the model and go on for a new one.",
            "tokens": [
                50952,
                6713,
                13,
                400,
                562,
                264,
                6713,
                2170,
                886,
                1090,
                11,
                291,
                31597,
                264,
                2316,
                293,
                352,
                322,
                337,
                257,
                777,
                472,
                13,
                51288
            ],
            "temperature": 0.0,
            "avg_logprob": -0.24474897443271074,
            "compression_ratio": 1.7084870848708487,
            "no_speech_prob": 0.009857177734375,
            "words": [
                {
                    "start": 346.62,
                    "end": 346.86,
                    "word": " error",
                    "probability": 0.82666015625
                },
                {
                    "start": 346.86,
                    "end": 347.28,
                    "word": ".",
                    "probability": 0.429443359375
                },
                {
                    "start": 349.02,
                    "end": 349.34,
                    "word": " And",
                    "probability": 0.8603515625
                },
                {
                    "start": 349.34,
                    "end": 349.52,
                    "word": " when",
                    "probability": 0.84228515625
                },
                {
                    "start": 349.52,
                    "end": 349.7,
                    "word": " the",
                    "probability": 0.82275390625
                },
                {
                    "start": 349.7,
                    "end": 349.88,
                    "word": " error",
                    "probability": 0.83154296875
                },
                {
                    "start": 349.88,
                    "end": 350.06,
                    "word": " gets",
                    "probability": 0.8115234375
                },
                {
                    "start": 350.06,
                    "end": 350.3,
                    "word": " too",
                    "probability": 0.8076171875
                },
                {
                    "start": 350.3,
                    "end": 350.78,
                    "word": " high",
                    "probability": 0.859375
                },
                {
                    "start": 350.78,
                    "end": 351.06,
                    "word": ",",
                    "probability": 0.8642578125
                },
                {
                    "start": 351.06,
                    "end": 351.54,
                    "word": " you",
                    "probability": 0.8896484375
                },
                {
                    "start": 351.54,
                    "end": 351.96,
                    "word": " discard",
                    "probability": 0.70703125
                },
                {
                    "start": 351.96,
                    "end": 352.1,
                    "word": " the",
                    "probability": 0.81298828125
                },
                {
                    "start": 352.1,
                    "end": 352.26,
                    "word": " model",
                    "probability": 0.78955078125
                },
                {
                    "start": 352.26,
                    "end": 352.38,
                    "word": " and",
                    "probability": 0.88134765625
                },
                {
                    "start": 352.38,
                    "end": 352.52,
                    "word": " go",
                    "probability": 0.90966796875
                },
                {
                    "start": 352.52,
                    "end": 352.62,
                    "word": " on",
                    "probability": 0.75830078125
                },
                {
                    "start": 352.62,
                    "end": 352.72,
                    "word": " for",
                    "probability": 0.88916015625
                },
                {
                    "start": 352.8,
                    "end": 352.88,
                    "word": " a new",
                    "probability": 0.81982421875
                },
                {
                    "start": 352.88,
                    "end": 353.06,
                    "word": " one",
                    "probability": 0.8291015625
                },
                {
                    "start": 353.06,
                    "end": 353.34,
                    "word": ".",
                    "probability": 0.9755859375
                }
            ]
        },
        {
            "id": 38,
            "seek": 22278,
            "start": 353.34,
            "end": 358.46,
            "text": " But the problem is, I don't trust VirusTotalLabels, and I don't think anyone really should trust",
            "tokens": [
                51288,
                583,
                264,
                1154,
                307,
                11,
                286,
                500,
                380,
                3361,
                39790,
                51,
                38180,
                37880,
                1625,
                11,
                293,
                286,
                500,
                380,
                519,
                2878,
                534,
                820,
                3361,
                51540
            ],
            "temperature": 0.0,
            "avg_logprob": -0.24474897443271074,
            "compression_ratio": 1.7084870848708487,
            "no_speech_prob": 0.009857177734375,
            "words": [
                {
                    "start": 353.34,
                    "end": 353.5,
                    "word": " But",
                    "probability": 0.87890625
                },
                {
                    "start": 353.5,
                    "end": 353.68,
                    "word": " the",
                    "probability": 0.70703125
                },
                {
                    "start": 353.68,
                    "end": 353.82,
                    "word": " problem",
                    "probability": 0.76953125
                },
                {
                    "start": 353.82,
                    "end": 354.1,
                    "word": " is",
                    "probability": 0.90673828125
                },
                {
                    "start": 354.1,
                    "end": 354.34,
                    "word": ",",
                    "probability": 0.482421875
                },
                {
                    "start": 354.34,
                    "end": 354.54,
                    "word": " I",
                    "probability": 0.99462890625
                },
                {
                    "start": 354.54,
                    "end": 354.82,
                    "word": " don't",
                    "probability": 0.94921875
                },
                {
                    "start": 354.82,
                    "end": 355.26,
                    "word": " trust",
                    "probability": 0.8056640625
                },
                {
                    "start": 355.26,
                    "end": 356.16,
                    "word": " VirusTotalLabels",
                    "probability": 0.663916015625
                },
                {
                    "start": 356.16,
                    "end": 356.24,
                    "word": ",",
                    "probability": 0.431396484375
                },
                {
                    "start": 356.24,
                    "end": 356.38,
                    "word": " and",
                    "probability": 0.88037109375
                },
                {
                    "start": 356.38,
                    "end": 356.6,
                    "word": " I",
                    "probability": 0.9970703125
                },
                {
                    "start": 356.6,
                    "end": 356.72,
                    "word": " don't",
                    "probability": 0.94677734375
                },
                {
                    "start": 356.72,
                    "end": 356.9,
                    "word": " think",
                    "probability": 0.80322265625
                },
                {
                    "start": 356.9,
                    "end": 357.22,
                    "word": " anyone",
                    "probability": 0.90478515625
                },
                {
                    "start": 357.22,
                    "end": 357.56,
                    "word": " really",
                    "probability": 0.763671875
                },
                {
                    "start": 357.56,
                    "end": 357.88,
                    "word": " should",
                    "probability": 0.93212890625
                },
                {
                    "start": 357.88,
                    "end": 358.46,
                    "word": " trust",
                    "probability": 0.81884765625
                }
            ]
        },
        {
            "id": 39,
            "seek": 22278,
            "start": 359.4,
            "end": 363.0,
            "text": " VirusTotalLabels. And that's the best way we have because of the sheer volume of data",
            "tokens": [
                51568,
                39790,
                51,
                38180,
                37880,
                1625,
                13,
                400,
                300,
                311,
                264,
                1151,
                636,
                321,
                362,
                570,
                295,
                264,
                23061,
                5523,
                295,
                1412,
                51792
            ],
            "temperature": 0.0,
            "avg_logprob": -0.24474897443271074,
            "compression_ratio": 1.7084870848708487,
            "no_speech_prob": 0.009857177734375,
            "words": [
                {
                    "start": 359.4,
                    "end": 359.82,
                    "word": " VirusTotalLabels",
                    "probability": 0.9470703125
                },
                {
                    "start": 359.82,
                    "end": 359.94,
                    "word": ".",
                    "probability": 0.55859375
                },
                {
                    "start": 359.94,
                    "end": 360.04,
                    "word": " And",
                    "probability": 0.892578125
                },
                {
                    "start": 360.04,
                    "end": 360.26,
                    "word": " that's",
                    "probability": 0.9384765625
                },
                {
                    "start": 360.26,
                    "end": 360.46,
                    "word": " the",
                    "probability": 0.80615234375
                },
                {
                    "start": 360.46,
                    "end": 360.74,
                    "word": " best",
                    "probability": 0.81591796875
                },
                {
                    "start": 360.74,
                    "end": 360.9,
                    "word": " way",
                    "probability": 0.896484375
                },
                {
                    "start": 360.9,
                    "end": 361.04,
                    "word": " we",
                    "probability": 0.89990234375
                },
                {
                    "start": 361.04,
                    "end": 361.28,
                    "word": " have",
                    "probability": 0.87939453125
                },
                {
                    "start": 361.28,
                    "end": 361.54,
                    "word": " because",
                    "probability": 0.42578125
                },
                {
                    "start": 361.54,
                    "end": 361.68,
                    "word": " of",
                    "probability": 0.455810546875
                },
                {
                    "start": 361.68,
                    "end": 361.78,
                    "word": " the",
                    "probability": 0.7724609375
                },
                {
                    "start": 361.78,
                    "end": 362.0,
                    "word": " sheer",
                    "probability": 0.77294921875
                },
                {
                    "start": 362.0,
                    "end": 362.5,
                    "word": " volume",
                    "probability": 0.79541015625
                },
                {
                    "start": 362.5,
                    "end": 362.74,
                    "word": " of",
                    "probability": 0.90673828125
                },
                {
                    "start": 362.74,
                    "end": 363.0,
                    "word": " data",
                    "probability": 0.81396484375
                }
            ]
        },
        {
            "id": 40,
            "seek": 24970,
            "start": 363.0,
            "end": 370.48,
            "text": " we have to kind of trust VirusTotalLabels. So, and for other things, I wouldn't trust",
            "tokens": [
                50380,
                321,
                362,
                281,
                733,
                295,
                3361,
                39790,
                51,
                38180,
                37880,
                1625,
                13,
                407,
                11,
                293,
                337,
                661,
                721,
                11,
                286,
                2759,
                380,
                3361,
                50740
            ],
            "temperature": 0.0,
            "avg_logprob": -0.2334374986588955,
            "compression_ratio": 1.5833333333333333,
            "no_speech_prob": 0.0003859996795654297,
            "words": [
                {
                    "start": 363.0,
                    "end": 363.54,
                    "word": " we",
                    "probability": 0.330810546875
                },
                {
                    "start": 363.54,
                    "end": 363.86,
                    "word": " have",
                    "probability": 0.86767578125
                },
                {
                    "start": 363.86,
                    "end": 363.98,
                    "word": " to",
                    "probability": 0.90283203125
                },
                {
                    "start": 363.98,
                    "end": 364.2,
                    "word": " kind",
                    "probability": 0.3798828125
                },
                {
                    "start": 364.2,
                    "end": 364.3,
                    "word": " of",
                    "probability": 0.91015625
                },
                {
                    "start": 364.3,
                    "end": 364.58,
                    "word": " trust",
                    "probability": 0.7900390625
                },
                {
                    "start": 364.58,
                    "end": 365.52,
                    "word": " VirusTotalLabels",
                    "probability": 0.389202880859375
                },
                {
                    "start": 365.52,
                    "end": 366.0,
                    "word": ".",
                    "probability": 0.8486328125
                },
                {
                    "start": 366.22,
                    "end": 366.32,
                    "word": " So",
                    "probability": 0.3232421875
                },
                {
                    "start": 366.32,
                    "end": 366.94,
                    "word": ",",
                    "probability": 0.51171875
                },
                {
                    "start": 366.94,
                    "end": 368.82,
                    "word": " and",
                    "probability": 0.6630859375
                },
                {
                    "start": 368.82,
                    "end": 368.98,
                    "word": " for",
                    "probability": 0.8974609375
                },
                {
                    "start": 368.98,
                    "end": 369.18,
                    "word": " other",
                    "probability": 0.78125
                },
                {
                    "start": 369.18,
                    "end": 369.48,
                    "word": " things",
                    "probability": 0.73388671875
                },
                {
                    "start": 369.48,
                    "end": 369.6,
                    "word": ",",
                    "probability": 0.69921875
                },
                {
                    "start": 369.6,
                    "end": 369.7,
                    "word": " I",
                    "probability": 0.99951171875
                },
                {
                    "start": 369.7,
                    "end": 369.98,
                    "word": " wouldn't",
                    "probability": 0.8984375
                },
                {
                    "start": 369.98,
                    "end": 370.48,
                    "word": " trust",
                    "probability": 0.8125
                }
            ]
        },
        {
            "id": 41,
            "seek": 24970,
            "start": 370.48,
            "end": 376.66,
            "text": " early labels on data because we have to move very quickly and sometimes with whitelisting and stuff.",
            "tokens": [
                50740,
                2440,
                16949,
                322,
                1412,
                570,
                321,
                362,
                281,
                1286,
                588,
                2661,
                293,
                2171,
                365,
                315,
                38906,
                468,
                278,
                293,
                1507,
                13,
                51020
            ],
            "temperature": 0.0,
            "avg_logprob": -0.2334374986588955,
            "compression_ratio": 1.5833333333333333,
            "no_speech_prob": 0.0003859996795654297,
            "words": [
                {
                    "start": 370.48,
                    "end": 371.0,
                    "word": " early",
                    "probability": 0.91845703125
                },
                {
                    "start": 371.0,
                    "end": 371.4,
                    "word": " labels",
                    "probability": 0.8662109375
                },
                {
                    "start": 371.4,
                    "end": 371.98,
                    "word": " on",
                    "probability": 0.85107421875
                },
                {
                    "start": 371.98,
                    "end": 372.28,
                    "word": " data",
                    "probability": 0.7841796875
                },
                {
                    "start": 372.28,
                    "end": 373.6,
                    "word": " because",
                    "probability": 0.491455078125
                },
                {
                    "start": 373.6,
                    "end": 373.78,
                    "word": " we",
                    "probability": 0.90625
                },
                {
                    "start": 373.78,
                    "end": 373.9,
                    "word": " have",
                    "probability": 0.88330078125
                },
                {
                    "start": 373.9,
                    "end": 374.0,
                    "word": " to",
                    "probability": 0.9091796875
                },
                {
                    "start": 374.0,
                    "end": 374.12,
                    "word": " move",
                    "probability": 0.86572265625
                },
                {
                    "start": 374.12,
                    "end": 374.34,
                    "word": " very",
                    "probability": 0.72265625
                },
                {
                    "start": 374.34,
                    "end": 374.68,
                    "word": " quickly",
                    "probability": 0.87255859375
                },
                {
                    "start": 374.68,
                    "end": 374.88,
                    "word": " and",
                    "probability": 0.63818359375
                },
                {
                    "start": 374.88,
                    "end": 375.2,
                    "word": " sometimes",
                    "probability": 0.91357421875
                },
                {
                    "start": 375.2,
                    "end": 375.38,
                    "word": " with",
                    "probability": 0.7880859375
                },
                {
                    "start": 375.38,
                    "end": 375.86,
                    "word": " whitelisting",
                    "probability": 0.85125732421875
                },
                {
                    "start": 375.86,
                    "end": 375.98,
                    "word": " and",
                    "probability": 0.89453125
                },
                {
                    "start": 375.98,
                    "end": 376.18,
                    "word": " stuff",
                    "probability": 0.80029296875
                },
                {
                    "start": 376.18,
                    "end": 376.66,
                    "word": ".",
                    "probability": 0.59912109375
                }
            ]
        },
        {
            "id": 42,
            "seek": 24970,
            "start": 378.38,
            "end": 384.52,
            "text": " And, you know, well, and sometimes we only get like a label when our customer complains and we'd rather",
            "tokens": [
                51116,
                400,
                11,
                291,
                458,
                11,
                731,
                11,
                293,
                2171,
                321,
                787,
                483,
                411,
                257,
                7645,
                562,
                527,
                5474,
                1209,
                2315,
                293,
                321,
                1116,
                2831,
                51436
            ],
            "temperature": 0.0,
            "avg_logprob": -0.2334374986588955,
            "compression_ratio": 1.5833333333333333,
            "no_speech_prob": 0.0003859996795654297,
            "words": [
                {
                    "start": 378.38,
                    "end": 378.38,
                    "word": " And",
                    "probability": 0.79150390625
                },
                {
                    "start": 378.38,
                    "end": 378.78,
                    "word": ",",
                    "probability": 0.4453125
                },
                {
                    "start": 378.78,
                    "end": 378.96,
                    "word": " you",
                    "probability": 0.66162109375
                },
                {
                    "start": 378.96,
                    "end": 379.2,
                    "word": " know",
                    "probability": 0.81787109375
                },
                {
                    "start": 379.2,
                    "end": 379.28,
                    "word": ",",
                    "probability": 0.95263671875
                },
                {
                    "start": 379.28,
                    "end": 379.58,
                    "word": " well",
                    "probability": 0.740234375
                },
                {
                    "start": 379.58,
                    "end": 379.86,
                    "word": ",",
                    "probability": 0.84228515625
                },
                {
                    "start": 379.86,
                    "end": 380.1,
                    "word": " and",
                    "probability": 0.4501953125
                },
                {
                    "start": 380.1,
                    "end": 380.68,
                    "word": " sometimes",
                    "probability": 0.90771484375
                },
                {
                    "start": 380.68,
                    "end": 381.3,
                    "word": " we",
                    "probability": 0.86083984375
                },
                {
                    "start": 381.3,
                    "end": 381.52,
                    "word": " only",
                    "probability": 0.8642578125
                },
                {
                    "start": 381.52,
                    "end": 381.82,
                    "word": " get",
                    "probability": 0.8984375
                },
                {
                    "start": 381.82,
                    "end": 382.04,
                    "word": " like",
                    "probability": 0.763671875
                },
                {
                    "start": 382.36,
                    "end": 382.6,
                    "word": " a label",
                    "probability": 0.81884765625
                },
                {
                    "start": 382.6,
                    "end": 382.8,
                    "word": " when",
                    "probability": 0.83544921875
                },
                {
                    "start": 382.8,
                    "end": 382.96,
                    "word": " our",
                    "probability": 0.7958984375
                },
                {
                    "start": 382.96,
                    "end": 383.28,
                    "word": " customer",
                    "probability": 0.6796875
                },
                {
                    "start": 383.28,
                    "end": 383.82,
                    "word": " complains",
                    "probability": 0.912841796875
                },
                {
                    "start": 383.82,
                    "end": 383.94,
                    "word": " and",
                    "probability": 0.60986328125
                },
                {
                    "start": 383.94,
                    "end": 384.18,
                    "word": " we'd",
                    "probability": 0.888427734375
                },
                {
                    "start": 384.18,
                    "end": 384.52,
                    "word": " rather",
                    "probability": 0.81689453125
                }
            ]
        },
        {
            "id": 43,
            "seek": 24970,
            "start": 385.16,
            "end": 389.92,
            "text": " get ahead of that and figure out like, hey, where's the drift happening? What's going on?",
            "tokens": [
                51472,
                483,
                2286,
                295,
                300,
                293,
                2573,
                484,
                411,
                11,
                4177,
                11,
                689,
                311,
                264,
                19699,
                2737,
                30,
                708,
                311,
                516,
                322,
                30,
                51680
            ],
            "temperature": 0.0,
            "avg_logprob": -0.2334374986588955,
            "compression_ratio": 1.5833333333333333,
            "no_speech_prob": 0.0003859996795654297,
            "words": [
                {
                    "start": 385.16,
                    "end": 385.46,
                    "word": " get",
                    "probability": 0.88525390625
                },
                {
                    "start": 385.46,
                    "end": 385.72,
                    "word": " ahead",
                    "probability": 0.8359375
                },
                {
                    "start": 385.72,
                    "end": 385.88,
                    "word": " of",
                    "probability": 0.90966796875
                },
                {
                    "start": 385.88,
                    "end": 386.12,
                    "word": " that",
                    "probability": 0.8818359375
                },
                {
                    "start": 386.12,
                    "end": 386.28,
                    "word": " and",
                    "probability": 0.88232421875
                },
                {
                    "start": 386.28,
                    "end": 386.46,
                    "word": " figure",
                    "probability": 0.92138671875
                },
                {
                    "start": 386.46,
                    "end": 386.64,
                    "word": " out",
                    "probability": 0.8017578125
                },
                {
                    "start": 386.64,
                    "end": 386.82,
                    "word": " like",
                    "probability": 0.81689453125
                },
                {
                    "start": 386.82,
                    "end": 386.92,
                    "word": ",",
                    "probability": 0.93603515625
                },
                {
                    "start": 386.92,
                    "end": 387.08,
                    "word": " hey",
                    "probability": 0.52099609375
                },
                {
                    "start": 387.08,
                    "end": 387.14,
                    "word": ",",
                    "probability": 0.87890625
                },
                {
                    "start": 387.14,
                    "end": 387.42,
                    "word": " where's",
                    "probability": 0.9375
                },
                {
                    "start": 387.42,
                    "end": 387.54,
                    "word": " the",
                    "probability": 0.81591796875
                },
                {
                    "start": 387.54,
                    "end": 387.74,
                    "word": " drift",
                    "probability": 0.86279296875
                },
                {
                    "start": 387.74,
                    "end": 388.14,
                    "word": " happening",
                    "probability": 0.716796875
                },
                {
                    "start": 388.14,
                    "end": 388.62,
                    "word": "?",
                    "probability": 0.7177734375
                },
                {
                    "start": 388.68,
                    "end": 389.0,
                    "word": " What's",
                    "probability": 0.90380859375
                },
                {
                    "start": 389.0,
                    "end": 389.16,
                    "word": " going",
                    "probability": 0.8544921875
                },
                {
                    "start": 389.16,
                    "end": 389.44,
                    "word": " on",
                    "probability": 0.91162109375
                },
                {
                    "start": 389.44,
                    "end": 389.92,
                    "word": "?",
                    "probability": 0.8837890625
                }
            ]
        },
        {
            "id": 44,
            "seek": 27580,
            "start": 389.92,
            "end": 394.9,
            "text": " And then additionally, aside from the fact that things are just changing all the time,",
            "tokens": [
                50368,
                400,
                550,
                43181,
                11,
                7359,
                490,
                264,
                1186,
                300,
                721,
                366,
                445,
                4473,
                439,
                264,
                565,
                11,
                50612
            ],
            "temperature": 0.0,
            "avg_logprob": -0.22485118479955765,
            "compression_ratio": 1.7451737451737452,
            "no_speech_prob": 0.0128173828125,
            "words": [
                {
                    "start": 389.92,
                    "end": 390.18,
                    "word": " And",
                    "probability": 0.54638671875
                },
                {
                    "start": 390.18,
                    "end": 390.36,
                    "word": " then",
                    "probability": 0.71484375
                },
                {
                    "start": 390.36,
                    "end": 391.02,
                    "word": " additionally",
                    "probability": 0.5712890625
                },
                {
                    "start": 391.02,
                    "end": 391.32,
                    "word": ",",
                    "probability": 0.9033203125
                },
                {
                    "start": 391.32,
                    "end": 391.7,
                    "word": " aside",
                    "probability": 0.81494140625
                },
                {
                    "start": 391.7,
                    "end": 392.0,
                    "word": " from",
                    "probability": 0.81103515625
                },
                {
                    "start": 392.0,
                    "end": 392.2,
                    "word": " the",
                    "probability": 0.81005859375
                },
                {
                    "start": 392.2,
                    "end": 392.78,
                    "word": " fact",
                    "probability": 0.79931640625
                },
                {
                    "start": 392.78,
                    "end": 393.02,
                    "word": " that",
                    "probability": 0.87353515625
                },
                {
                    "start": 393.02,
                    "end": 393.26,
                    "word": " things",
                    "probability": 0.74951171875
                },
                {
                    "start": 393.26,
                    "end": 393.4,
                    "word": " are",
                    "probability": 0.89697265625
                },
                {
                    "start": 393.4,
                    "end": 393.54,
                    "word": " just",
                    "probability": 0.87841796875
                },
                {
                    "start": 393.54,
                    "end": 394.0,
                    "word": " changing",
                    "probability": 0.7705078125
                },
                {
                    "start": 394.0,
                    "end": 394.2,
                    "word": " all",
                    "probability": 0.90087890625
                },
                {
                    "start": 394.2,
                    "end": 394.3,
                    "word": " the",
                    "probability": 0.81396484375
                },
                {
                    "start": 394.3,
                    "end": 394.58,
                    "word": " time",
                    "probability": 0.806640625
                },
                {
                    "start": 394.58,
                    "end": 394.9,
                    "word": ",",
                    "probability": 0.9033203125
                }
            ]
        },
        {
            "id": 45,
            "seek": 27580,
            "start": 394.9,
            "end": 399.06,
            "text": " our models are also under attack. People are trying to bypass them by doing new and weird",
            "tokens": [
                50612,
                527,
                5245,
                366,
                611,
                833,
                2690,
                13,
                3432,
                366,
                1382,
                281,
                24996,
                552,
                538,
                884,
                777,
                293,
                3657,
                50824
            ],
            "temperature": 0.0,
            "avg_logprob": -0.22485118479955765,
            "compression_ratio": 1.7451737451737452,
            "no_speech_prob": 0.0128173828125,
            "words": [
                {
                    "start": 394.9,
                    "end": 395.14,
                    "word": " our",
                    "probability": 0.80810546875
                },
                {
                    "start": 395.14,
                    "end": 395.44,
                    "word": " models",
                    "probability": 0.80322265625
                },
                {
                    "start": 395.44,
                    "end": 395.62,
                    "word": " are",
                    "probability": 0.87841796875
                },
                {
                    "start": 395.62,
                    "end": 395.8,
                    "word": " also",
                    "probability": 0.82666015625
                },
                {
                    "start": 395.8,
                    "end": 395.98,
                    "word": " under",
                    "probability": 0.7265625
                },
                {
                    "start": 395.98,
                    "end": 396.26,
                    "word": " attack",
                    "probability": 0.8544921875
                },
                {
                    "start": 396.26,
                    "end": 396.4,
                    "word": ".",
                    "probability": 0.931640625
                },
                {
                    "start": 396.4,
                    "end": 396.56,
                    "word": " People",
                    "probability": 0.9189453125
                },
                {
                    "start": 396.56,
                    "end": 396.72,
                    "word": " are",
                    "probability": 0.88916015625
                },
                {
                    "start": 396.72,
                    "end": 396.94,
                    "word": " trying",
                    "probability": 0.80224609375
                },
                {
                    "start": 396.94,
                    "end": 397.14,
                    "word": " to",
                    "probability": 0.90380859375
                },
                {
                    "start": 397.14,
                    "end": 397.48,
                    "word": " bypass",
                    "probability": 0.841796875
                },
                {
                    "start": 397.48,
                    "end": 397.92,
                    "word": " them",
                    "probability": 0.73193359375
                },
                {
                    "start": 397.92,
                    "end": 398.16,
                    "word": " by",
                    "probability": 0.87548828125
                },
                {
                    "start": 398.16,
                    "end": 398.42,
                    "word": " doing",
                    "probability": 0.89990234375
                },
                {
                    "start": 398.42,
                    "end": 398.74,
                    "word": " new",
                    "probability": 0.81787109375
                },
                {
                    "start": 398.74,
                    "end": 398.88,
                    "word": " and",
                    "probability": 0.89697265625
                },
                {
                    "start": 398.88,
                    "end": 399.06,
                    "word": " weird",
                    "probability": 0.8564453125
                }
            ]
        },
        {
            "id": 46,
            "seek": 27580,
            "start": 399.06,
            "end": 406.58,
            "text": " things to their data, like packing a spam filter and seeing if that works, or like posting a",
            "tokens": [
                50824,
                721,
                281,
                641,
                1412,
                11,
                411,
                20815,
                257,
                24028,
                6608,
                293,
                2577,
                498,
                300,
                1985,
                11,
                420,
                411,
                15978,
                257,
                51180
            ],
            "temperature": 0.0,
            "avg_logprob": -0.22485118479955765,
            "compression_ratio": 1.7451737451737452,
            "no_speech_prob": 0.0128173828125,
            "words": [
                {
                    "start": 399.06,
                    "end": 399.44,
                    "word": " things",
                    "probability": 0.74365234375
                },
                {
                    "start": 399.44,
                    "end": 399.58,
                    "word": " to",
                    "probability": 0.900390625
                },
                {
                    "start": 399.58,
                    "end": 399.72,
                    "word": " their",
                    "probability": 0.6298828125
                },
                {
                    "start": 399.72,
                    "end": 399.98,
                    "word": " data",
                    "probability": 0.806640625
                },
                {
                    "start": 399.98,
                    "end": 400.36,
                    "word": ",",
                    "probability": 0.642578125
                },
                {
                    "start": 400.36,
                    "end": 401.08,
                    "word": " like",
                    "probability": 0.86572265625
                },
                {
                    "start": 401.08,
                    "end": 402.0,
                    "word": " packing",
                    "probability": 0.63818359375
                },
                {
                    "start": 402.22,
                    "end": 402.48,
                    "word": " a spam",
                    "probability": 0.931640625
                },
                {
                    "start": 402.48,
                    "end": 402.88,
                    "word": " filter",
                    "probability": 0.900390625
                },
                {
                    "start": 402.88,
                    "end": 403.76,
                    "word": " and",
                    "probability": 0.53271484375
                },
                {
                    "start": 403.76,
                    "end": 404.38,
                    "word": " seeing",
                    "probability": 0.76220703125
                },
                {
                    "start": 404.38,
                    "end": 404.52,
                    "word": " if",
                    "probability": 0.8984375
                },
                {
                    "start": 404.52,
                    "end": 404.7,
                    "word": " that",
                    "probability": 0.88525390625
                },
                {
                    "start": 404.7,
                    "end": 405.0,
                    "word": " works",
                    "probability": 0.72216796875
                },
                {
                    "start": 405.0,
                    "end": 405.08,
                    "word": ",",
                    "probability": 0.58740234375
                },
                {
                    "start": 405.08,
                    "end": 405.22,
                    "word": " or",
                    "probability": 0.90576171875
                },
                {
                    "start": 405.22,
                    "end": 405.46,
                    "word": " like",
                    "probability": 0.44921875
                },
                {
                    "start": 405.46,
                    "end": 406.18,
                    "word": " posting",
                    "probability": 0.72412109375
                },
                {
                    "start": 406.34,
                    "end": 406.58,
                    "word": " a different",
                    "probability": 0.69287109375
                }
            ]
        },
        {
            "id": 47,
            "seek": 27580,
            "start": 406.58,
            "end": 411.1,
            "text": " different pattern to see if that works, to see if they can get past like the Facebook spam filter.",
            "tokens": [
                51180,
                819,
                5102,
                281,
                536,
                498,
                300,
                1985,
                11,
                281,
                536,
                498,
                436,
                393,
                483,
                1791,
                411,
                264,
                4384,
                24028,
                6608,
                13,
                51408
            ],
            "temperature": 0.0,
            "avg_logprob": -0.22485118479955765,
            "compression_ratio": 1.7451737451737452,
            "no_speech_prob": 0.0128173828125,
            "words": [
                {
                    "start": 406.58,
                    "end": 407.04,
                    "word": " pattern",
                    "probability": 0.73681640625
                },
                {
                    "start": 407.04,
                    "end": 407.26,
                    "word": " to",
                    "probability": 0.7578125
                },
                {
                    "start": 407.26,
                    "end": 407.68,
                    "word": " see",
                    "probability": 0.8134765625
                },
                {
                    "start": 407.68,
                    "end": 407.84,
                    "word": " if",
                    "probability": 0.89599609375
                },
                {
                    "start": 407.84,
                    "end": 407.98,
                    "word": " that",
                    "probability": 0.79638671875
                },
                {
                    "start": 407.98,
                    "end": 408.32,
                    "word": " works",
                    "probability": 0.7353515625
                },
                {
                    "start": 408.32,
                    "end": 408.38,
                    "word": ",",
                    "probability": 0.61279296875
                },
                {
                    "start": 408.38,
                    "end": 408.48,
                    "word": " to",
                    "probability": 0.90576171875
                },
                {
                    "start": 408.48,
                    "end": 408.6,
                    "word": " see",
                    "probability": 0.806640625
                },
                {
                    "start": 408.6,
                    "end": 408.72,
                    "word": " if",
                    "probability": 0.8984375
                },
                {
                    "start": 408.72,
                    "end": 408.8,
                    "word": " they",
                    "probability": 0.73486328125
                },
                {
                    "start": 408.8,
                    "end": 408.92,
                    "word": " can",
                    "probability": 0.89892578125
                },
                {
                    "start": 408.92,
                    "end": 409.08,
                    "word": " get",
                    "probability": 0.88916015625
                },
                {
                    "start": 409.08,
                    "end": 409.38,
                    "word": " past",
                    "probability": 0.78369140625
                },
                {
                    "start": 409.38,
                    "end": 409.56,
                    "word": " like",
                    "probability": 0.408935546875
                },
                {
                    "start": 409.56,
                    "end": 409.68,
                    "word": " the",
                    "probability": 0.7919921875
                },
                {
                    "start": 409.68,
                    "end": 409.92,
                    "word": " Facebook",
                    "probability": 0.7939453125
                },
                {
                    "start": 409.92,
                    "end": 410.2,
                    "word": " spam",
                    "probability": 0.88623046875
                },
                {
                    "start": 410.2,
                    "end": 410.58,
                    "word": " filter",
                    "probability": 0.9482421875
                },
                {
                    "start": 410.58,
                    "end": 410.92,
                    "word": ".",
                    "probability": 0.90087890625
                },
                {
                    "start": 410.92,
                    "end": 411.1,
                    "word": " They're",
                    "probability": 0.67138671875
                }
            ]
        },
        {
            "id": 48,
            "seek": 27580,
            "start": 411.1,
            "end": 416.02,
            "text": " They're constantly innovating with a specific goal in mind of bypassing our models.",
            "tokens": [
                51408,
                814,
                434,
                6460,
                5083,
                990,
                365,
                257,
                2685,
                3387,
                294,
                1575,
                295,
                24996,
                278,
                527,
                5245,
                13,
                51652
            ],
            "temperature": 0.0,
            "avg_logprob": -0.22485118479955765,
            "compression_ratio": 1.7451737451737452,
            "no_speech_prob": 0.0128173828125,
            "words": [
                {
                    "start": 411.1,
                    "end": 411.5,
                    "word": " constantly",
                    "probability": 0.8818359375
                },
                {
                    "start": 411.5,
                    "end": 412.12,
                    "word": " innovating",
                    "probability": 0.902099609375
                },
                {
                    "start": 412.12,
                    "end": 412.64,
                    "word": " with",
                    "probability": 0.77734375
                },
                {
                    "start": 412.78,
                    "end": 413.1,
                    "word": " a specific",
                    "probability": 0.77783203125
                },
                {
                    "start": 413.1,
                    "end": 413.4,
                    "word": " goal",
                    "probability": 0.83740234375
                },
                {
                    "start": 413.4,
                    "end": 413.54,
                    "word": " in",
                    "probability": 0.8740234375
                },
                {
                    "start": 413.54,
                    "end": 413.74,
                    "word": " mind",
                    "probability": 0.81298828125
                },
                {
                    "start": 413.74,
                    "end": 413.9,
                    "word": " of",
                    "probability": 0.89111328125
                },
                {
                    "start": 413.9,
                    "end": 414.56,
                    "word": " bypassing",
                    "probability": 0.897705078125
                },
                {
                    "start": 414.56,
                    "end": 414.72,
                    "word": " our",
                    "probability": 0.8251953125
                },
                {
                    "start": 414.72,
                    "end": 415.54,
                    "word": " models",
                    "probability": 0.373291015625
                },
                {
                    "start": 415.54,
                    "end": 416.02,
                    "word": ".",
                    "probability": 0.99560546875
                }
            ]
        },
        {
            "id": 49,
            "seek": 30180,
            "start": 416.54,
            "end": 422.14,
            "text": " This may not result be an adversarial example per se from the literature, but it sure",
            "tokens": [
                50400,
                639,
                815,
                406,
                1874,
                312,
                364,
                17641,
                44745,
                1365,
                680,
                369,
                490,
                264,
                10394,
                11,
                457,
                309,
                988,
                50688
            ],
            "temperature": 0.0,
            "avg_logprob": -0.2284059002158347,
            "compression_ratio": 1.555084745762712,
            "no_speech_prob": 0.0078125,
            "words": [
                {
                    "start": 416.54,
                    "end": 417.02,
                    "word": " This",
                    "probability": 0.482177734375
                },
                {
                    "start": 417.02,
                    "end": 417.2,
                    "word": " may",
                    "probability": 0.8818359375
                },
                {
                    "start": 417.2,
                    "end": 417.4,
                    "word": " not",
                    "probability": 0.89453125
                },
                {
                    "start": 417.4,
                    "end": 417.9,
                    "word": " result",
                    "probability": 0.5673828125
                },
                {
                    "start": 417.9,
                    "end": 418.76,
                    "word": " be",
                    "probability": 0.36328125
                },
                {
                    "start": 418.76,
                    "end": 419.2,
                    "word": " an",
                    "probability": 0.83935546875
                },
                {
                    "start": 419.2,
                    "end": 419.62,
                    "word": " adversarial",
                    "probability": 0.9169921875
                },
                {
                    "start": 419.62,
                    "end": 420.12,
                    "word": " example",
                    "probability": 0.9169921875
                },
                {
                    "start": 420.12,
                    "end": 420.4,
                    "word": " per",
                    "probability": 0.71826171875
                },
                {
                    "start": 420.4,
                    "end": 420.58,
                    "word": " se",
                    "probability": 0.80126953125
                },
                {
                    "start": 420.58,
                    "end": 420.84,
                    "word": " from",
                    "probability": 0.69482421875
                },
                {
                    "start": 420.84,
                    "end": 420.98,
                    "word": " the",
                    "probability": 0.7578125
                },
                {
                    "start": 420.98,
                    "end": 421.26,
                    "word": " literature",
                    "probability": 0.849609375
                },
                {
                    "start": 421.26,
                    "end": 421.52,
                    "word": ",",
                    "probability": 0.6015625
                },
                {
                    "start": 421.52,
                    "end": 421.64,
                    "word": " but",
                    "probability": 0.8935546875
                },
                {
                    "start": 421.64,
                    "end": 421.84,
                    "word": " it",
                    "probability": 0.84814453125
                },
                {
                    "start": 421.84,
                    "end": 422.14,
                    "word": " sure",
                    "probability": 0.70849609375
                }
            ]
        },
        {
            "id": 50,
            "seek": 30180,
            "start": 422.14,
            "end": 428.5,
            "text": " kind of acts like that sometimes. And additionally, we could be under attack via a poisoning system.",
            "tokens": [
                50688,
                733,
                295,
                10672,
                411,
                300,
                2171,
                13,
                400,
                43181,
                11,
                321,
                727,
                312,
                833,
                2690,
                5766,
                257,
                36778,
                1185,
                13,
                50964
            ],
            "temperature": 0.0,
            "avg_logprob": -0.2284059002158347,
            "compression_ratio": 1.555084745762712,
            "no_speech_prob": 0.0078125,
            "words": [
                {
                    "start": 422.14,
                    "end": 422.78,
                    "word": " kind",
                    "probability": 0.77490234375
                },
                {
                    "start": 422.78,
                    "end": 422.86,
                    "word": " of",
                    "probability": 0.90869140625
                },
                {
                    "start": 422.86,
                    "end": 423.12,
                    "word": " acts",
                    "probability": 0.75244140625
                },
                {
                    "start": 423.12,
                    "end": 423.32,
                    "word": " like",
                    "probability": 0.87646484375
                },
                {
                    "start": 423.32,
                    "end": 423.46,
                    "word": " that",
                    "probability": 0.8427734375
                },
                {
                    "start": 423.46,
                    "end": 423.82,
                    "word": " sometimes",
                    "probability": 0.8837890625
                },
                {
                    "start": 423.82,
                    "end": 424.3,
                    "word": ".",
                    "probability": 0.7998046875
                },
                {
                    "start": 424.4,
                    "end": 424.66,
                    "word": " And",
                    "probability": 0.73095703125
                },
                {
                    "start": 424.66,
                    "end": 425.64,
                    "word": " additionally",
                    "probability": 0.69580078125
                },
                {
                    "start": 425.64,
                    "end": 425.7,
                    "word": ",",
                    "probability": 0.7900390625
                },
                {
                    "start": 425.7,
                    "end": 425.94,
                    "word": " we",
                    "probability": 0.8974609375
                },
                {
                    "start": 425.94,
                    "end": 426.12,
                    "word": " could",
                    "probability": 0.82958984375
                },
                {
                    "start": 426.12,
                    "end": 426.32,
                    "word": " be",
                    "probability": 0.9091796875
                },
                {
                    "start": 426.32,
                    "end": 426.58,
                    "word": " under",
                    "probability": 0.84375
                },
                {
                    "start": 426.58,
                    "end": 426.9,
                    "word": " attack",
                    "probability": 0.7705078125
                },
                {
                    "start": 426.9,
                    "end": 427.14,
                    "word": " via",
                    "probability": 0.498046875
                },
                {
                    "start": 427.32,
                    "end": 427.6,
                    "word": " a poisoning",
                    "probability": 0.623046875
                },
                {
                    "start": 427.6,
                    "end": 428.02,
                    "word": " system",
                    "probability": 0.86865234375
                },
                {
                    "start": 428.02,
                    "end": 428.5,
                    "word": ".",
                    "probability": 0.974609375
                }
            ]
        },
        {
            "id": 51,
            "seek": 30180,
            "start": 429.42,
            "end": 434.76,
            "text": " Maybe if the detector is fast enough, we can get ahead of those things beforehand.",
            "tokens": [
                51032,
                2704,
                498,
                264,
                25712,
                307,
                2370,
                1547,
                11,
                321,
                393,
                483,
                2286,
                295,
                729,
                721,
                22893,
                13,
                51284
            ],
            "temperature": 0.0,
            "avg_logprob": -0.2284059002158347,
            "compression_ratio": 1.555084745762712,
            "no_speech_prob": 0.0078125,
            "words": [
                {
                    "start": 429.42,
                    "end": 429.9,
                    "word": " Maybe",
                    "probability": 0.92333984375
                },
                {
                    "start": 429.9,
                    "end": 430.38,
                    "word": " if",
                    "probability": 0.7705078125
                },
                {
                    "start": 430.38,
                    "end": 430.62,
                    "word": " the",
                    "probability": 0.486572265625
                },
                {
                    "start": 430.62,
                    "end": 431.78,
                    "word": " detector",
                    "probability": 0.75
                },
                {
                    "start": 431.78,
                    "end": 431.92,
                    "word": " is",
                    "probability": 0.85986328125
                },
                {
                    "start": 431.92,
                    "end": 432.18,
                    "word": " fast",
                    "probability": 0.82763671875
                },
                {
                    "start": 432.18,
                    "end": 432.36,
                    "word": " enough",
                    "probability": 0.7900390625
                },
                {
                    "start": 432.36,
                    "end": 432.42,
                    "word": ",",
                    "probability": 0.8193359375
                },
                {
                    "start": 432.42,
                    "end": 432.52,
                    "word": " we",
                    "probability": 0.8974609375
                },
                {
                    "start": 432.52,
                    "end": 432.68,
                    "word": " can",
                    "probability": 0.609375
                },
                {
                    "start": 432.68,
                    "end": 433.14,
                    "word": " get",
                    "probability": 0.88916015625
                },
                {
                    "start": 433.14,
                    "end": 433.36,
                    "word": " ahead",
                    "probability": 0.83642578125
                },
                {
                    "start": 433.36,
                    "end": 433.46,
                    "word": " of",
                    "probability": 0.90673828125
                },
                {
                    "start": 433.46,
                    "end": 433.62,
                    "word": " those",
                    "probability": 0.82421875
                },
                {
                    "start": 433.62,
                    "end": 433.84,
                    "word": " things",
                    "probability": 0.69580078125
                },
                {
                    "start": 433.84,
                    "end": 434.28,
                    "word": " beforehand",
                    "probability": 0.7744140625
                },
                {
                    "start": 434.28,
                    "end": 434.76,
                    "word": ".",
                    "probability": 0.994140625
                }
            ]
        },
        {
            "id": 52,
            "seek": 30180,
            "start": 435.64,
            "end": 442.02,
            "text": " So here's a trivial example. So we have these two data sets. One is our training data set. That's",
            "tokens": [
                51356,
                407,
                510,
                311,
                257,
                26703,
                1365,
                13,
                407,
                321,
                362,
                613,
                732,
                1412,
                6352,
                13,
                1485,
                307,
                527,
                3097,
                1412,
                992,
                13,
                663,
                311,
                51660
            ],
            "temperature": 0.0,
            "avg_logprob": -0.2284059002158347,
            "compression_ratio": 1.555084745762712,
            "no_speech_prob": 0.0078125,
            "words": [
                {
                    "start": 435.64,
                    "end": 436.12,
                    "word": " So",
                    "probability": 0.873046875
                },
                {
                    "start": 436.12,
                    "end": 436.46,
                    "word": " here's",
                    "probability": 0.86767578125
                },
                {
                    "start": 436.64,
                    "end": 437.08,
                    "word": " a trivial",
                    "probability": 0.89892578125
                },
                {
                    "start": 437.08,
                    "end": 437.56,
                    "word": " example",
                    "probability": 0.927734375
                },
                {
                    "start": 437.56,
                    "end": 438.04,
                    "word": ".",
                    "probability": 0.9755859375
                },
                {
                    "start": 438.68,
                    "end": 439.16,
                    "word": " So",
                    "probability": 0.8349609375
                },
                {
                    "start": 439.16,
                    "end": 439.32,
                    "word": " we",
                    "probability": 0.88525390625
                },
                {
                    "start": 439.32,
                    "end": 439.5,
                    "word": " have",
                    "probability": 0.8798828125
                },
                {
                    "start": 439.5,
                    "end": 439.68,
                    "word": " these",
                    "probability": 0.70947265625
                },
                {
                    "start": 439.68,
                    "end": 439.88,
                    "word": " two",
                    "probability": 0.8046875
                },
                {
                    "start": 439.88,
                    "end": 440.1,
                    "word": " data",
                    "probability": 0.48974609375
                },
                {
                    "start": 440.1,
                    "end": 440.34,
                    "word": " sets",
                    "probability": 0.81494140625
                },
                {
                    "start": 440.34,
                    "end": 440.46,
                    "word": ".",
                    "probability": 0.6708984375
                },
                {
                    "start": 440.46,
                    "end": 440.66,
                    "word": " One",
                    "probability": 0.6455078125
                },
                {
                    "start": 440.66,
                    "end": 440.82,
                    "word": " is",
                    "probability": 0.908203125
                },
                {
                    "start": 440.82,
                    "end": 440.96,
                    "word": " our",
                    "probability": 0.8017578125
                },
                {
                    "start": 440.96,
                    "end": 441.26,
                    "word": " training",
                    "probability": 0.7880859375
                },
                {
                    "start": 441.26,
                    "end": 441.48,
                    "word": " data",
                    "probability": 0.7431640625
                },
                {
                    "start": 441.48,
                    "end": 441.66,
                    "word": " set",
                    "probability": 0.86474609375
                },
                {
                    "start": 441.66,
                    "end": 441.76,
                    "word": ".",
                    "probability": 0.748046875
                },
                {
                    "start": 441.76,
                    "end": 442.02,
                    "word": " That's",
                    "probability": 0.78955078125
                }
            ]
        },
        {
            "id": 53,
            "seek": 32874,
            "start": 442.02,
            "end": 449.5,
            "text": " we've got 1, 1, 2, 1, 1. And so our training data set has 70% 1s, 20% 2s, and 10% 3s.",
            "tokens": [
                50376,
                321,
                600,
                658,
                502,
                11,
                502,
                11,
                568,
                11,
                502,
                11,
                502,
                13,
                400,
                370,
                527,
                3097,
                1412,
                992,
                575,
                5285,
                4,
                502,
                82,
                11,
                945,
                4,
                568,
                82,
                11,
                293,
                1266,
                4,
                805,
                82,
                13,
                50712
            ],
            "temperature": 0.0,
            "avg_logprob": -0.19417735348399887,
            "compression_ratio": 1.660633484162896,
            "no_speech_prob": 0.0279998779296875,
            "words": [
                {
                    "start": 442.02,
                    "end": 442.54,
                    "word": " we've",
                    "probability": 0.15282440185546875
                },
                {
                    "start": 442.54,
                    "end": 442.7,
                    "word": " got",
                    "probability": 0.900390625
                },
                {
                    "start": 442.7,
                    "end": 443.24,
                    "word": " 1",
                    "probability": 0.58349609375
                },
                {
                    "start": 443.24,
                    "end": 443.34,
                    "word": ",",
                    "probability": 0.76025390625
                },
                {
                    "start": 443.34,
                    "end": 443.54,
                    "word": " 1",
                    "probability": 0.853515625
                },
                {
                    "start": 443.54,
                    "end": 443.66,
                    "word": ",",
                    "probability": 0.9677734375
                },
                {
                    "start": 443.66,
                    "end": 443.9,
                    "word": " 2",
                    "probability": 0.98388671875
                },
                {
                    "start": 443.9,
                    "end": 443.98,
                    "word": ",",
                    "probability": 0.95947265625
                },
                {
                    "start": 443.98,
                    "end": 444.18,
                    "word": " 1",
                    "probability": 0.9619140625
                },
                {
                    "start": 444.18,
                    "end": 444.24,
                    "word": ",",
                    "probability": 0.888671875
                },
                {
                    "start": 444.24,
                    "end": 444.38,
                    "word": " 1",
                    "probability": 0.94970703125
                },
                {
                    "start": 444.38,
                    "end": 444.54,
                    "word": ".",
                    "probability": 0.5927734375
                },
                {
                    "start": 444.54,
                    "end": 444.62,
                    "word": " And",
                    "probability": 0.66943359375
                },
                {
                    "start": 444.62,
                    "end": 444.82,
                    "word": " so",
                    "probability": 0.89013671875
                },
                {
                    "start": 444.82,
                    "end": 445.22,
                    "word": " our",
                    "probability": 0.68017578125
                },
                {
                    "start": 445.22,
                    "end": 445.5,
                    "word": " training",
                    "probability": 0.81884765625
                },
                {
                    "start": 445.5,
                    "end": 445.68,
                    "word": " data",
                    "probability": 0.482666015625
                },
                {
                    "start": 445.68,
                    "end": 445.84,
                    "word": " set",
                    "probability": 0.8154296875
                },
                {
                    "start": 445.84,
                    "end": 446.0,
                    "word": " has",
                    "probability": 0.88720703125
                },
                {
                    "start": 446.0,
                    "end": 446.36,
                    "word": " 70",
                    "probability": 0.869140625
                },
                {
                    "start": 446.36,
                    "end": 446.74,
                    "word": "%",
                    "probability": 0.83642578125
                },
                {
                    "start": 446.74,
                    "end": 447.18,
                    "word": " 1s",
                    "probability": 0.740478515625
                },
                {
                    "start": 447.18,
                    "end": 447.24,
                    "word": ",",
                    "probability": 0.947265625
                },
                {
                    "start": 447.24,
                    "end": 447.46,
                    "word": " 20",
                    "probability": 0.90673828125
                },
                {
                    "start": 447.46,
                    "end": 447.74,
                    "word": "%",
                    "probability": 0.99169921875
                },
                {
                    "start": 447.74,
                    "end": 448.06,
                    "word": " 2s",
                    "probability": 0.9951171875
                },
                {
                    "start": 448.06,
                    "end": 448.08,
                    "word": ",",
                    "probability": 0.63818359375
                },
                {
                    "start": 448.08,
                    "end": 448.18,
                    "word": " and",
                    "probability": 0.88134765625
                },
                {
                    "start": 448.18,
                    "end": 448.36,
                    "word": " 10",
                    "probability": 0.90771484375
                },
                {
                    "start": 448.36,
                    "end": 448.58,
                    "word": "%",
                    "probability": 0.99365234375
                },
                {
                    "start": 448.58,
                    "end": 449.1,
                    "word": " 3s",
                    "probability": 0.998046875
                },
                {
                    "start": 449.1,
                    "end": 449.5,
                    "word": ".",
                    "probability": 0.87939453125
                }
            ]
        },
        {
            "id": 54,
            "seek": 32874,
            "start": 449.9,
            "end": 455.42,
            "text": " And this is what we trained on. Realistically, there's no point in training a machine learning",
            "tokens": [
                50764,
                400,
                341,
                307,
                437,
                321,
                8895,
                322,
                13,
                8467,
                20458,
                11,
                456,
                311,
                572,
                935,
                294,
                3097,
                257,
                3479,
                2539,
                51032
            ],
            "temperature": 0.0,
            "avg_logprob": -0.19417735348399887,
            "compression_ratio": 1.660633484162896,
            "no_speech_prob": 0.0279998779296875,
            "words": [
                {
                    "start": 449.9,
                    "end": 450.3,
                    "word": " And",
                    "probability": 0.8056640625
                },
                {
                    "start": 450.3,
                    "end": 450.52,
                    "word": " this",
                    "probability": 0.87255859375
                },
                {
                    "start": 450.52,
                    "end": 450.98,
                    "word": " is",
                    "probability": 0.751953125
                },
                {
                    "start": 450.98,
                    "end": 451.14,
                    "word": " what",
                    "probability": 0.88232421875
                },
                {
                    "start": 451.14,
                    "end": 451.26,
                    "word": " we",
                    "probability": 0.90576171875
                },
                {
                    "start": 451.26,
                    "end": 451.56,
                    "word": " trained",
                    "probability": 0.71337890625
                },
                {
                    "start": 451.56,
                    "end": 451.74,
                    "word": " on",
                    "probability": 0.90380859375
                },
                {
                    "start": 451.74,
                    "end": 452.14,
                    "word": ".",
                    "probability": 0.96826171875
                },
                {
                    "start": 452.68,
                    "end": 453.08,
                    "word": " Realistically",
                    "probability": 0.9482421875
                },
                {
                    "start": 453.08,
                    "end": 453.32,
                    "word": ",",
                    "probability": 0.83154296875
                },
                {
                    "start": 453.32,
                    "end": 453.74,
                    "word": " there's",
                    "probability": 0.90673828125
                },
                {
                    "start": 453.74,
                    "end": 453.88,
                    "word": " no",
                    "probability": 0.9072265625
                },
                {
                    "start": 453.88,
                    "end": 454.14,
                    "word": " point",
                    "probability": 0.849609375
                },
                {
                    "start": 454.14,
                    "end": 454.26,
                    "word": " in",
                    "probability": 0.89404296875
                },
                {
                    "start": 454.26,
                    "end": 454.46,
                    "word": " training",
                    "probability": 0.7861328125
                },
                {
                    "start": 454.64,
                    "end": 455.18,
                    "word": " a machine",
                    "probability": 0.6796875
                },
                {
                    "start": 455.18,
                    "end": 455.42,
                    "word": " learning",
                    "probability": 0.8203125
                }
            ]
        },
        {
            "id": 55,
            "seek": 32874,
            "start": 455.42,
            "end": 463.14,
            "text": " model on a sequence of numbers, but suppose we do. But then we deploy this model, and we get",
            "tokens": [
                51032,
                2316,
                322,
                257,
                8310,
                295,
                3547,
                11,
                457,
                7297,
                321,
                360,
                13,
                583,
                550,
                321,
                7274,
                341,
                2316,
                11,
                293,
                321,
                483,
                51412
            ],
            "temperature": 0.0,
            "avg_logprob": -0.19417735348399887,
            "compression_ratio": 1.660633484162896,
            "no_speech_prob": 0.0279998779296875,
            "words": [
                {
                    "start": 455.42,
                    "end": 455.72,
                    "word": " model",
                    "probability": 0.8125
                },
                {
                    "start": 455.72,
                    "end": 456.32,
                    "word": " on",
                    "probability": 0.900390625
                },
                {
                    "start": 457.42,
                    "end": 457.68,
                    "word": " a sequence",
                    "probability": 0.85302734375
                },
                {
                    "start": 457.68,
                    "end": 457.82,
                    "word": " of",
                    "probability": 0.91015625
                },
                {
                    "start": 457.82,
                    "end": 458.06,
                    "word": " numbers",
                    "probability": 0.86376953125
                },
                {
                    "start": 458.06,
                    "end": 458.42,
                    "word": ",",
                    "probability": 0.6962890625
                },
                {
                    "start": 458.42,
                    "end": 458.72,
                    "word": " but",
                    "probability": 0.8935546875
                },
                {
                    "start": 458.72,
                    "end": 459.14,
                    "word": " suppose",
                    "probability": 0.8466796875
                },
                {
                    "start": 459.14,
                    "end": 459.3,
                    "word": " we",
                    "probability": 0.90625
                },
                {
                    "start": 459.3,
                    "end": 459.64,
                    "word": " do",
                    "probability": 0.8515625
                },
                {
                    "start": 459.64,
                    "end": 460.04,
                    "word": ".",
                    "probability": 0.90576171875
                },
                {
                    "start": 461.08,
                    "end": 461.48,
                    "word": " But",
                    "probability": 0.87744140625
                },
                {
                    "start": 461.48,
                    "end": 461.6,
                    "word": " then",
                    "probability": 0.73193359375
                },
                {
                    "start": 461.6,
                    "end": 461.72,
                    "word": " we",
                    "probability": 0.88427734375
                },
                {
                    "start": 461.72,
                    "end": 462.0,
                    "word": " deploy",
                    "probability": 0.73681640625
                },
                {
                    "start": 462.0,
                    "end": 462.16,
                    "word": " this",
                    "probability": 0.88427734375
                },
                {
                    "start": 462.16,
                    "end": 462.38,
                    "word": " model",
                    "probability": 0.81884765625
                },
                {
                    "start": 462.38,
                    "end": 462.62,
                    "word": ",",
                    "probability": 0.3837890625
                },
                {
                    "start": 462.62,
                    "end": 462.82,
                    "word": " and",
                    "probability": 0.8984375
                },
                {
                    "start": 462.82,
                    "end": 462.96,
                    "word": " we",
                    "probability": 0.90478515625
                },
                {
                    "start": 462.96,
                    "end": 463.14,
                    "word": " get",
                    "probability": 0.90283203125
                }
            ]
        },
        {
            "id": 56,
            "seek": 32874,
            "start": 463.14,
            "end": 468.96,
            "text": " this other sequence of data. 2, 2, 1, 2, so on, so on. And when we actually deploy it, we see",
            "tokens": [
                51412,
                341,
                661,
                8310,
                295,
                1412,
                13,
                568,
                11,
                568,
                11,
                502,
                11,
                568,
                11,
                370,
                322,
                11,
                370,
                322,
                13,
                400,
                562,
                321,
                767,
                7274,
                309,
                11,
                321,
                536,
                51724
            ],
            "temperature": 0.0,
            "avg_logprob": -0.19417735348399887,
            "compression_ratio": 1.660633484162896,
            "no_speech_prob": 0.0279998779296875,
            "words": [
                {
                    "start": 463.14,
                    "end": 463.3,
                    "word": " this",
                    "probability": 0.8779296875
                },
                {
                    "start": 463.3,
                    "end": 463.46,
                    "word": " other",
                    "probability": 0.79150390625
                },
                {
                    "start": 463.46,
                    "end": 463.78,
                    "word": " sequence",
                    "probability": 0.857421875
                },
                {
                    "start": 463.78,
                    "end": 463.9,
                    "word": " of",
                    "probability": 0.890625
                },
                {
                    "start": 463.9,
                    "end": 464.04,
                    "word": " data",
                    "probability": 0.81787109375
                },
                {
                    "start": 464.04,
                    "end": 464.22,
                    "word": ".",
                    "probability": 0.34228515625
                },
                {
                    "start": 464.22,
                    "end": 464.38,
                    "word": " 2",
                    "probability": 0.953125
                },
                {
                    "start": 464.38,
                    "end": 464.46,
                    "word": ",",
                    "probability": 0.9794921875
                },
                {
                    "start": 464.46,
                    "end": 464.62,
                    "word": " 2",
                    "probability": 0.98583984375
                },
                {
                    "start": 464.62,
                    "end": 464.7,
                    "word": ",",
                    "probability": 0.994140625
                },
                {
                    "start": 464.7,
                    "end": 464.96,
                    "word": " 1",
                    "probability": 0.99365234375
                },
                {
                    "start": 464.96,
                    "end": 465.14,
                    "word": ",",
                    "probability": 0.9873046875
                },
                {
                    "start": 465.14,
                    "end": 465.42,
                    "word": " 2",
                    "probability": 0.982421875
                },
                {
                    "start": 465.42,
                    "end": 465.56,
                    "word": ",",
                    "probability": 0.90771484375
                },
                {
                    "start": 465.56,
                    "end": 465.68,
                    "word": " so",
                    "probability": 0.74560546875
                },
                {
                    "start": 465.68,
                    "end": 465.8,
                    "word": " on",
                    "probability": 0.89453125
                },
                {
                    "start": 465.8,
                    "end": 465.8,
                    "word": ",",
                    "probability": 0.380615234375
                },
                {
                    "start": 465.8,
                    "end": 465.86,
                    "word": " so",
                    "probability": 0.84033203125
                },
                {
                    "start": 465.86,
                    "end": 466.04,
                    "word": " on",
                    "probability": 0.90380859375
                },
                {
                    "start": 466.04,
                    "end": 466.38,
                    "word": ".",
                    "probability": 0.9580078125
                },
                {
                    "start": 466.52,
                    "end": 466.92,
                    "word": " And",
                    "probability": 0.82275390625
                },
                {
                    "start": 466.92,
                    "end": 467.68,
                    "word": " when",
                    "probability": 0.80615234375
                },
                {
                    "start": 467.68,
                    "end": 467.82,
                    "word": " we",
                    "probability": 0.90380859375
                },
                {
                    "start": 467.82,
                    "end": 468.1,
                    "word": " actually",
                    "probability": 0.7587890625
                },
                {
                    "start": 468.1,
                    "end": 468.46,
                    "word": " deploy",
                    "probability": 0.75390625
                },
                {
                    "start": 468.46,
                    "end": 468.62,
                    "word": " it",
                    "probability": 0.89990234375
                },
                {
                    "start": 468.62,
                    "end": 468.72,
                    "word": ",",
                    "probability": 0.95654296875
                },
                {
                    "start": 468.72,
                    "end": 468.78,
                    "word": " we",
                    "probability": 0.89990234375
                },
                {
                    "start": 468.78,
                    "end": 468.96,
                    "word": " see",
                    "probability": 0.70458984375
                }
            ]
        },
        {
            "id": 57,
            "seek": 35738,
            "start": 468.96,
            "end": 474.42,
            "text": " data coming over the wire. We have 30% 1s, 60% 2s, and 10% 3s. So that's a little bit different",
            "tokens": [
                50376,
                1412,
                1348,
                670,
                264,
                6234,
                13,
                492,
                362,
                2217,
                4,
                502,
                82,
                11,
                4060,
                4,
                568,
                82,
                11,
                293,
                1266,
                4,
                805,
                82,
                13,
                407,
                300,
                311,
                257,
                707,
                857,
                819,
                50632
            ],
            "temperature": 0.0,
            "avg_logprob": -0.18888608594575235,
            "compression_ratio": 1.7472527472527473,
            "no_speech_prob": 0.0010480880737304688,
            "words": [
                {
                    "start": 468.96,
                    "end": 469.54,
                    "word": " data",
                    "probability": 0.220458984375
                },
                {
                    "start": 469.54,
                    "end": 469.8,
                    "word": " coming",
                    "probability": 0.81103515625
                },
                {
                    "start": 469.8,
                    "end": 469.94,
                    "word": " over",
                    "probability": 0.751953125
                },
                {
                    "start": 469.94,
                    "end": 470.08,
                    "word": " the",
                    "probability": 0.8134765625
                },
                {
                    "start": 470.08,
                    "end": 470.3,
                    "word": " wire",
                    "probability": 0.78173828125
                },
                {
                    "start": 470.3,
                    "end": 470.44,
                    "word": ".",
                    "probability": 0.0732421875
                },
                {
                    "start": 470.44,
                    "end": 470.44,
                    "word": " We",
                    "probability": 0.896484375
                },
                {
                    "start": 470.44,
                    "end": 470.6,
                    "word": " have",
                    "probability": 0.876953125
                },
                {
                    "start": 470.6,
                    "end": 470.84,
                    "word": " 30",
                    "probability": 0.8349609375
                },
                {
                    "start": 470.84,
                    "end": 471.06,
                    "word": "%",
                    "probability": 0.69580078125
                },
                {
                    "start": 471.06,
                    "end": 471.46,
                    "word": " 1s",
                    "probability": 0.6982421875
                },
                {
                    "start": 471.46,
                    "end": 471.5,
                    "word": ",",
                    "probability": 0.865234375
                },
                {
                    "start": 471.5,
                    "end": 471.7,
                    "word": " 60",
                    "probability": 0.9111328125
                },
                {
                    "start": 471.7,
                    "end": 471.94,
                    "word": "%",
                    "probability": 0.9892578125
                },
                {
                    "start": 471.94,
                    "end": 472.26,
                    "word": " 2s",
                    "probability": 0.992431640625
                },
                {
                    "start": 472.26,
                    "end": 472.32,
                    "word": ",",
                    "probability": 0.493408203125
                },
                {
                    "start": 472.32,
                    "end": 472.4,
                    "word": " and",
                    "probability": 0.8681640625
                },
                {
                    "start": 472.4,
                    "end": 472.64,
                    "word": " 10",
                    "probability": 0.888671875
                },
                {
                    "start": 472.64,
                    "end": 472.84,
                    "word": "%",
                    "probability": 0.98876953125
                },
                {
                    "start": 472.84,
                    "end": 473.26,
                    "word": " 3s",
                    "probability": 0.99609375
                },
                {
                    "start": 473.26,
                    "end": 473.42,
                    "word": ".",
                    "probability": 0.8818359375
                },
                {
                    "start": 473.42,
                    "end": 473.7,
                    "word": " So",
                    "probability": 0.272216796875
                },
                {
                    "start": 473.7,
                    "end": 473.86,
                    "word": " that's",
                    "probability": 0.88623046875
                },
                {
                    "start": 473.94,
                    "end": 474.02,
                    "word": " a little",
                    "probability": 0.802734375
                },
                {
                    "start": 474.02,
                    "end": 474.18,
                    "word": " bit",
                    "probability": 0.888671875
                },
                {
                    "start": 474.18,
                    "end": 474.42,
                    "word": " different",
                    "probability": 0.6953125
                }
            ]
        },
        {
            "id": 58,
            "seek": 35738,
            "start": 474.42,
            "end": 481.34,
            "text": " from our training set. So the real world is different from our training set. So in this case,",
            "tokens": [
                50632,
                490,
                527,
                3097,
                992,
                13,
                407,
                264,
                957,
                1002,
                307,
                819,
                490,
                527,
                3097,
                992,
                13,
                407,
                294,
                341,
                1389,
                11,
                50972
            ],
            "temperature": 0.0,
            "avg_logprob": -0.18888608594575235,
            "compression_ratio": 1.7472527472527473,
            "no_speech_prob": 0.0010480880737304688,
            "words": [
                {
                    "start": 474.42,
                    "end": 474.68,
                    "word": " from",
                    "probability": 0.8212890625
                },
                {
                    "start": 474.68,
                    "end": 474.88,
                    "word": " our",
                    "probability": 0.80810546875
                },
                {
                    "start": 474.88,
                    "end": 475.1,
                    "word": " training",
                    "probability": 0.79833984375
                },
                {
                    "start": 475.1,
                    "end": 475.28,
                    "word": " set",
                    "probability": 0.8515625
                },
                {
                    "start": 475.28,
                    "end": 475.4,
                    "word": ".",
                    "probability": 0.85595703125
                },
                {
                    "start": 475.4,
                    "end": 475.58,
                    "word": " So",
                    "probability": 0.73828125
                },
                {
                    "start": 475.58,
                    "end": 475.82,
                    "word": " the",
                    "probability": 0.7509765625
                },
                {
                    "start": 475.82,
                    "end": 476.58,
                    "word": " real",
                    "probability": 0.87939453125
                },
                {
                    "start": 476.58,
                    "end": 476.88,
                    "word": " world",
                    "probability": 0.86767578125
                },
                {
                    "start": 476.88,
                    "end": 477.5,
                    "word": " is",
                    "probability": 0.90478515625
                },
                {
                    "start": 477.5,
                    "end": 478.32,
                    "word": " different",
                    "probability": 0.6826171875
                },
                {
                    "start": 478.32,
                    "end": 478.5,
                    "word": " from",
                    "probability": 0.80810546875
                },
                {
                    "start": 478.5,
                    "end": 478.64,
                    "word": " our",
                    "probability": 0.8017578125
                },
                {
                    "start": 478.64,
                    "end": 478.8,
                    "word": " training",
                    "probability": 0.7666015625
                },
                {
                    "start": 478.8,
                    "end": 479.04,
                    "word": " set",
                    "probability": 0.861328125
                },
                {
                    "start": 479.04,
                    "end": 479.48,
                    "word": ".",
                    "probability": 0.9716796875
                },
                {
                    "start": 480.0,
                    "end": 480.44,
                    "word": " So",
                    "probability": 0.83837890625
                },
                {
                    "start": 480.44,
                    "end": 480.78,
                    "word": " in",
                    "probability": 0.77734375
                },
                {
                    "start": 480.78,
                    "end": 480.94,
                    "word": " this",
                    "probability": 0.8798828125
                },
                {
                    "start": 480.94,
                    "end": 481.22,
                    "word": " case",
                    "probability": 0.80615234375
                },
                {
                    "start": 481.22,
                    "end": 481.34,
                    "word": ",",
                    "probability": 0.69580078125
                }
            ]
        },
        {
            "id": 59,
            "seek": 35738,
            "start": 481.34,
            "end": 486.24,
            "text": " this is a... we can model this distribution very easily because it's a discrete space.",
            "tokens": [
                50972,
                341,
                307,
                257,
                485,
                321,
                393,
                2316,
                341,
                7316,
                588,
                3612,
                570,
                309,
                311,
                257,
                27706,
                1901,
                13,
                51220
            ],
            "temperature": 0.0,
            "avg_logprob": -0.18888608594575235,
            "compression_ratio": 1.7472527472527473,
            "no_speech_prob": 0.0010480880737304688,
            "words": [
                {
                    "start": 481.34,
                    "end": 481.44,
                    "word": " this",
                    "probability": 0.329345703125
                },
                {
                    "start": 481.44,
                    "end": 481.56,
                    "word": " is",
                    "probability": 0.91064453125
                },
                {
                    "start": 481.56,
                    "end": 481.94,
                    "word": " a...",
                    "probability": 0.54144287109375
                },
                {
                    "start": 481.94,
                    "end": 482.14,
                    "word": " we",
                    "probability": 0.489501953125
                },
                {
                    "start": 482.14,
                    "end": 482.34,
                    "word": " can",
                    "probability": 0.8974609375
                },
                {
                    "start": 482.34,
                    "end": 482.68,
                    "word": " model",
                    "probability": 0.8203125
                },
                {
                    "start": 482.68,
                    "end": 483.1,
                    "word": " this",
                    "probability": 0.8701171875
                },
                {
                    "start": 483.1,
                    "end": 483.58,
                    "word": " distribution",
                    "probability": 0.71630859375
                },
                {
                    "start": 483.58,
                    "end": 483.88,
                    "word": " very",
                    "probability": 0.73486328125
                },
                {
                    "start": 483.88,
                    "end": 484.22,
                    "word": " easily",
                    "probability": 0.8544921875
                },
                {
                    "start": 484.22,
                    "end": 484.52,
                    "word": " because",
                    "probability": 0.66845703125
                },
                {
                    "start": 484.52,
                    "end": 484.7,
                    "word": " it's",
                    "probability": 0.94140625
                },
                {
                    "start": 484.84,
                    "end": 485.28,
                    "word": " a discrete",
                    "probability": 0.7880859375
                },
                {
                    "start": 485.28,
                    "end": 485.94,
                    "word": " space",
                    "probability": 0.81689453125
                },
                {
                    "start": 485.94,
                    "end": 486.24,
                    "word": ".",
                    "probability": 0.7958984375
                }
            ]
        },
        {
            "id": 60,
            "seek": 35738,
            "start": 486.24,
            "end": 492.64,
            "text": " There are only three things that our data could be. There's only... it isn't continuous like a lot",
            "tokens": [
                51220,
                821,
                366,
                787,
                1045,
                721,
                300,
                527,
                1412,
                727,
                312,
                13,
                821,
                311,
                787,
                485,
                309,
                1943,
                380,
                10957,
                411,
                257,
                688,
                51540
            ],
            "temperature": 0.0,
            "avg_logprob": -0.18888608594575235,
            "compression_ratio": 1.7472527472527473,
            "no_speech_prob": 0.0010480880737304688,
            "words": [
                {
                    "start": 486.24,
                    "end": 486.44,
                    "word": " There",
                    "probability": 0.8017578125
                },
                {
                    "start": 486.44,
                    "end": 486.56,
                    "word": " are",
                    "probability": 0.89208984375
                },
                {
                    "start": 486.56,
                    "end": 486.92,
                    "word": " only",
                    "probability": 0.845703125
                },
                {
                    "start": 486.92,
                    "end": 487.3,
                    "word": " three",
                    "probability": 0.78173828125
                },
                {
                    "start": 487.3,
                    "end": 488.06,
                    "word": " things",
                    "probability": 0.546875
                },
                {
                    "start": 488.06,
                    "end": 488.26,
                    "word": " that",
                    "probability": 0.83642578125
                },
                {
                    "start": 488.26,
                    "end": 488.5,
                    "word": " our",
                    "probability": 0.74365234375
                },
                {
                    "start": 488.5,
                    "end": 488.66,
                    "word": " data",
                    "probability": 0.833984375
                },
                {
                    "start": 488.66,
                    "end": 488.82,
                    "word": " could",
                    "probability": 0.8076171875
                },
                {
                    "start": 488.82,
                    "end": 489.08,
                    "word": " be",
                    "probability": 0.90380859375
                },
                {
                    "start": 489.08,
                    "end": 489.52,
                    "word": ".",
                    "probability": 0.9287109375
                },
                {
                    "start": 489.7,
                    "end": 489.9,
                    "word": " There's",
                    "probability": 0.801025390625
                },
                {
                    "start": 489.9,
                    "end": 490.92,
                    "word": " only...",
                    "probability": 0.804443359375
                },
                {
                    "start": 490.92,
                    "end": 491.32,
                    "word": " it",
                    "probability": 0.67333984375
                },
                {
                    "start": 491.32,
                    "end": 491.58,
                    "word": " isn't",
                    "probability": 0.89892578125
                },
                {
                    "start": 491.58,
                    "end": 492.02,
                    "word": " continuous",
                    "probability": 0.69677734375
                },
                {
                    "start": 492.02,
                    "end": 492.36,
                    "word": " like",
                    "probability": 0.83056640625
                },
                {
                    "start": 492.5,
                    "end": 492.64,
                    "word": " a lot",
                    "probability": 0.88818359375
                }
            ]
        },
        {
            "id": 61,
            "seek": 35738,
            "start": 492.64,
            "end": 497.6,
            "text": " of the stuff we have to deal with in data science. So we can actually compute the distribution fairly",
            "tokens": [
                51540,
                295,
                264,
                1507,
                321,
                362,
                281,
                2028,
                365,
                294,
                1412,
                3497,
                13,
                407,
                321,
                393,
                767,
                14722,
                264,
                7316,
                6457,
                51792
            ],
            "temperature": 0.0,
            "avg_logprob": -0.18888608594575235,
            "compression_ratio": 1.7472527472527473,
            "no_speech_prob": 0.0010480880737304688,
            "words": [
                {
                    "start": 492.64,
                    "end": 492.78,
                    "word": " of",
                    "probability": 0.90966796875
                },
                {
                    "start": 492.78,
                    "end": 492.94,
                    "word": " the",
                    "probability": 0.8095703125
                },
                {
                    "start": 492.94,
                    "end": 493.42,
                    "word": " stuff",
                    "probability": 0.6875
                },
                {
                    "start": 493.42,
                    "end": 493.58,
                    "word": " we",
                    "probability": 0.89697265625
                },
                {
                    "start": 493.58,
                    "end": 493.76,
                    "word": " have",
                    "probability": 0.88232421875
                },
                {
                    "start": 493.76,
                    "end": 493.86,
                    "word": " to",
                    "probability": 0.90673828125
                },
                {
                    "start": 493.86,
                    "end": 494.0,
                    "word": " deal",
                    "probability": 0.92919921875
                },
                {
                    "start": 494.0,
                    "end": 494.12,
                    "word": " with",
                    "probability": 0.8095703125
                },
                {
                    "start": 494.12,
                    "end": 494.2,
                    "word": " in",
                    "probability": 0.8486328125
                },
                {
                    "start": 494.2,
                    "end": 494.38,
                    "word": " data",
                    "probability": 0.78369140625
                },
                {
                    "start": 494.38,
                    "end": 494.66,
                    "word": " science",
                    "probability": 0.87548828125
                },
                {
                    "start": 494.66,
                    "end": 494.84,
                    "word": ".",
                    "probability": 0.65185546875
                },
                {
                    "start": 495.04,
                    "end": 495.48,
                    "word": " So",
                    "probability": 0.90087890625
                },
                {
                    "start": 495.48,
                    "end": 495.66,
                    "word": " we",
                    "probability": 0.896484375
                },
                {
                    "start": 495.66,
                    "end": 495.8,
                    "word": " can",
                    "probability": 0.89501953125
                },
                {
                    "start": 495.8,
                    "end": 496.02,
                    "word": " actually",
                    "probability": 0.78662109375
                },
                {
                    "start": 496.02,
                    "end": 496.4,
                    "word": " compute",
                    "probability": 0.859375
                },
                {
                    "start": 496.4,
                    "end": 496.72,
                    "word": " the",
                    "probability": 0.81591796875
                },
                {
                    "start": 496.72,
                    "end": 497.16,
                    "word": " distribution",
                    "probability": 0.71240234375
                },
                {
                    "start": 497.16,
                    "end": 497.6,
                    "word": " fairly",
                    "probability": 0.837890625
                }
            ]
        },
        {
            "id": 62,
            "seek": 38024,
            "start": 497.6,
            "end": 502.98,
            "text": " easily. And this is a pretty good... we can feel confident in our estimate of the distribution.",
            "tokens": [
                50364,
                3612,
                13,
                400,
                341,
                307,
                257,
                1238,
                665,
                485,
                321,
                393,
                841,
                6679,
                294,
                527,
                12539,
                295,
                264,
                7316,
                13,
                50608
            ],
            "temperature": 0.0,
            "avg_logprob": -0.2204649419319339,
            "compression_ratio": 1.6255924170616114,
            "no_speech_prob": 0.00327301025390625,
            "words": [
                {
                    "start": 497.6,
                    "end": 497.94,
                    "word": " easily",
                    "probability": 0.794921875
                },
                {
                    "start": 497.94,
                    "end": 498.1,
                    "word": ".",
                    "probability": 0.199462890625
                },
                {
                    "start": 498.1,
                    "end": 498.22,
                    "word": " And",
                    "probability": 0.765625
                },
                {
                    "start": 498.22,
                    "end": 498.46,
                    "word": " this",
                    "probability": 0.8017578125
                },
                {
                    "start": 498.46,
                    "end": 498.58,
                    "word": " is",
                    "probability": 0.90576171875
                },
                {
                    "start": 498.68,
                    "end": 498.84,
                    "word": " a pretty",
                    "probability": 0.86962890625
                },
                {
                    "start": 498.84,
                    "end": 499.26,
                    "word": " good...",
                    "probability": 0.488922119140625
                },
                {
                    "start": 499.26,
                    "end": 499.42,
                    "word": " we",
                    "probability": 0.45703125
                },
                {
                    "start": 499.42,
                    "end": 499.68,
                    "word": " can",
                    "probability": 0.70263671875
                },
                {
                    "start": 499.68,
                    "end": 500.28,
                    "word": " feel",
                    "probability": 0.916015625
                },
                {
                    "start": 500.28,
                    "end": 500.72,
                    "word": " confident",
                    "probability": 0.9384765625
                },
                {
                    "start": 500.72,
                    "end": 500.9,
                    "word": " in",
                    "probability": 0.8046875
                },
                {
                    "start": 500.9,
                    "end": 501.04,
                    "word": " our",
                    "probability": 0.82177734375
                },
                {
                    "start": 501.04,
                    "end": 501.36,
                    "word": " estimate",
                    "probability": 0.78564453125
                },
                {
                    "start": 501.36,
                    "end": 501.56,
                    "word": " of",
                    "probability": 0.90869140625
                },
                {
                    "start": 501.56,
                    "end": 501.86,
                    "word": " the",
                    "probability": 0.81005859375
                },
                {
                    "start": 501.86,
                    "end": 502.46,
                    "word": " distribution",
                    "probability": 0.7587890625
                },
                {
                    "start": 502.46,
                    "end": 502.98,
                    "word": ".",
                    "probability": 0.84033203125
                }
            ]
        },
        {
            "id": 63,
            "seek": 38024,
            "start": 503.18,
            "end": 506.76,
            "text": " We have to do some Bayesian statistics to do this properly,",
            "tokens": [
                50648,
                492,
                362,
                281,
                360,
                512,
                7840,
                42434,
                12523,
                281,
                360,
                341,
                6108,
                11,
                50780
            ],
            "temperature": 0.0,
            "avg_logprob": -0.2204649419319339,
            "compression_ratio": 1.6255924170616114,
            "no_speech_prob": 0.00327301025390625,
            "words": [
                {
                    "start": 503.18,
                    "end": 503.5,
                    "word": " We",
                    "probability": 0.8759765625
                },
                {
                    "start": 503.5,
                    "end": 503.64,
                    "word": " have",
                    "probability": 0.88623046875
                },
                {
                    "start": 503.64,
                    "end": 503.74,
                    "word": " to",
                    "probability": 0.90673828125
                },
                {
                    "start": 503.74,
                    "end": 503.84,
                    "word": " do",
                    "probability": 0.826171875
                },
                {
                    "start": 503.84,
                    "end": 503.98,
                    "word": " some",
                    "probability": 0.83251953125
                },
                {
                    "start": 503.98,
                    "end": 504.2,
                    "word": " Bayesian",
                    "probability": 0.9326171875
                },
                {
                    "start": 504.2,
                    "end": 504.54,
                    "word": " statistics",
                    "probability": 0.86962890625
                },
                {
                    "start": 504.54,
                    "end": 505.2,
                    "word": " to",
                    "probability": 0.88134765625
                },
                {
                    "start": 505.2,
                    "end": 505.36,
                    "word": " do",
                    "probability": 0.90869140625
                },
                {
                    "start": 505.36,
                    "end": 505.54,
                    "word": " this",
                    "probability": 0.884765625
                },
                {
                    "start": 505.54,
                    "end": 505.94,
                    "word": " properly",
                    "probability": 0.81201171875
                },
                {
                    "start": 505.94,
                    "end": 506.76,
                    "word": ",",
                    "probability": 0.4208984375
                }
            ]
        },
        {
            "id": 64,
            "seek": 38024,
            "start": 508.0,
            "end": 515.1,
            "text": " which we're not getting into this talk. But this is basically how it works. And then once we have",
            "tokens": [
                50892,
                597,
                321,
                434,
                406,
                1242,
                666,
                341,
                751,
                13,
                583,
                341,
                307,
                1936,
                577,
                309,
                1985,
                13,
                400,
                550,
                1564,
                321,
                362,
                51236
            ],
            "temperature": 0.0,
            "avg_logprob": -0.2204649419319339,
            "compression_ratio": 1.6255924170616114,
            "no_speech_prob": 0.00327301025390625,
            "words": [
                {
                    "start": 508.0,
                    "end": 508.52,
                    "word": " which",
                    "probability": 0.82421875
                },
                {
                    "start": 508.52,
                    "end": 508.64,
                    "word": " we're",
                    "probability": 0.907470703125
                },
                {
                    "start": 508.64,
                    "end": 508.82,
                    "word": " not",
                    "probability": 0.900390625
                },
                {
                    "start": 508.82,
                    "end": 509.0,
                    "word": " getting",
                    "probability": 0.9072265625
                },
                {
                    "start": 509.0,
                    "end": 509.2,
                    "word": " into",
                    "probability": 0.7412109375
                },
                {
                    "start": 509.2,
                    "end": 509.42,
                    "word": " this",
                    "probability": 0.7470703125
                },
                {
                    "start": 509.42,
                    "end": 509.68,
                    "word": " talk",
                    "probability": 0.80859375
                },
                {
                    "start": 509.68,
                    "end": 509.8,
                    "word": ".",
                    "probability": 0.337158203125
                },
                {
                    "start": 509.8,
                    "end": 510.22,
                    "word": " But",
                    "probability": 0.88330078125
                },
                {
                    "start": 510.22,
                    "end": 511.44,
                    "word": " this",
                    "probability": 0.8017578125
                },
                {
                    "start": 511.44,
                    "end": 511.7,
                    "word": " is",
                    "probability": 0.91162109375
                },
                {
                    "start": 511.7,
                    "end": 512.76,
                    "word": " basically",
                    "probability": 0.87841796875
                },
                {
                    "start": 512.76,
                    "end": 513.02,
                    "word": " how",
                    "probability": 0.89990234375
                },
                {
                    "start": 513.02,
                    "end": 513.12,
                    "word": " it",
                    "probability": 0.90673828125
                },
                {
                    "start": 513.12,
                    "end": 513.36,
                    "word": " works",
                    "probability": 0.7197265625
                },
                {
                    "start": 513.36,
                    "end": 513.88,
                    "word": ".",
                    "probability": 0.90869140625
                },
                {
                    "start": 514.0,
                    "end": 514.26,
                    "word": " And",
                    "probability": 0.8935546875
                },
                {
                    "start": 514.26,
                    "end": 514.38,
                    "word": " then",
                    "probability": 0.732421875
                },
                {
                    "start": 514.38,
                    "end": 514.54,
                    "word": " once",
                    "probability": 0.76171875
                },
                {
                    "start": 514.54,
                    "end": 514.7,
                    "word": " we",
                    "probability": 0.90673828125
                },
                {
                    "start": 514.7,
                    "end": 515.1,
                    "word": " have",
                    "probability": 0.8798828125
                }
            ]
        },
        {
            "id": 65,
            "seek": 38024,
            "start": 515.1,
            "end": 520.46,
            "text": " like a concept of our training distribution, and our real world distribution, or our test",
            "tokens": [
                51236,
                411,
                257,
                3410,
                295,
                527,
                3097,
                7316,
                11,
                293,
                527,
                957,
                1002,
                7316,
                11,
                420,
                527,
                1500,
                51524
            ],
            "temperature": 0.0,
            "avg_logprob": -0.2204649419319339,
            "compression_ratio": 1.6255924170616114,
            "no_speech_prob": 0.00327301025390625,
            "words": [
                {
                    "start": 515.1,
                    "end": 515.36,
                    "word": " like",
                    "probability": 0.51025390625
                },
                {
                    "start": 515.8,
                    "end": 516.08,
                    "word": " a concept",
                    "probability": 0.791015625
                },
                {
                    "start": 516.08,
                    "end": 516.3,
                    "word": " of",
                    "probability": 0.90966796875
                },
                {
                    "start": 516.3,
                    "end": 516.52,
                    "word": " our",
                    "probability": 0.8037109375
                },
                {
                    "start": 516.52,
                    "end": 516.98,
                    "word": " training",
                    "probability": 0.82763671875
                },
                {
                    "start": 516.98,
                    "end": 517.48,
                    "word": " distribution",
                    "probability": 0.767578125
                },
                {
                    "start": 517.48,
                    "end": 517.78,
                    "word": ",",
                    "probability": 0.2291259765625
                },
                {
                    "start": 517.78,
                    "end": 518.08,
                    "word": " and",
                    "probability": 0.8896484375
                },
                {
                    "start": 518.08,
                    "end": 518.56,
                    "word": " our",
                    "probability": 0.8154296875
                },
                {
                    "start": 518.56,
                    "end": 519.22,
                    "word": " real",
                    "probability": 0.84423828125
                },
                {
                    "start": 519.22,
                    "end": 519.48,
                    "word": " world",
                    "probability": 0.6328125
                },
                {
                    "start": 519.48,
                    "end": 519.92,
                    "word": " distribution",
                    "probability": 0.74609375
                },
                {
                    "start": 519.92,
                    "end": 520.04,
                    "word": ",",
                    "probability": 0.54443359375
                },
                {
                    "start": 520.04,
                    "end": 520.1,
                    "word": " or",
                    "probability": 0.74169921875
                },
                {
                    "start": 520.1,
                    "end": 520.24,
                    "word": " our",
                    "probability": 0.734375
                },
                {
                    "start": 520.24,
                    "end": 520.46,
                    "word": " test",
                    "probability": 0.6572265625
                }
            ]
        },
        {
            "id": 66,
            "seek": 41012,
            "start": 520.46,
            "end": 528.72,
            "text": " distribution, as I'll be calling it today, we can take the Kublai-Klepch divergence. So",
            "tokens": [
                50380,
                7316,
                11,
                382,
                286,
                603,
                312,
                5141,
                309,
                965,
                11,
                321,
                393,
                747,
                264,
                35805,
                875,
                72,
                12,
                42,
                306,
                79,
                339,
                47387,
                13,
                407,
                50772
            ],
            "temperature": 0.0,
            "avg_logprob": -0.2492108688209996,
            "compression_ratio": 1.6192660550458715,
            "no_speech_prob": 0.05108642578125,
            "words": [
                {
                    "start": 520.46,
                    "end": 521.22,
                    "word": " distribution",
                    "probability": 0.07366943359375
                },
                {
                    "start": 521.22,
                    "end": 521.38,
                    "word": ",",
                    "probability": 0.186279296875
                },
                {
                    "start": 521.38,
                    "end": 521.54,
                    "word": " as",
                    "probability": 0.7431640625
                },
                {
                    "start": 521.54,
                    "end": 521.64,
                    "word": " I'll",
                    "probability": 0.907470703125
                },
                {
                    "start": 521.64,
                    "end": 521.74,
                    "word": " be",
                    "probability": 0.908203125
                },
                {
                    "start": 521.74,
                    "end": 522.0,
                    "word": " calling",
                    "probability": 0.67626953125
                },
                {
                    "start": 522.0,
                    "end": 522.2,
                    "word": " it",
                    "probability": 0.89404296875
                },
                {
                    "start": 522.2,
                    "end": 522.94,
                    "word": " today",
                    "probability": 0.494873046875
                },
                {
                    "start": 522.94,
                    "end": 523.8,
                    "word": ",",
                    "probability": 0.70458984375
                },
                {
                    "start": 523.8,
                    "end": 524.36,
                    "word": " we",
                    "probability": 0.88671875
                },
                {
                    "start": 524.36,
                    "end": 524.8,
                    "word": " can",
                    "probability": 0.89306640625
                },
                {
                    "start": 524.8,
                    "end": 525.2,
                    "word": " take",
                    "probability": 0.79833984375
                },
                {
                    "start": 525.2,
                    "end": 525.52,
                    "word": " the",
                    "probability": 0.79638671875
                },
                {
                    "start": 525.52,
                    "end": 526.54,
                    "word": " Kublai",
                    "probability": 0.6394856770833334
                },
                {
                    "start": 526.54,
                    "end": 526.92,
                    "word": "-Klepch",
                    "probability": 0.48092041015625
                },
                {
                    "start": 526.92,
                    "end": 527.9,
                    "word": " divergence",
                    "probability": 0.426025390625
                },
                {
                    "start": 527.9,
                    "end": 528.48,
                    "word": ".",
                    "probability": 0.90478515625
                },
                {
                    "start": 528.52,
                    "end": 528.72,
                    "word": " So",
                    "probability": 0.67919921875
                }
            ]
        },
        {
            "id": 67,
            "seek": 41012,
            "start": 529.5,
            "end": 535.93,
            "text": " this is a sort of a measure of the distance between this distribution and this distribution.",
            "tokens": [
                50816,
                341,
                307,
                257,
                1333,
                295,
                257,
                3481,
                295,
                264,
                4560,
                1296,
                341,
                7316,
                293,
                341,
                7316,
                13,
                51124
            ],
            "temperature": 0.0,
            "avg_logprob": -0.2492108688209996,
            "compression_ratio": 1.6192660550458715,
            "no_speech_prob": 0.05108642578125,
            "words": [
                {
                    "start": 529.5,
                    "end": 529.8,
                    "word": " this",
                    "probability": 0.6943359375
                },
                {
                    "start": 529.8,
                    "end": 530.4,
                    "word": " is",
                    "probability": 0.90185546875
                },
                {
                    "start": 531.28,
                    "end": 531.68,
                    "word": " a sort",
                    "probability": 0.53662109375
                },
                {
                    "start": 531.68,
                    "end": 531.76,
                    "word": " of",
                    "probability": 0.9052734375
                },
                {
                    "start": 531.96,
                    "end": 532.24,
                    "word": " a measure",
                    "probability": 0.77880859375
                },
                {
                    "start": 532.24,
                    "end": 532.68,
                    "word": " of",
                    "probability": 0.8994140625
                },
                {
                    "start": 532.68,
                    "end": 532.9,
                    "word": " the",
                    "probability": 0.8193359375
                },
                {
                    "start": 532.9,
                    "end": 533.26,
                    "word": " distance",
                    "probability": 0.79443359375
                },
                {
                    "start": 533.26,
                    "end": 533.56,
                    "word": " between",
                    "probability": 0.7880859375
                },
                {
                    "start": 533.56,
                    "end": 533.82,
                    "word": " this",
                    "probability": 0.875
                },
                {
                    "start": 533.82,
                    "end": 534.34,
                    "word": " distribution",
                    "probability": 0.6923828125
                },
                {
                    "start": 534.34,
                    "end": 534.82,
                    "word": " and",
                    "probability": 0.87744140625
                },
                {
                    "start": 534.82,
                    "end": 535.1,
                    "word": " this",
                    "probability": 0.8798828125
                },
                {
                    "start": 535.1,
                    "end": 535.64,
                    "word": " distribution",
                    "probability": 0.72216796875
                },
                {
                    "start": 535.64,
                    "end": 535.93,
                    "word": ".",
                    "probability": 0.560546875
                }
            ]
        },
        {
            "id": 68,
            "seek": 41012,
            "start": 536.2,
            "end": 543.76,
            "text": " And it's very easy to calculate on this categorical discrete data because it's just",
            "tokens": [
                51156,
                400,
                309,
                311,
                588,
                1858,
                281,
                8873,
                322,
                341,
                19250,
                804,
                27706,
                1412,
                570,
                309,
                311,
                445,
                51524
            ],
            "temperature": 0.0,
            "avg_logprob": -0.2492108688209996,
            "compression_ratio": 1.6192660550458715,
            "no_speech_prob": 0.05108642578125,
            "words": [
                {
                    "start": 536.2,
                    "end": 536.62,
                    "word": " And",
                    "probability": 0.82373046875
                },
                {
                    "start": 536.62,
                    "end": 536.74,
                    "word": " it's",
                    "probability": 0.94873046875
                },
                {
                    "start": 536.74,
                    "end": 536.9,
                    "word": " very",
                    "probability": 0.744140625
                },
                {
                    "start": 536.9,
                    "end": 537.14,
                    "word": " easy",
                    "probability": 0.80419921875
                },
                {
                    "start": 537.14,
                    "end": 537.32,
                    "word": " to",
                    "probability": 0.90673828125
                },
                {
                    "start": 537.32,
                    "end": 537.62,
                    "word": " calculate",
                    "probability": 0.8583984375
                },
                {
                    "start": 537.62,
                    "end": 538.1,
                    "word": " on",
                    "probability": 0.89990234375
                },
                {
                    "start": 538.1,
                    "end": 538.56,
                    "word": " this",
                    "probability": 0.8818359375
                },
                {
                    "start": 538.56,
                    "end": 539.92,
                    "word": " categorical",
                    "probability": 0.886962890625
                },
                {
                    "start": 539.92,
                    "end": 541.52,
                    "word": " discrete",
                    "probability": 0.66748046875
                },
                {
                    "start": 541.52,
                    "end": 542.84,
                    "word": " data",
                    "probability": 0.78076171875
                },
                {
                    "start": 542.84,
                    "end": 543.44,
                    "word": " because",
                    "probability": 0.35693359375
                },
                {
                    "start": 543.44,
                    "end": 543.64,
                    "word": " it's",
                    "probability": 0.950927734375
                },
                {
                    "start": 543.64,
                    "end": 543.76,
                    "word": " just",
                    "probability": 0.875
                }
            ]
        },
        {
            "id": 69,
            "seek": 41012,
            "start": 543.76,
            "end": 550.34,
            "text": " going to be, well, 0.7 times log of 0.7 over 0.3. So that's the category for 1s over the",
            "tokens": [
                51524,
                516,
                281,
                312,
                11,
                731,
                11,
                1958,
                13,
                22,
                1413,
                3565,
                295,
                1958,
                13,
                22,
                670,
                1958,
                13,
                18,
                13,
                407,
                300,
                311,
                264,
                7719,
                337,
                502,
                82,
                670,
                264,
                51852
            ],
            "temperature": 0.0,
            "avg_logprob": -0.2492108688209996,
            "compression_ratio": 1.6192660550458715,
            "no_speech_prob": 0.05108642578125,
            "words": [
                {
                    "start": 543.76,
                    "end": 543.9,
                    "word": " going",
                    "probability": 0.841796875
                },
                {
                    "start": 543.9,
                    "end": 543.96,
                    "word": " to",
                    "probability": 0.9091796875
                },
                {
                    "start": 543.96,
                    "end": 544.2,
                    "word": " be",
                    "probability": 0.9052734375
                },
                {
                    "start": 544.2,
                    "end": 544.28,
                    "word": ",",
                    "probability": 0.515625
                },
                {
                    "start": 544.28,
                    "end": 544.48,
                    "word": " well",
                    "probability": 0.857421875
                },
                {
                    "start": 544.48,
                    "end": 544.9,
                    "word": ",",
                    "probability": 0.80078125
                },
                {
                    "start": 544.9,
                    "end": 545.2,
                    "word": " 0",
                    "probability": 0.86328125
                },
                {
                    "start": 545.2,
                    "end": 545.7,
                    "word": ".7",
                    "probability": 0.99658203125
                },
                {
                    "start": 545.7,
                    "end": 546.84,
                    "word": " times",
                    "probability": 0.712890625
                },
                {
                    "start": 546.84,
                    "end": 547.22,
                    "word": " log",
                    "probability": 0.779296875
                },
                {
                    "start": 547.22,
                    "end": 547.34,
                    "word": " of",
                    "probability": 0.65283203125
                },
                {
                    "start": 547.34,
                    "end": 547.46,
                    "word": " 0",
                    "probability": 0.970703125
                },
                {
                    "start": 547.46,
                    "end": 547.64,
                    "word": ".7",
                    "probability": 0.99658203125
                },
                {
                    "start": 547.64,
                    "end": 547.78,
                    "word": " over",
                    "probability": 0.79931640625
                },
                {
                    "start": 547.78,
                    "end": 547.96,
                    "word": " 0",
                    "probability": 0.97705078125
                },
                {
                    "start": 547.96,
                    "end": 548.16,
                    "word": ".3",
                    "probability": 0.995849609375
                },
                {
                    "start": 548.16,
                    "end": 548.26,
                    "word": ".",
                    "probability": 0.80810546875
                },
                {
                    "start": 548.26,
                    "end": 548.36,
                    "word": " So",
                    "probability": 0.8896484375
                },
                {
                    "start": 548.36,
                    "end": 548.64,
                    "word": " that's",
                    "probability": 0.9375
                },
                {
                    "start": 548.64,
                    "end": 548.9,
                    "word": " the",
                    "probability": 0.81201171875
                },
                {
                    "start": 548.9,
                    "end": 549.36,
                    "word": " category",
                    "probability": 0.78271484375
                },
                {
                    "start": 549.36,
                    "end": 549.6,
                    "word": " for",
                    "probability": 0.85791015625
                },
                {
                    "start": 549.6,
                    "end": 549.96,
                    "word": " 1s",
                    "probability": 0.38916015625
                },
                {
                    "start": 549.96,
                    "end": 550.2,
                    "word": " over",
                    "probability": 0.7998046875
                },
                {
                    "start": 550.2,
                    "end": 550.34,
                    "word": " the",
                    "probability": 0.80908203125
                }
            ]
        },
        {
            "id": 70,
            "seek": 43384,
            "start": 551.16,
            "end": 556.12,
            "text": " 1s. Take the log and then multiply it for the category for 1s. And that gives you the first",
            "tokens": [
                50400,
                502,
                82,
                13,
                3664,
                264,
                3565,
                293,
                550,
                12972,
                309,
                337,
                264,
                7719,
                337,
                502,
                82,
                13,
                400,
                300,
                2709,
                291,
                264,
                700,
                50652
            ],
            "temperature": 0.0,
            "avg_logprob": -0.2306621263522913,
            "compression_ratio": 1.701834862385321,
            "no_speech_prob": 0.0021152496337890625,
            "words": [
                {
                    "start": 551.16,
                    "end": 551.6,
                    "word": " 1s",
                    "probability": 0.2597770690917969
                },
                {
                    "start": 551.6,
                    "end": 552.04,
                    "word": ".",
                    "probability": 0.2330322265625
                },
                {
                    "start": 552.56,
                    "end": 552.68,
                    "word": " Take",
                    "probability": 0.2364501953125
                },
                {
                    "start": 552.68,
                    "end": 552.86,
                    "word": " the",
                    "probability": 0.80029296875
                },
                {
                    "start": 552.86,
                    "end": 553.1,
                    "word": " log",
                    "probability": 0.85400390625
                },
                {
                    "start": 553.1,
                    "end": 553.24,
                    "word": " and",
                    "probability": 0.7939453125
                },
                {
                    "start": 553.24,
                    "end": 553.36,
                    "word": " then",
                    "probability": 0.69873046875
                },
                {
                    "start": 553.36,
                    "end": 553.64,
                    "word": " multiply",
                    "probability": 0.77099609375
                },
                {
                    "start": 553.64,
                    "end": 553.84,
                    "word": " it",
                    "probability": 0.828125
                },
                {
                    "start": 553.84,
                    "end": 553.94,
                    "word": " for",
                    "probability": 0.72021484375
                },
                {
                    "start": 553.94,
                    "end": 554.02,
                    "word": " the",
                    "probability": 0.7490234375
                },
                {
                    "start": 554.02,
                    "end": 554.28,
                    "word": " category",
                    "probability": 0.73583984375
                },
                {
                    "start": 554.28,
                    "end": 554.68,
                    "word": " for",
                    "probability": 0.8115234375
                },
                {
                    "start": 554.68,
                    "end": 555.08,
                    "word": " 1s",
                    "probability": 0.911376953125
                },
                {
                    "start": 555.08,
                    "end": 555.22,
                    "word": ".",
                    "probability": 0.40771484375
                },
                {
                    "start": 555.22,
                    "end": 555.36,
                    "word": " And",
                    "probability": 0.75048828125
                },
                {
                    "start": 555.36,
                    "end": 555.5,
                    "word": " that",
                    "probability": 0.86376953125
                },
                {
                    "start": 555.5,
                    "end": 555.66,
                    "word": " gives",
                    "probability": 0.85693359375
                },
                {
                    "start": 555.66,
                    "end": 555.78,
                    "word": " you",
                    "probability": 0.8916015625
                },
                {
                    "start": 555.78,
                    "end": 555.9,
                    "word": " the",
                    "probability": 0.814453125
                },
                {
                    "start": 555.9,
                    "end": 556.12,
                    "word": " first",
                    "probability": 0.7763671875
                }
            ]
        },
        {
            "id": 71,
            "seek": 43384,
            "start": 556.12,
            "end": 561.14,
            "text": " term. Second term is the same thing for 2s. And the third term is the same thing for 3s.",
            "tokens": [
                50652,
                1433,
                13,
                5736,
                1433,
                307,
                264,
                912,
                551,
                337,
                568,
                82,
                13,
                400,
                264,
                2636,
                1433,
                307,
                264,
                912,
                551,
                337,
                805,
                82,
                13,
                50876
            ],
            "temperature": 0.0,
            "avg_logprob": -0.2306621263522913,
            "compression_ratio": 1.701834862385321,
            "no_speech_prob": 0.0021152496337890625,
            "words": [
                {
                    "start": 556.12,
                    "end": 556.44,
                    "word": " term",
                    "probability": 0.82568359375
                },
                {
                    "start": 556.44,
                    "end": 556.7,
                    "word": ".",
                    "probability": 0.59912109375
                },
                {
                    "start": 556.7,
                    "end": 557.08,
                    "word": " Second",
                    "probability": 0.57763671875
                },
                {
                    "start": 557.08,
                    "end": 557.42,
                    "word": " term",
                    "probability": 0.82080078125
                },
                {
                    "start": 557.42,
                    "end": 557.56,
                    "word": " is",
                    "probability": 0.8759765625
                },
                {
                    "start": 557.56,
                    "end": 557.78,
                    "word": " the",
                    "probability": 0.81103515625
                },
                {
                    "start": 557.78,
                    "end": 558.0,
                    "word": " same",
                    "probability": 0.8212890625
                },
                {
                    "start": 558.0,
                    "end": 558.2,
                    "word": " thing",
                    "probability": 0.8134765625
                },
                {
                    "start": 558.2,
                    "end": 558.38,
                    "word": " for",
                    "probability": 0.89404296875
                },
                {
                    "start": 558.38,
                    "end": 558.72,
                    "word": " 2s",
                    "probability": 0.931884765625
                },
                {
                    "start": 558.72,
                    "end": 558.74,
                    "word": ".",
                    "probability": 0.28173828125
                },
                {
                    "start": 558.74,
                    "end": 558.84,
                    "word": " And",
                    "probability": 0.85400390625
                },
                {
                    "start": 558.84,
                    "end": 558.98,
                    "word": " the",
                    "probability": 0.80078125
                },
                {
                    "start": 558.98,
                    "end": 559.18,
                    "word": " third",
                    "probability": 0.86767578125
                },
                {
                    "start": 559.18,
                    "end": 559.48,
                    "word": " term",
                    "probability": 0.8125
                },
                {
                    "start": 559.48,
                    "end": 559.62,
                    "word": " is",
                    "probability": 0.8564453125
                },
                {
                    "start": 559.62,
                    "end": 559.72,
                    "word": " the",
                    "probability": 0.771484375
                },
                {
                    "start": 559.72,
                    "end": 559.86,
                    "word": " same",
                    "probability": 0.81201171875
                },
                {
                    "start": 559.86,
                    "end": 560.04,
                    "word": " thing",
                    "probability": 0.81396484375
                },
                {
                    "start": 560.04,
                    "end": 560.22,
                    "word": " for",
                    "probability": 0.89794921875
                },
                {
                    "start": 560.22,
                    "end": 560.7,
                    "word": " 3s",
                    "probability": 0.98681640625
                },
                {
                    "start": 560.7,
                    "end": 561.14,
                    "word": ".",
                    "probability": 0.9443359375
                }
            ]
        },
        {
            "id": 72,
            "seek": 43384,
            "start": 561.24,
            "end": 566.52,
            "text": " And then we add those all up and you get the KL divergence between these two distributions is",
            "tokens": [
                50904,
                400,
                550,
                321,
                909,
                729,
                439,
                493,
                293,
                291,
                483,
                264,
                47991,
                47387,
                1296,
                613,
                732,
                37870,
                307,
                51168
            ],
            "temperature": 0.0,
            "avg_logprob": -0.2306621263522913,
            "compression_ratio": 1.701834862385321,
            "no_speech_prob": 0.0021152496337890625,
            "words": [
                {
                    "start": 561.24,
                    "end": 561.42,
                    "word": " And",
                    "probability": 0.8701171875
                },
                {
                    "start": 561.42,
                    "end": 561.64,
                    "word": " then",
                    "probability": 0.7265625
                },
                {
                    "start": 561.64,
                    "end": 562.08,
                    "word": " we",
                    "probability": 0.2744140625
                },
                {
                    "start": 562.08,
                    "end": 562.38,
                    "word": " add",
                    "probability": 0.80859375
                },
                {
                    "start": 562.38,
                    "end": 562.6,
                    "word": " those",
                    "probability": 0.84375
                },
                {
                    "start": 562.6,
                    "end": 562.74,
                    "word": " all",
                    "probability": 0.89453125
                },
                {
                    "start": 562.74,
                    "end": 562.94,
                    "word": " up",
                    "probability": 0.90966796875
                },
                {
                    "start": 562.94,
                    "end": 563.14,
                    "word": " and",
                    "probability": 0.75439453125
                },
                {
                    "start": 563.14,
                    "end": 563.24,
                    "word": " you",
                    "probability": 0.88671875
                },
                {
                    "start": 563.24,
                    "end": 563.38,
                    "word": " get",
                    "probability": 0.90185546875
                },
                {
                    "start": 563.38,
                    "end": 563.62,
                    "word": " the",
                    "probability": 0.73583984375
                },
                {
                    "start": 563.62,
                    "end": 564.06,
                    "word": " KL",
                    "probability": 0.669921875
                },
                {
                    "start": 564.06,
                    "end": 564.54,
                    "word": " divergence",
                    "probability": 0.75732421875
                },
                {
                    "start": 564.54,
                    "end": 565.08,
                    "word": " between",
                    "probability": 0.76171875
                },
                {
                    "start": 565.08,
                    "end": 565.42,
                    "word": " these",
                    "probability": 0.71630859375
                },
                {
                    "start": 565.42,
                    "end": 565.62,
                    "word": " two",
                    "probability": 0.7998046875
                },
                {
                    "start": 565.62,
                    "end": 566.16,
                    "word": " distributions",
                    "probability": 0.84326171875
                },
                {
                    "start": 566.16,
                    "end": 566.52,
                    "word": " is",
                    "probability": 0.83056640625
                }
            ]
        },
        {
            "id": 73,
            "seek": 43384,
            "start": 566.52,
            "end": 574.06,
            "text": " 0.16. So that's all great. But the problem is our data looks like this. And I don't know whether",
            "tokens": [
                51168,
                1958,
                13,
                6866,
                13,
                407,
                300,
                311,
                439,
                869,
                13,
                583,
                264,
                1154,
                307,
                527,
                1412,
                1542,
                411,
                341,
                13,
                400,
                286,
                500,
                380,
                458,
                1968,
                51564
            ],
            "temperature": 0.0,
            "avg_logprob": -0.2306621263522913,
            "compression_ratio": 1.701834862385321,
            "no_speech_prob": 0.0021152496337890625,
            "words": [
                {
                    "start": 566.52,
                    "end": 566.7,
                    "word": " 0",
                    "probability": 0.78564453125
                },
                {
                    "start": 566.7,
                    "end": 567.04,
                    "word": ".16",
                    "probability": 0.993896484375
                },
                {
                    "start": 567.04,
                    "end": 567.48,
                    "word": ".",
                    "probability": 0.9677734375
                },
                {
                    "start": 568.78,
                    "end": 569.22,
                    "word": " So",
                    "probability": 0.8857421875
                },
                {
                    "start": 569.22,
                    "end": 569.46,
                    "word": " that's",
                    "probability": 0.9111328125
                },
                {
                    "start": 569.46,
                    "end": 569.6,
                    "word": " all",
                    "probability": 0.89306640625
                },
                {
                    "start": 569.6,
                    "end": 569.96,
                    "word": " great",
                    "probability": 0.87646484375
                },
                {
                    "start": 569.96,
                    "end": 570.08,
                    "word": ".",
                    "probability": 0.443603515625
                },
                {
                    "start": 570.08,
                    "end": 570.22,
                    "word": " But",
                    "probability": 0.88623046875
                },
                {
                    "start": 570.22,
                    "end": 570.38,
                    "word": " the",
                    "probability": 0.8154296875
                },
                {
                    "start": 570.38,
                    "end": 570.66,
                    "word": " problem",
                    "probability": 0.76513671875
                },
                {
                    "start": 570.66,
                    "end": 570.88,
                    "word": " is",
                    "probability": 0.90625
                },
                {
                    "start": 570.88,
                    "end": 571.12,
                    "word": " our",
                    "probability": 0.6201171875
                },
                {
                    "start": 571.12,
                    "end": 571.3,
                    "word": " data",
                    "probability": 0.8125
                },
                {
                    "start": 571.3,
                    "end": 571.54,
                    "word": " looks",
                    "probability": 0.79931640625
                },
                {
                    "start": 571.54,
                    "end": 571.74,
                    "word": " like",
                    "probability": 0.876953125
                },
                {
                    "start": 571.74,
                    "end": 572.1,
                    "word": " this",
                    "probability": 0.8837890625
                },
                {
                    "start": 572.1,
                    "end": 572.54,
                    "word": ".",
                    "probability": 0.77978515625
                },
                {
                    "start": 572.64,
                    "end": 573.08,
                    "word": " And",
                    "probability": 0.8203125
                },
                {
                    "start": 573.08,
                    "end": 573.24,
                    "word": " I",
                    "probability": 0.99267578125
                },
                {
                    "start": 573.24,
                    "end": 573.44,
                    "word": " don't",
                    "probability": 0.946533203125
                },
                {
                    "start": 573.44,
                    "end": 573.8,
                    "word": " know",
                    "probability": 0.81298828125
                },
                {
                    "start": 573.8,
                    "end": 574.06,
                    "word": " whether",
                    "probability": 0.8056640625
                }
            ]
        },
        {
            "id": 74,
            "seek": 45362,
            "start": 574.06,
            "end": 582.2,
            "text": " this has any difference. Are my blue points sampled from the same distribution as my orange",
            "tokens": [
                50380,
                341,
                575,
                604,
                2649,
                13,
                2014,
                452,
                3344,
                2793,
                3247,
                15551,
                490,
                264,
                912,
                7316,
                382,
                452,
                7671,
                50772
            ],
            "temperature": 0.0,
            "avg_logprob": -0.20852481376598864,
            "compression_ratio": 1.5235294117647058,
            "no_speech_prob": 0.0038852691650390625,
            "words": [
                {
                    "start": 574.06,
                    "end": 574.66,
                    "word": " this",
                    "probability": 0.10980224609375
                },
                {
                    "start": 574.66,
                    "end": 576.5,
                    "word": " has",
                    "probability": 0.477294921875
                },
                {
                    "start": 576.5,
                    "end": 576.8,
                    "word": " any",
                    "probability": 0.7607421875
                },
                {
                    "start": 576.8,
                    "end": 577.3,
                    "word": " difference",
                    "probability": 0.64306640625
                },
                {
                    "start": 577.3,
                    "end": 577.58,
                    "word": ".",
                    "probability": 0.488525390625
                },
                {
                    "start": 577.58,
                    "end": 577.82,
                    "word": " Are",
                    "probability": 0.84375
                },
                {
                    "start": 577.82,
                    "end": 578.98,
                    "word": " my",
                    "probability": 0.7451171875
                },
                {
                    "start": 578.98,
                    "end": 579.26,
                    "word": " blue",
                    "probability": 0.91455078125
                },
                {
                    "start": 579.26,
                    "end": 579.66,
                    "word": " points",
                    "probability": 0.7998046875
                },
                {
                    "start": 579.66,
                    "end": 580.34,
                    "word": " sampled",
                    "probability": 0.89453125
                },
                {
                    "start": 580.34,
                    "end": 580.64,
                    "word": " from",
                    "probability": 0.82177734375
                },
                {
                    "start": 580.64,
                    "end": 580.78,
                    "word": " the",
                    "probability": 0.8095703125
                },
                {
                    "start": 580.78,
                    "end": 580.96,
                    "word": " same",
                    "probability": 0.78857421875
                },
                {
                    "start": 580.96,
                    "end": 581.52,
                    "word": " distribution",
                    "probability": 0.76171875
                },
                {
                    "start": 581.52,
                    "end": 581.78,
                    "word": " as",
                    "probability": 0.59423828125
                },
                {
                    "start": 581.78,
                    "end": 581.94,
                    "word": " my",
                    "probability": 0.8974609375
                },
                {
                    "start": 581.94,
                    "end": 582.2,
                    "word": " orange",
                    "probability": 0.83447265625
                }
            ]
        },
        {
            "id": 75,
            "seek": 45362,
            "start": 582.2,
            "end": 588.02,
            "text": " points? I don't know. It's really hard to tell by just looking at this picture.",
            "tokens": [
                50772,
                2793,
                30,
                286,
                500,
                380,
                458,
                13,
                467,
                311,
                534,
                1152,
                281,
                980,
                538,
                445,
                1237,
                412,
                341,
                3036,
                13,
                51032
            ],
            "temperature": 0.0,
            "avg_logprob": -0.20852481376598864,
            "compression_ratio": 1.5235294117647058,
            "no_speech_prob": 0.0038852691650390625,
            "words": [
                {
                    "start": 582.2,
                    "end": 582.56,
                    "word": " points",
                    "probability": 0.7548828125
                },
                {
                    "start": 582.56,
                    "end": 583.0,
                    "word": "?",
                    "probability": 0.96044921875
                },
                {
                    "start": 583.12,
                    "end": 583.68,
                    "word": " I",
                    "probability": 0.9697265625
                },
                {
                    "start": 583.68,
                    "end": 583.84,
                    "word": " don't",
                    "probability": 0.9482421875
                },
                {
                    "start": 583.84,
                    "end": 584.12,
                    "word": " know",
                    "probability": 0.81787109375
                },
                {
                    "start": 584.12,
                    "end": 584.68,
                    "word": ".",
                    "probability": 0.75537109375
                },
                {
                    "start": 584.94,
                    "end": 585.46,
                    "word": " It's",
                    "probability": 0.9384765625
                },
                {
                    "start": 585.46,
                    "end": 585.68,
                    "word": " really",
                    "probability": 0.7822265625
                },
                {
                    "start": 585.68,
                    "end": 585.86,
                    "word": " hard",
                    "probability": 0.81787109375
                },
                {
                    "start": 585.86,
                    "end": 586.02,
                    "word": " to",
                    "probability": 0.904296875
                },
                {
                    "start": 586.02,
                    "end": 586.18,
                    "word": " tell",
                    "probability": 0.822265625
                },
                {
                    "start": 586.18,
                    "end": 586.34,
                    "word": " by",
                    "probability": 0.88232421875
                },
                {
                    "start": 586.34,
                    "end": 586.5,
                    "word": " just",
                    "probability": 0.88134765625
                },
                {
                    "start": 586.5,
                    "end": 586.78,
                    "word": " looking",
                    "probability": 0.8037109375
                },
                {
                    "start": 586.78,
                    "end": 586.96,
                    "word": " at",
                    "probability": 0.9091796875
                },
                {
                    "start": 586.96,
                    "end": 587.1,
                    "word": " this",
                    "probability": 0.86865234375
                },
                {
                    "start": 587.1,
                    "end": 587.46,
                    "word": " picture",
                    "probability": 0.7607421875
                },
                {
                    "start": 587.46,
                    "end": 588.02,
                    "word": ".",
                    "probability": 0.810546875
                }
            ]
        },
        {
            "id": 76,
            "seek": 45362,
            "start": 588.6,
            "end": 593.84,
            "text": " And now, in this case, they are. So my orange points are sampled from a two-dimensional",
            "tokens": [
                51100,
                400,
                586,
                11,
                294,
                341,
                1389,
                11,
                436,
                366,
                13,
                407,
                452,
                7671,
                2793,
                366,
                3247,
                15551,
                490,
                257,
                732,
                12,
                18759,
                51376
            ],
            "temperature": 0.0,
            "avg_logprob": -0.20852481376598864,
            "compression_ratio": 1.5235294117647058,
            "no_speech_prob": 0.0038852691650390625,
            "words": [
                {
                    "start": 588.6,
                    "end": 588.62,
                    "word": " And",
                    "probability": 0.59130859375
                },
                {
                    "start": 588.62,
                    "end": 589.24,
                    "word": " now",
                    "probability": 0.86669921875
                },
                {
                    "start": 589.24,
                    "end": 589.52,
                    "word": ",",
                    "probability": 0.375
                },
                {
                    "start": 589.52,
                    "end": 589.66,
                    "word": " in",
                    "probability": 0.90576171875
                },
                {
                    "start": 589.66,
                    "end": 589.82,
                    "word": " this",
                    "probability": 0.8798828125
                },
                {
                    "start": 589.82,
                    "end": 590.12,
                    "word": " case",
                    "probability": 0.8046875
                },
                {
                    "start": 590.12,
                    "end": 590.38,
                    "word": ",",
                    "probability": 0.84619140625
                },
                {
                    "start": 590.38,
                    "end": 590.62,
                    "word": " they",
                    "probability": 0.73681640625
                },
                {
                    "start": 590.62,
                    "end": 590.96,
                    "word": " are",
                    "probability": 0.8935546875
                },
                {
                    "start": 590.96,
                    "end": 591.14,
                    "word": ".",
                    "probability": 0.8828125
                },
                {
                    "start": 591.14,
                    "end": 591.24,
                    "word": " So",
                    "probability": 0.86328125
                },
                {
                    "start": 591.24,
                    "end": 591.44,
                    "word": " my",
                    "probability": 0.8134765625
                },
                {
                    "start": 591.44,
                    "end": 591.8,
                    "word": " orange",
                    "probability": 0.8251953125
                },
                {
                    "start": 591.8,
                    "end": 592.16,
                    "word": " points",
                    "probability": 0.79150390625
                },
                {
                    "start": 592.16,
                    "end": 592.3,
                    "word": " are",
                    "probability": 0.88720703125
                },
                {
                    "start": 592.3,
                    "end": 592.6,
                    "word": " sampled",
                    "probability": 0.89404296875
                },
                {
                    "start": 592.6,
                    "end": 592.94,
                    "word": " from",
                    "probability": 0.81884765625
                },
                {
                    "start": 593.24,
                    "end": 593.42,
                    "word": " a two",
                    "probability": 0.78076171875
                },
                {
                    "start": 593.42,
                    "end": 593.84,
                    "word": "-dimensional",
                    "probability": 0.794677734375
                }
            ]
        },
        {
            "id": 77,
            "seek": 48336,
            "start": 594.34,
            "end": 602.04,
            "text": " Gaussian, two-dimensional normal distribution, where it's just, it's a nice sphere. And they're",
            "tokens": [
                50388,
                39148,
                11,
                732,
                12,
                18759,
                2710,
                7316,
                11,
                689,
                309,
                311,
                445,
                11,
                309,
                311,
                257,
                1481,
                16687,
                13,
                400,
                436,
                434,
                50760
            ],
            "temperature": 0.0,
            "avg_logprob": -0.24711538415688736,
            "compression_ratio": 1.750915750915751,
            "no_speech_prob": 0.381103515625,
            "words": [
                {
                    "start": 594.34,
                    "end": 594.72,
                    "word": " Gaussian",
                    "probability": 0.2109375
                },
                {
                    "start": 594.72,
                    "end": 594.9,
                    "word": ",",
                    "probability": 0.2218017578125
                },
                {
                    "start": 594.9,
                    "end": 595.04,
                    "word": " two",
                    "probability": 0.63427734375
                },
                {
                    "start": 595.04,
                    "end": 595.26,
                    "word": "-dimensional",
                    "probability": 0.84130859375
                },
                {
                    "start": 595.26,
                    "end": 595.8,
                    "word": " normal",
                    "probability": 0.6748046875
                },
                {
                    "start": 595.8,
                    "end": 596.94,
                    "word": " distribution",
                    "probability": 0.74658203125
                },
                {
                    "start": 596.94,
                    "end": 597.9,
                    "word": ",",
                    "probability": 0.34912109375
                },
                {
                    "start": 597.9,
                    "end": 599.16,
                    "word": " where",
                    "probability": 0.84033203125
                },
                {
                    "start": 599.16,
                    "end": 599.66,
                    "word": " it's",
                    "probability": 0.930908203125
                },
                {
                    "start": 599.66,
                    "end": 599.84,
                    "word": " just",
                    "probability": 0.72021484375
                },
                {
                    "start": 599.84,
                    "end": 599.98,
                    "word": ",",
                    "probability": 0.351318359375
                },
                {
                    "start": 599.98,
                    "end": 600.42,
                    "word": " it's",
                    "probability": 0.94677734375
                },
                {
                    "start": 600.48,
                    "end": 600.7,
                    "word": " a nice",
                    "probability": 0.8271484375
                },
                {
                    "start": 600.7,
                    "end": 601.32,
                    "word": " sphere",
                    "probability": 0.76806640625
                },
                {
                    "start": 601.32,
                    "end": 601.64,
                    "word": ".",
                    "probability": 0.1551513671875
                },
                {
                    "start": 601.64,
                    "end": 601.68,
                    "word": " And",
                    "probability": 0.78662109375
                },
                {
                    "start": 601.68,
                    "end": 602.04,
                    "word": " they're",
                    "probability": 0.619873046875
                }
            ]
        },
        {
            "id": 78,
            "seek": 48336,
            "start": 603.2,
            "end": 608.18,
            "text": " both, the covariance matrix is 1, 1. And the covariance matrix with blues is 1, 1. And it",
            "tokens": [
                50832,
                1293,
                11,
                264,
                49851,
                719,
                8141,
                307,
                502,
                11,
                502,
                13,
                400,
                264,
                49851,
                719,
                8141,
                365,
                24244,
                307,
                502,
                11,
                502,
                13,
                400,
                309,
                51072
            ],
            "temperature": 0.0,
            "avg_logprob": -0.24711538415688736,
            "compression_ratio": 1.750915750915751,
            "no_speech_prob": 0.381103515625,
            "words": [
                {
                    "start": 603.2,
                    "end": 603.58,
                    "word": " both",
                    "probability": 0.62890625
                },
                {
                    "start": 603.58,
                    "end": 603.76,
                    "word": ",",
                    "probability": 0.66650390625
                },
                {
                    "start": 603.76,
                    "end": 604.1,
                    "word": " the",
                    "probability": 0.6171875
                },
                {
                    "start": 604.1,
                    "end": 604.6,
                    "word": " covariance",
                    "probability": 0.824462890625
                },
                {
                    "start": 604.6,
                    "end": 604.88,
                    "word": " matrix",
                    "probability": 0.83642578125
                },
                {
                    "start": 604.88,
                    "end": 605.1,
                    "word": " is",
                    "probability": 0.890625
                },
                {
                    "start": 605.1,
                    "end": 605.58,
                    "word": " 1",
                    "probability": 0.5322265625
                },
                {
                    "start": 605.58,
                    "end": 605.6,
                    "word": ",",
                    "probability": 0.5966796875
                },
                {
                    "start": 605.6,
                    "end": 605.78,
                    "word": " 1",
                    "probability": 0.544921875
                },
                {
                    "start": 605.78,
                    "end": 606.1,
                    "word": ".",
                    "probability": 0.155517578125
                },
                {
                    "start": 606.1,
                    "end": 606.36,
                    "word": " And",
                    "probability": 0.8671875
                },
                {
                    "start": 606.36,
                    "end": 606.46,
                    "word": " the",
                    "probability": 0.81005859375
                },
                {
                    "start": 606.46,
                    "end": 606.76,
                    "word": " covariance",
                    "probability": 0.82470703125
                },
                {
                    "start": 606.76,
                    "end": 607.04,
                    "word": " matrix",
                    "probability": 0.85595703125
                },
                {
                    "start": 607.04,
                    "end": 607.22,
                    "word": " with",
                    "probability": 0.66015625
                },
                {
                    "start": 607.22,
                    "end": 607.44,
                    "word": " blues",
                    "probability": 0.6142578125
                },
                {
                    "start": 607.44,
                    "end": 607.58,
                    "word": " is",
                    "probability": 0.90087890625
                },
                {
                    "start": 607.58,
                    "end": 607.76,
                    "word": " 1",
                    "probability": 0.98828125
                },
                {
                    "start": 607.76,
                    "end": 607.78,
                    "word": ",",
                    "probability": 0.97607421875
                },
                {
                    "start": 607.78,
                    "end": 607.92,
                    "word": " 1",
                    "probability": 0.9345703125
                },
                {
                    "start": 607.92,
                    "end": 608.0,
                    "word": ".",
                    "probability": 0.62109375
                },
                {
                    "start": 608.0,
                    "end": 608.08,
                    "word": " And",
                    "probability": 0.8916015625
                },
                {
                    "start": 608.08,
                    "end": 608.18,
                    "word": " it",
                    "probability": 0.8935546875
                }
            ]
        },
        {
            "id": 79,
            "seek": 48336,
            "start": 608.18,
            "end": 612.56,
            "text": " gives you this nice distribution. And it looks okay. Now, what I'm going to do is I'm going to",
            "tokens": [
                51072,
                2709,
                291,
                341,
                1481,
                7316,
                13,
                400,
                309,
                1542,
                1392,
                13,
                823,
                11,
                437,
                286,
                478,
                516,
                281,
                360,
                307,
                286,
                478,
                516,
                281,
                51304
            ],
            "temperature": 0.0,
            "avg_logprob": -0.24711538415688736,
            "compression_ratio": 1.750915750915751,
            "no_speech_prob": 0.381103515625,
            "words": [
                {
                    "start": 608.18,
                    "end": 608.3,
                    "word": " gives",
                    "probability": 0.85009765625
                },
                {
                    "start": 608.3,
                    "end": 608.4,
                    "word": " you",
                    "probability": 0.89013671875
                },
                {
                    "start": 608.4,
                    "end": 608.54,
                    "word": " this",
                    "probability": 0.8798828125
                },
                {
                    "start": 608.54,
                    "end": 608.78,
                    "word": " nice",
                    "probability": 0.84814453125
                },
                {
                    "start": 608.78,
                    "end": 609.34,
                    "word": " distribution",
                    "probability": 0.7197265625
                },
                {
                    "start": 609.34,
                    "end": 609.44,
                    "word": ".",
                    "probability": 0.299560546875
                },
                {
                    "start": 609.44,
                    "end": 609.56,
                    "word": " And",
                    "probability": 0.89501953125
                },
                {
                    "start": 609.56,
                    "end": 609.7,
                    "word": " it",
                    "probability": 0.892578125
                },
                {
                    "start": 609.7,
                    "end": 610.02,
                    "word": " looks",
                    "probability": 0.78759765625
                },
                {
                    "start": 610.02,
                    "end": 610.74,
                    "word": " okay",
                    "probability": 0.361083984375
                },
                {
                    "start": 610.74,
                    "end": 611.12,
                    "word": ".",
                    "probability": 0.982421875
                },
                {
                    "start": 611.54,
                    "end": 611.66,
                    "word": " Now",
                    "probability": 0.64599609375
                },
                {
                    "start": 611.66,
                    "end": 611.66,
                    "word": ",",
                    "probability": 0.398681640625
                },
                {
                    "start": 611.66,
                    "end": 611.74,
                    "word": " what",
                    "probability": 0.8837890625
                },
                {
                    "start": 611.74,
                    "end": 611.84,
                    "word": " I'm",
                    "probability": 0.99365234375
                },
                {
                    "start": 611.84,
                    "end": 611.92,
                    "word": " going",
                    "probability": 0.759765625
                },
                {
                    "start": 611.92,
                    "end": 612.0,
                    "word": " to",
                    "probability": 0.9091796875
                },
                {
                    "start": 612.0,
                    "end": 612.16,
                    "word": " do",
                    "probability": 0.9091796875
                },
                {
                    "start": 612.16,
                    "end": 612.28,
                    "word": " is",
                    "probability": 0.8505859375
                },
                {
                    "start": 612.28,
                    "end": 612.4,
                    "word": " I'm",
                    "probability": 0.89404296875
                },
                {
                    "start": 612.4,
                    "end": 612.46,
                    "word": " going",
                    "probability": 0.8349609375
                },
                {
                    "start": 612.46,
                    "end": 612.56,
                    "word": " to",
                    "probability": 0.8994140625
                }
            ]
        },
        {
            "id": 80,
            "seek": 48336,
            "start": 612.56,
            "end": 619.22,
            "text": " drift the blues slightly over. And can you tell, is this correct? Well, you know, every single time",
            "tokens": [
                51304,
                19699,
                264,
                24244,
                4748,
                670,
                13,
                400,
                393,
                291,
                980,
                11,
                307,
                341,
                3006,
                30,
                1042,
                11,
                291,
                458,
                11,
                633,
                2167,
                565,
                51624
            ],
            "temperature": 0.0,
            "avg_logprob": -0.24711538415688736,
            "compression_ratio": 1.750915750915751,
            "no_speech_prob": 0.381103515625,
            "words": [
                {
                    "start": 612.56,
                    "end": 613.04,
                    "word": " drift",
                    "probability": 0.67236328125
                },
                {
                    "start": 613.04,
                    "end": 613.46,
                    "word": " the",
                    "probability": 0.81005859375
                },
                {
                    "start": 613.46,
                    "end": 614.1,
                    "word": " blues",
                    "probability": 0.81005859375
                },
                {
                    "start": 614.1,
                    "end": 614.6,
                    "word": " slightly",
                    "probability": 0.75537109375
                },
                {
                    "start": 614.6,
                    "end": 614.98,
                    "word": " over",
                    "probability": 0.88427734375
                },
                {
                    "start": 614.98,
                    "end": 615.36,
                    "word": ".",
                    "probability": 0.79150390625
                },
                {
                    "start": 615.54,
                    "end": 615.92,
                    "word": " And",
                    "probability": 0.87841796875
                },
                {
                    "start": 615.92,
                    "end": 616.08,
                    "word": " can",
                    "probability": 0.85498046875
                },
                {
                    "start": 616.08,
                    "end": 616.18,
                    "word": " you",
                    "probability": 0.896484375
                },
                {
                    "start": 616.18,
                    "end": 616.4,
                    "word": " tell",
                    "probability": 0.83349609375
                },
                {
                    "start": 616.4,
                    "end": 616.62,
                    "word": ",",
                    "probability": 0.462646484375
                },
                {
                    "start": 616.62,
                    "end": 616.8,
                    "word": " is",
                    "probability": 0.892578125
                },
                {
                    "start": 616.8,
                    "end": 616.94,
                    "word": " this",
                    "probability": 0.87890625
                },
                {
                    "start": 616.94,
                    "end": 617.3,
                    "word": " correct",
                    "probability": 0.86865234375
                },
                {
                    "start": 617.3,
                    "end": 617.48,
                    "word": "?",
                    "probability": 0.9892578125
                },
                {
                    "start": 617.48,
                    "end": 617.68,
                    "word": " Well",
                    "probability": 0.7353515625
                },
                {
                    "start": 617.68,
                    "end": 618.22,
                    "word": ",",
                    "probability": 0.95458984375
                },
                {
                    "start": 618.22,
                    "end": 618.32,
                    "word": " you",
                    "probability": 0.787109375
                },
                {
                    "start": 618.32,
                    "end": 618.44,
                    "word": " know",
                    "probability": 0.81689453125
                },
                {
                    "start": 618.44,
                    "end": 618.48,
                    "word": ",",
                    "probability": 0.9521484375
                },
                {
                    "start": 618.48,
                    "end": 618.7,
                    "word": " every",
                    "probability": 0.7529296875
                },
                {
                    "start": 618.7,
                    "end": 618.92,
                    "word": " single",
                    "probability": 0.8857421875
                },
                {
                    "start": 618.92,
                    "end": 619.22,
                    "word": " time",
                    "probability": 0.80908203125
                }
            ]
        },
        {
            "id": 81,
            "seek": 48336,
            "start": 619.22,
            "end": 623.58,
            "text": " that I put a blue, it's right next to some orange points. So this looks like it could be from the",
            "tokens": [
                51624,
                300,
                286,
                829,
                257,
                3344,
                11,
                309,
                311,
                558,
                958,
                281,
                512,
                7671,
                2793,
                13,
                407,
                341,
                1542,
                411,
                309,
                727,
                312,
                490,
                264,
                51844
            ],
            "temperature": 0.0,
            "avg_logprob": -0.24711538415688736,
            "compression_ratio": 1.750915750915751,
            "no_speech_prob": 0.381103515625,
            "words": [
                {
                    "start": 619.22,
                    "end": 619.34,
                    "word": " that",
                    "probability": 0.8779296875
                },
                {
                    "start": 619.34,
                    "end": 619.48,
                    "word": " I",
                    "probability": 0.998046875
                },
                {
                    "start": 619.48,
                    "end": 619.6,
                    "word": " put",
                    "probability": 0.67529296875
                },
                {
                    "start": 619.72,
                    "end": 619.96,
                    "word": " a blue",
                    "probability": 0.9130859375
                },
                {
                    "start": 619.96,
                    "end": 620.14,
                    "word": ",",
                    "probability": 0.57568359375
                },
                {
                    "start": 620.14,
                    "end": 620.36,
                    "word": " it's",
                    "probability": 0.935546875
                },
                {
                    "start": 620.36,
                    "end": 620.58,
                    "word": " right",
                    "probability": 0.8447265625
                },
                {
                    "start": 620.58,
                    "end": 620.92,
                    "word": " next",
                    "probability": 0.84521484375
                },
                {
                    "start": 620.92,
                    "end": 621.06,
                    "word": " to",
                    "probability": 0.90673828125
                },
                {
                    "start": 621.06,
                    "end": 621.26,
                    "word": " some",
                    "probability": 0.82373046875
                },
                {
                    "start": 621.26,
                    "end": 621.48,
                    "word": " orange",
                    "probability": 0.85107421875
                },
                {
                    "start": 621.48,
                    "end": 621.82,
                    "word": " points",
                    "probability": 0.442138671875
                },
                {
                    "start": 621.82,
                    "end": 621.94,
                    "word": ".",
                    "probability": 0.8798828125
                },
                {
                    "start": 621.94,
                    "end": 622.0,
                    "word": " So",
                    "probability": 0.8837890625
                },
                {
                    "start": 622.0,
                    "end": 622.14,
                    "word": " this",
                    "probability": 0.73876953125
                },
                {
                    "start": 622.14,
                    "end": 622.48,
                    "word": " looks",
                    "probability": 0.78662109375
                },
                {
                    "start": 622.48,
                    "end": 622.74,
                    "word": " like",
                    "probability": 0.876953125
                },
                {
                    "start": 622.74,
                    "end": 622.88,
                    "word": " it",
                    "probability": 0.8857421875
                },
                {
                    "start": 622.88,
                    "end": 623.04,
                    "word": " could",
                    "probability": 0.81396484375
                },
                {
                    "start": 623.04,
                    "end": 623.24,
                    "word": " be",
                    "probability": 0.9013671875
                },
                {
                    "start": 623.24,
                    "end": 623.44,
                    "word": " from",
                    "probability": 0.8251953125
                },
                {
                    "start": 623.44,
                    "end": 623.58,
                    "word": " the",
                    "probability": 0.755859375
                }
            ]
        },
        {
            "id": 82,
            "seek": 50998,
            "start": 623.58,
            "end": 628.5,
            "text": " distribution. I can't really tell with my eyes. And one of the techniques we might do is like,",
            "tokens": [
                50372,
                7316,
                13,
                286,
                393,
                380,
                534,
                980,
                365,
                452,
                2575,
                13,
                400,
                472,
                295,
                264,
                7512,
                321,
                1062,
                360,
                307,
                411,
                11,
                50604
            ],
            "temperature": 0.0,
            "avg_logprob": -0.19969224004131375,
            "compression_ratio": 1.8181818181818181,
            "no_speech_prob": 0.06280517578125,
            "words": [
                {
                    "start": 623.58,
                    "end": 624.02,
                    "word": " distribution",
                    "probability": 0.1318359375
                },
                {
                    "start": 624.02,
                    "end": 624.38,
                    "word": ".",
                    "probability": 0.5361328125
                },
                {
                    "start": 625.12,
                    "end": 625.48,
                    "word": " I",
                    "probability": 0.974609375
                },
                {
                    "start": 625.48,
                    "end": 625.72,
                    "word": " can't",
                    "probability": 0.950927734375
                },
                {
                    "start": 625.72,
                    "end": 625.92,
                    "word": " really",
                    "probability": 0.78125
                },
                {
                    "start": 625.92,
                    "end": 626.12,
                    "word": " tell",
                    "probability": 0.8193359375
                },
                {
                    "start": 626.12,
                    "end": 626.24,
                    "word": " with",
                    "probability": 0.8134765625
                },
                {
                    "start": 626.24,
                    "end": 626.4,
                    "word": " my",
                    "probability": 0.89501953125
                },
                {
                    "start": 626.4,
                    "end": 626.68,
                    "word": " eyes",
                    "probability": 0.81591796875
                },
                {
                    "start": 626.68,
                    "end": 626.8,
                    "word": ".",
                    "probability": 0.34814453125
                },
                {
                    "start": 626.8,
                    "end": 626.86,
                    "word": " And",
                    "probability": 0.498046875
                },
                {
                    "start": 626.86,
                    "end": 627.16,
                    "word": " one",
                    "probability": 0.810546875
                },
                {
                    "start": 627.16,
                    "end": 627.26,
                    "word": " of",
                    "probability": 0.9013671875
                },
                {
                    "start": 627.26,
                    "end": 627.4,
                    "word": " the",
                    "probability": 0.81640625
                },
                {
                    "start": 627.4,
                    "end": 627.72,
                    "word": " techniques",
                    "probability": 0.83154296875
                },
                {
                    "start": 627.72,
                    "end": 627.9,
                    "word": " we",
                    "probability": 0.8955078125
                },
                {
                    "start": 627.9,
                    "end": 628.08,
                    "word": " might",
                    "probability": 0.80810546875
                },
                {
                    "start": 628.08,
                    "end": 628.22,
                    "word": " do",
                    "probability": 0.90185546875
                },
                {
                    "start": 628.22,
                    "end": 628.32,
                    "word": " is",
                    "probability": 0.87841796875
                },
                {
                    "start": 628.32,
                    "end": 628.44,
                    "word": " like",
                    "probability": 0.73095703125
                },
                {
                    "start": 628.44,
                    "end": 628.5,
                    "word": ",",
                    "probability": 0.5966796875
                }
            ]
        },
        {
            "id": 83,
            "seek": 50998,
            "start": 628.5,
            "end": 633.04,
            "text": " oh, well, let me take the k-nearest neighbors and see if the distance from my k-nearest neighbors",
            "tokens": [
                50604,
                1954,
                11,
                731,
                11,
                718,
                385,
                747,
                264,
                350,
                12,
                716,
                17363,
                12512,
                293,
                536,
                498,
                264,
                4560,
                490,
                452,
                350,
                12,
                716,
                17363,
                12512,
                50840
            ],
            "temperature": 0.0,
            "avg_logprob": -0.19969224004131375,
            "compression_ratio": 1.8181818181818181,
            "no_speech_prob": 0.06280517578125,
            "words": [
                {
                    "start": 628.5,
                    "end": 628.62,
                    "word": " oh",
                    "probability": 0.52734375
                },
                {
                    "start": 628.62,
                    "end": 628.7,
                    "word": ",",
                    "probability": 0.50927734375
                },
                {
                    "start": 628.7,
                    "end": 628.84,
                    "word": " well",
                    "probability": 0.873046875
                },
                {
                    "start": 628.84,
                    "end": 628.86,
                    "word": ",",
                    "probability": 0.703125
                },
                {
                    "start": 628.86,
                    "end": 629.0,
                    "word": " let",
                    "probability": 0.68994140625
                },
                {
                    "start": 629.0,
                    "end": 629.06,
                    "word": " me",
                    "probability": 0.89892578125
                },
                {
                    "start": 629.06,
                    "end": 629.34,
                    "word": " take",
                    "probability": 0.78515625
                },
                {
                    "start": 629.34,
                    "end": 629.5,
                    "word": " the",
                    "probability": 0.818359375
                },
                {
                    "start": 629.5,
                    "end": 629.7,
                    "word": " k",
                    "probability": 0.79541015625
                },
                {
                    "start": 629.7,
                    "end": 629.96,
                    "word": "-nearest",
                    "probability": 0.6653645833333334
                },
                {
                    "start": 629.96,
                    "end": 630.24,
                    "word": " neighbors",
                    "probability": 0.6533203125
                },
                {
                    "start": 630.24,
                    "end": 630.48,
                    "word": " and",
                    "probability": 0.8671875
                },
                {
                    "start": 630.48,
                    "end": 630.66,
                    "word": " see",
                    "probability": 0.8037109375
                },
                {
                    "start": 630.66,
                    "end": 630.84,
                    "word": " if",
                    "probability": 0.90234375
                },
                {
                    "start": 630.84,
                    "end": 631.12,
                    "word": " the",
                    "probability": 0.80908203125
                },
                {
                    "start": 631.12,
                    "end": 632.02,
                    "word": " distance",
                    "probability": 0.408447265625
                },
                {
                    "start": 632.02,
                    "end": 632.22,
                    "word": " from",
                    "probability": 0.66259765625
                },
                {
                    "start": 632.22,
                    "end": 632.36,
                    "word": " my",
                    "probability": 0.837890625
                },
                {
                    "start": 632.36,
                    "end": 632.5,
                    "word": " k",
                    "probability": 0.98095703125
                },
                {
                    "start": 632.5,
                    "end": 632.68,
                    "word": "-nearest",
                    "probability": 0.89599609375
                },
                {
                    "start": 632.68,
                    "end": 633.04,
                    "word": " neighbors",
                    "probability": 0.81396484375
                }
            ]
        },
        {
            "id": 84,
            "seek": 50998,
            "start": 633.64,
            "end": 639.14,
            "text": " over time goes up, then I'm out of distribution. But if you tell here, that's not going to tell",
            "tokens": [
                50868,
                670,
                565,
                1709,
                493,
                11,
                550,
                286,
                478,
                484,
                295,
                7316,
                13,
                583,
                498,
                291,
                980,
                510,
                11,
                300,
                311,
                406,
                516,
                281,
                980,
                51140
            ],
            "temperature": 0.0,
            "avg_logprob": -0.19969224004131375,
            "compression_ratio": 1.8181818181818181,
            "no_speech_prob": 0.06280517578125,
            "words": [
                {
                    "start": 633.64,
                    "end": 634.0,
                    "word": " over",
                    "probability": 0.77001953125
                },
                {
                    "start": 634.0,
                    "end": 634.36,
                    "word": " time",
                    "probability": 0.8017578125
                },
                {
                    "start": 634.36,
                    "end": 634.76,
                    "word": " goes",
                    "probability": 0.837890625
                },
                {
                    "start": 634.76,
                    "end": 635.08,
                    "word": " up",
                    "probability": 0.912109375
                },
                {
                    "start": 635.08,
                    "end": 635.2,
                    "word": ",",
                    "probability": 0.489990234375
                },
                {
                    "start": 635.2,
                    "end": 635.4,
                    "word": " then",
                    "probability": 0.48291015625
                },
                {
                    "start": 635.4,
                    "end": 635.82,
                    "word": " I'm",
                    "probability": 0.98291015625
                },
                {
                    "start": 635.82,
                    "end": 635.96,
                    "word": " out",
                    "probability": 0.82373046875
                },
                {
                    "start": 635.96,
                    "end": 636.06,
                    "word": " of",
                    "probability": 0.88671875
                },
                {
                    "start": 636.06,
                    "end": 636.52,
                    "word": " distribution",
                    "probability": 0.7353515625
                },
                {
                    "start": 636.52,
                    "end": 636.7,
                    "word": ".",
                    "probability": 0.91357421875
                },
                {
                    "start": 636.7,
                    "end": 636.86,
                    "word": " But",
                    "probability": 0.8701171875
                },
                {
                    "start": 636.86,
                    "end": 637.1,
                    "word": " if",
                    "probability": 0.88232421875
                },
                {
                    "start": 637.1,
                    "end": 637.2,
                    "word": " you",
                    "probability": 0.896484375
                },
                {
                    "start": 637.2,
                    "end": 637.56,
                    "word": " tell",
                    "probability": 0.79931640625
                },
                {
                    "start": 637.56,
                    "end": 637.94,
                    "word": " here",
                    "probability": 0.81103515625
                },
                {
                    "start": 637.94,
                    "end": 638.14,
                    "word": ",",
                    "probability": 0.87939453125
                },
                {
                    "start": 638.14,
                    "end": 638.44,
                    "word": " that's",
                    "probability": 0.9345703125
                },
                {
                    "start": 638.44,
                    "end": 638.64,
                    "word": " not",
                    "probability": 0.8994140625
                },
                {
                    "start": 638.64,
                    "end": 638.78,
                    "word": " going",
                    "probability": 0.399169921875
                },
                {
                    "start": 638.78,
                    "end": 638.9,
                    "word": " to",
                    "probability": 0.90576171875
                },
                {
                    "start": 638.9,
                    "end": 639.14,
                    "word": " tell",
                    "probability": 0.83056640625
                }
            ]
        },
        {
            "id": 85,
            "seek": 50998,
            "start": 639.14,
            "end": 646.0,
            "text": " you much. If I go here, well, oh, I've got some outliers here now. So maybe that will work now.",
            "tokens": [
                51140,
                291,
                709,
                13,
                759,
                286,
                352,
                510,
                11,
                731,
                11,
                1954,
                11,
                286,
                600,
                658,
                512,
                484,
                23646,
                510,
                586,
                13,
                407,
                1310,
                300,
                486,
                589,
                586,
                13,
                51480
            ],
            "temperature": 0.0,
            "avg_logprob": -0.19969224004131375,
            "compression_ratio": 1.8181818181818181,
            "no_speech_prob": 0.06280517578125,
            "words": [
                {
                    "start": 639.14,
                    "end": 639.32,
                    "word": " you",
                    "probability": 0.89404296875
                },
                {
                    "start": 639.32,
                    "end": 639.64,
                    "word": " much",
                    "probability": 0.81396484375
                },
                {
                    "start": 639.64,
                    "end": 640.0,
                    "word": ".",
                    "probability": 0.98583984375
                },
                {
                    "start": 640.78,
                    "end": 640.9,
                    "word": " If",
                    "probability": 0.560546875
                },
                {
                    "start": 640.9,
                    "end": 640.94,
                    "word": " I",
                    "probability": 0.921875
                },
                {
                    "start": 640.94,
                    "end": 641.08,
                    "word": " go",
                    "probability": 0.90966796875
                },
                {
                    "start": 641.08,
                    "end": 641.34,
                    "word": " here",
                    "probability": 0.814453125
                },
                {
                    "start": 641.34,
                    "end": 641.48,
                    "word": ",",
                    "probability": 0.7451171875
                },
                {
                    "start": 641.48,
                    "end": 641.76,
                    "word": " well",
                    "probability": 0.83984375
                },
                {
                    "start": 641.76,
                    "end": 641.9,
                    "word": ",",
                    "probability": 0.9140625
                },
                {
                    "start": 641.9,
                    "end": 642.08,
                    "word": " oh",
                    "probability": 0.343994140625
                },
                {
                    "start": 642.08,
                    "end": 642.1,
                    "word": ",",
                    "probability": 0.8310546875
                },
                {
                    "start": 642.1,
                    "end": 642.26,
                    "word": " I've",
                    "probability": 0.9169921875
                },
                {
                    "start": 642.26,
                    "end": 642.38,
                    "word": " got",
                    "probability": 0.89599609375
                },
                {
                    "start": 642.38,
                    "end": 642.68,
                    "word": " some",
                    "probability": 0.8291015625
                },
                {
                    "start": 642.68,
                    "end": 643.18,
                    "word": " outliers",
                    "probability": 0.886962890625
                },
                {
                    "start": 643.18,
                    "end": 643.44,
                    "word": " here",
                    "probability": 0.81103515625
                },
                {
                    "start": 643.44,
                    "end": 643.78,
                    "word": " now",
                    "probability": 0.8916015625
                },
                {
                    "start": 643.78,
                    "end": 643.94,
                    "word": ".",
                    "probability": 0.411376953125
                },
                {
                    "start": 643.94,
                    "end": 644.18,
                    "word": " So",
                    "probability": 0.89892578125
                },
                {
                    "start": 644.18,
                    "end": 644.64,
                    "word": " maybe",
                    "probability": 0.80859375
                },
                {
                    "start": 644.64,
                    "end": 644.9,
                    "word": " that",
                    "probability": 0.88525390625
                },
                {
                    "start": 644.9,
                    "end": 645.02,
                    "word": " will",
                    "probability": 0.8154296875
                },
                {
                    "start": 645.02,
                    "end": 645.26,
                    "word": " work",
                    "probability": 0.81494140625
                },
                {
                    "start": 645.26,
                    "end": 645.64,
                    "word": " now",
                    "probability": 0.84814453125
                },
                {
                    "start": 645.64,
                    "end": 646.0,
                    "word": ".",
                    "probability": 0.97412109375
                }
            ]
        },
        {
            "id": 86,
            "seek": 50998,
            "start": 646.06,
            "end": 650.2,
            "text": " If I go here, well, this is really distributed. So this center, the center of this distribution",
            "tokens": [
                51480,
                759,
                286,
                352,
                510,
                11,
                731,
                11,
                341,
                307,
                534,
                12631,
                13,
                407,
                341,
                3056,
                11,
                264,
                3056,
                295,
                341,
                7316,
                51700
            ],
            "temperature": 0.0,
            "avg_logprob": -0.19969224004131375,
            "compression_ratio": 1.8181818181818181,
            "no_speech_prob": 0.06280517578125,
            "words": [
                {
                    "start": 646.06,
                    "end": 646.2,
                    "word": " If",
                    "probability": 0.84912109375
                },
                {
                    "start": 646.2,
                    "end": 646.24,
                    "word": " I",
                    "probability": 0.9228515625
                },
                {
                    "start": 646.24,
                    "end": 646.34,
                    "word": " go",
                    "probability": 0.91064453125
                },
                {
                    "start": 646.34,
                    "end": 646.62,
                    "word": " here",
                    "probability": 0.81396484375
                },
                {
                    "start": 646.62,
                    "end": 646.68,
                    "word": ",",
                    "probability": 0.92626953125
                },
                {
                    "start": 646.68,
                    "end": 646.78,
                    "word": " well",
                    "probability": 0.82666015625
                },
                {
                    "start": 646.78,
                    "end": 646.82,
                    "word": ",",
                    "probability": 0.90576171875
                },
                {
                    "start": 646.82,
                    "end": 647.04,
                    "word": " this",
                    "probability": 0.875
                },
                {
                    "start": 647.04,
                    "end": 647.16,
                    "word": " is",
                    "probability": 0.91162109375
                },
                {
                    "start": 647.16,
                    "end": 647.44,
                    "word": " really",
                    "probability": 0.78662109375
                },
                {
                    "start": 647.44,
                    "end": 647.94,
                    "word": " distributed",
                    "probability": 0.82373046875
                },
                {
                    "start": 647.94,
                    "end": 648.08,
                    "word": ".",
                    "probability": 0.767578125
                },
                {
                    "start": 648.08,
                    "end": 648.18,
                    "word": " So",
                    "probability": 0.88330078125
                },
                {
                    "start": 648.18,
                    "end": 648.4,
                    "word": " this",
                    "probability": 0.48388671875
                },
                {
                    "start": 648.4,
                    "end": 648.84,
                    "word": " center",
                    "probability": 0.57958984375
                },
                {
                    "start": 648.84,
                    "end": 648.96,
                    "word": ",",
                    "probability": 0.62548828125
                },
                {
                    "start": 648.96,
                    "end": 649.12,
                    "word": " the",
                    "probability": 0.81591796875
                },
                {
                    "start": 649.12,
                    "end": 649.34,
                    "word": " center",
                    "probability": 0.79296875
                },
                {
                    "start": 649.34,
                    "end": 649.5,
                    "word": " of",
                    "probability": 0.904296875
                },
                {
                    "start": 649.5,
                    "end": 649.74,
                    "word": " this",
                    "probability": 0.8798828125
                },
                {
                    "start": 649.74,
                    "end": 650.2,
                    "word": " distribution",
                    "probability": 0.74072265625
                }
            ]
        },
        {
            "id": 87,
            "seek": 53908,
            "start": 650.2,
            "end": 653.96,
            "text": " should be around here. And the center for this distribution is around here. So that's really",
            "tokens": [
                50368,
                820,
                312,
                926,
                510,
                13,
                400,
                264,
                3056,
                337,
                341,
                7316,
                307,
                926,
                510,
                13,
                407,
                300,
                311,
                534,
                50552
            ],
            "temperature": 0.0,
            "avg_logprob": -0.21383304904390882,
            "compression_ratio": 1.7492354740061162,
            "no_speech_prob": 0.00669097900390625,
            "words": [
                {
                    "start": 650.2,
                    "end": 650.48,
                    "word": " should",
                    "probability": 0.6455078125
                },
                {
                    "start": 650.48,
                    "end": 650.66,
                    "word": " be",
                    "probability": 0.90380859375
                },
                {
                    "start": 650.66,
                    "end": 650.9,
                    "word": " around",
                    "probability": 0.8759765625
                },
                {
                    "start": 650.9,
                    "end": 651.2,
                    "word": " here",
                    "probability": 0.8134765625
                },
                {
                    "start": 651.2,
                    "end": 651.3,
                    "word": ".",
                    "probability": 0.032989501953125
                },
                {
                    "start": 651.3,
                    "end": 651.34,
                    "word": " And",
                    "probability": 0.78662109375
                },
                {
                    "start": 651.34,
                    "end": 651.44,
                    "word": " the",
                    "probability": 0.8154296875
                },
                {
                    "start": 651.44,
                    "end": 651.62,
                    "word": " center",
                    "probability": 0.71630859375
                },
                {
                    "start": 651.62,
                    "end": 651.82,
                    "word": " for",
                    "probability": 0.88720703125
                },
                {
                    "start": 651.82,
                    "end": 651.98,
                    "word": " this",
                    "probability": 0.87548828125
                },
                {
                    "start": 651.98,
                    "end": 652.4,
                    "word": " distribution",
                    "probability": 0.71875
                },
                {
                    "start": 652.4,
                    "end": 652.62,
                    "word": " is",
                    "probability": 0.90234375
                },
                {
                    "start": 652.62,
                    "end": 652.78,
                    "word": " around",
                    "probability": 0.8447265625
                },
                {
                    "start": 652.78,
                    "end": 653.04,
                    "word": " here",
                    "probability": 0.81201171875
                },
                {
                    "start": 653.04,
                    "end": 653.1,
                    "word": ".",
                    "probability": 0.7158203125
                },
                {
                    "start": 653.1,
                    "end": 653.46,
                    "word": " So",
                    "probability": 0.89306640625
                },
                {
                    "start": 653.46,
                    "end": 653.64,
                    "word": " that's",
                    "probability": 0.92041015625
                },
                {
                    "start": 653.64,
                    "end": 653.96,
                    "word": " really",
                    "probability": 0.78271484375
                }
            ]
        },
        {
            "id": 88,
            "seek": 53908,
            "start": 653.96,
            "end": 659.48,
            "text": " drifted. But you've got some outliers over here. Now, what happens when there's like millions of",
            "tokens": [
                50552,
                19699,
                292,
                13,
                583,
                291,
                600,
                658,
                512,
                484,
                23646,
                670,
                510,
                13,
                823,
                11,
                437,
                2314,
                562,
                456,
                311,
                411,
                6803,
                295,
                50828
            ],
            "temperature": 0.0,
            "avg_logprob": -0.21383304904390882,
            "compression_ratio": 1.7492354740061162,
            "no_speech_prob": 0.00669097900390625,
            "words": [
                {
                    "start": 653.96,
                    "end": 654.48,
                    "word": " drifted",
                    "probability": 0.896240234375
                },
                {
                    "start": 654.48,
                    "end": 654.76,
                    "word": ".",
                    "probability": 0.7392578125
                },
                {
                    "start": 654.76,
                    "end": 655.06,
                    "word": " But",
                    "probability": 0.88330078125
                },
                {
                    "start": 655.06,
                    "end": 655.32,
                    "word": " you've",
                    "probability": 0.911376953125
                },
                {
                    "start": 655.32,
                    "end": 655.46,
                    "word": " got",
                    "probability": 0.89501953125
                },
                {
                    "start": 655.46,
                    "end": 655.7,
                    "word": " some",
                    "probability": 0.8203125
                },
                {
                    "start": 655.7,
                    "end": 656.04,
                    "word": " outliers",
                    "probability": 0.890380859375
                },
                {
                    "start": 656.04,
                    "end": 656.26,
                    "word": " over",
                    "probability": 0.88525390625
                },
                {
                    "start": 656.26,
                    "end": 656.46,
                    "word": " here",
                    "probability": 0.8095703125
                },
                {
                    "start": 656.46,
                    "end": 656.78,
                    "word": ".",
                    "probability": 0.935546875
                },
                {
                    "start": 656.78,
                    "end": 657.06,
                    "word": " Now",
                    "probability": 0.8798828125
                },
                {
                    "start": 657.06,
                    "end": 657.1,
                    "word": ",",
                    "probability": 0.17236328125
                },
                {
                    "start": 657.1,
                    "end": 657.18,
                    "word": " what",
                    "probability": 0.88525390625
                },
                {
                    "start": 657.18,
                    "end": 657.52,
                    "word": " happens",
                    "probability": 0.91650390625
                },
                {
                    "start": 657.52,
                    "end": 657.66,
                    "word": " when",
                    "probability": 0.87060546875
                },
                {
                    "start": 657.66,
                    "end": 657.82,
                    "word": " there's",
                    "probability": 0.91845703125
                },
                {
                    "start": 657.82,
                    "end": 658.02,
                    "word": " like",
                    "probability": 0.5498046875
                },
                {
                    "start": 658.02,
                    "end": 658.82,
                    "word": " millions",
                    "probability": 0.80126953125
                },
                {
                    "start": 658.82,
                    "end": 659.48,
                    "word": " of",
                    "probability": 0.91015625
                }
            ]
        },
        {
            "id": 89,
            "seek": 53908,
            "start": 659.48,
            "end": 663.76,
            "text": " orange points and there's only a few hundred blue points? Well, you might not be able to tell the",
            "tokens": [
                50828,
                7671,
                2793,
                293,
                456,
                311,
                787,
                257,
                1326,
                3262,
                3344,
                2793,
                30,
                1042,
                11,
                291,
                1062,
                406,
                312,
                1075,
                281,
                980,
                264,
                51036
            ],
            "temperature": 0.0,
            "avg_logprob": -0.21383304904390882,
            "compression_ratio": 1.7492354740061162,
            "no_speech_prob": 0.00669097900390625,
            "words": [
                {
                    "start": 659.48,
                    "end": 659.9,
                    "word": " orange",
                    "probability": 0.82958984375
                },
                {
                    "start": 659.9,
                    "end": 660.3,
                    "word": " points",
                    "probability": 0.77294921875
                },
                {
                    "start": 660.3,
                    "end": 660.46,
                    "word": " and",
                    "probability": 0.673828125
                },
                {
                    "start": 660.46,
                    "end": 660.62,
                    "word": " there's",
                    "probability": 0.92529296875
                },
                {
                    "start": 660.62,
                    "end": 660.84,
                    "word": " only",
                    "probability": 0.85205078125
                },
                {
                    "start": 661.1,
                    "end": 661.32,
                    "word": " a few",
                    "probability": 0.8125
                },
                {
                    "start": 661.32,
                    "end": 661.72,
                    "word": " hundred",
                    "probability": 0.84716796875
                },
                {
                    "start": 661.72,
                    "end": 662.04,
                    "word": " blue",
                    "probability": 0.91064453125
                },
                {
                    "start": 662.04,
                    "end": 662.38,
                    "word": " points",
                    "probability": 0.76513671875
                },
                {
                    "start": 662.38,
                    "end": 662.46,
                    "word": "?",
                    "probability": 0.83447265625
                },
                {
                    "start": 662.46,
                    "end": 662.58,
                    "word": " Well",
                    "probability": 0.744140625
                },
                {
                    "start": 662.58,
                    "end": 662.6,
                    "word": ",",
                    "probability": 0.76708984375
                },
                {
                    "start": 662.6,
                    "end": 662.74,
                    "word": " you",
                    "probability": 0.896484375
                },
                {
                    "start": 662.74,
                    "end": 662.94,
                    "word": " might",
                    "probability": 0.79541015625
                },
                {
                    "start": 662.94,
                    "end": 663.06,
                    "word": " not",
                    "probability": 0.90087890625
                },
                {
                    "start": 663.06,
                    "end": 663.18,
                    "word": " be",
                    "probability": 0.904296875
                },
                {
                    "start": 663.18,
                    "end": 663.26,
                    "word": " able",
                    "probability": 0.88525390625
                },
                {
                    "start": 663.26,
                    "end": 663.4,
                    "word": " to",
                    "probability": 0.904296875
                },
                {
                    "start": 663.4,
                    "end": 663.6,
                    "word": " tell",
                    "probability": 0.82080078125
                },
                {
                    "start": 663.6,
                    "end": 663.76,
                    "word": " the",
                    "probability": 0.81640625
                }
            ]
        },
        {
            "id": 90,
            "seek": 53908,
            "start": 663.76,
            "end": 670.06,
            "text": " difference with that in that case. So, and also like, there's all this, what I've described to",
            "tokens": [
                51036,
                2649,
                365,
                300,
                294,
                300,
                1389,
                13,
                407,
                11,
                293,
                611,
                411,
                11,
                456,
                311,
                439,
                341,
                11,
                437,
                286,
                600,
                7619,
                281,
                51348
            ],
            "temperature": 0.0,
            "avg_logprob": -0.21383304904390882,
            "compression_ratio": 1.7492354740061162,
            "no_speech_prob": 0.00669097900390625,
            "words": [
                {
                    "start": 663.76,
                    "end": 664.06,
                    "word": " difference",
                    "probability": 0.71142578125
                },
                {
                    "start": 664.06,
                    "end": 664.24,
                    "word": " with",
                    "probability": 0.591796875
                },
                {
                    "start": 664.24,
                    "end": 664.42,
                    "word": " that",
                    "probability": 0.8671875
                },
                {
                    "start": 664.42,
                    "end": 664.64,
                    "word": " in",
                    "probability": 0.59375
                },
                {
                    "start": 664.64,
                    "end": 664.8,
                    "word": " that",
                    "probability": 0.88232421875
                },
                {
                    "start": 664.8,
                    "end": 665.04,
                    "word": " case",
                    "probability": 0.81396484375
                },
                {
                    "start": 665.04,
                    "end": 665.44,
                    "word": ".",
                    "probability": 0.9794921875
                },
                {
                    "start": 665.86,
                    "end": 665.88,
                    "word": " So",
                    "probability": 0.703125
                },
                {
                    "start": 665.88,
                    "end": 666.52,
                    "word": ",",
                    "probability": 0.363037109375
                },
                {
                    "start": 666.52,
                    "end": 666.76,
                    "word": " and",
                    "probability": 0.74365234375
                },
                {
                    "start": 666.76,
                    "end": 666.98,
                    "word": " also",
                    "probability": 0.81787109375
                },
                {
                    "start": 666.98,
                    "end": 667.26,
                    "word": " like",
                    "probability": 0.55859375
                },
                {
                    "start": 667.26,
                    "end": 668.22,
                    "word": ",",
                    "probability": 0.336669921875
                },
                {
                    "start": 668.22,
                    "end": 668.42,
                    "word": " there's",
                    "probability": 0.703125
                },
                {
                    "start": 668.42,
                    "end": 668.64,
                    "word": " all",
                    "probability": 0.67236328125
                },
                {
                    "start": 668.64,
                    "end": 668.96,
                    "word": " this",
                    "probability": 0.77099609375
                },
                {
                    "start": 668.96,
                    "end": 669.14,
                    "word": ",",
                    "probability": 0.386474609375
                },
                {
                    "start": 669.14,
                    "end": 669.44,
                    "word": " what",
                    "probability": 0.8671875
                },
                {
                    "start": 669.44,
                    "end": 669.62,
                    "word": " I've",
                    "probability": 0.970458984375
                },
                {
                    "start": 669.62,
                    "end": 669.9,
                    "word": " described",
                    "probability": 0.818359375
                },
                {
                    "start": 669.9,
                    "end": 670.06,
                    "word": " to",
                    "probability": 0.904296875
                }
            ]
        },
        {
            "id": 91,
            "seek": 53908,
            "start": 670.06,
            "end": 674.02,
            "text": " you is kind of a feeling. And to actually get at the math of this thing, I have to like model the",
            "tokens": [
                51348,
                291,
                307,
                733,
                295,
                257,
                2633,
                13,
                400,
                281,
                767,
                483,
                412,
                264,
                5221,
                295,
                341,
                551,
                11,
                286,
                362,
                281,
                411,
                2316,
                264,
                51548
            ],
            "temperature": 0.0,
            "avg_logprob": -0.21383304904390882,
            "compression_ratio": 1.7492354740061162,
            "no_speech_prob": 0.00669097900390625,
            "words": [
                {
                    "start": 670.06,
                    "end": 670.12,
                    "word": " you",
                    "probability": 0.87890625
                },
                {
                    "start": 670.12,
                    "end": 670.18,
                    "word": " is",
                    "probability": 0.818359375
                },
                {
                    "start": 670.18,
                    "end": 670.36,
                    "word": " kind",
                    "probability": 0.80419921875
                },
                {
                    "start": 670.36,
                    "end": 670.44,
                    "word": " of",
                    "probability": 0.9072265625
                },
                {
                    "start": 670.58,
                    "end": 670.88,
                    "word": " a feeling",
                    "probability": 0.80224609375
                },
                {
                    "start": 670.88,
                    "end": 671.08,
                    "word": ".",
                    "probability": 0.346435546875
                },
                {
                    "start": 671.08,
                    "end": 671.18,
                    "word": " And",
                    "probability": 0.8740234375
                },
                {
                    "start": 671.18,
                    "end": 671.32,
                    "word": " to",
                    "probability": 0.59912109375
                },
                {
                    "start": 671.32,
                    "end": 671.52,
                    "word": " actually",
                    "probability": 0.77880859375
                },
                {
                    "start": 671.52,
                    "end": 671.7,
                    "word": " get",
                    "probability": 0.9033203125
                },
                {
                    "start": 671.7,
                    "end": 671.84,
                    "word": " at",
                    "probability": 0.86328125
                },
                {
                    "start": 671.84,
                    "end": 671.98,
                    "word": " the",
                    "probability": 0.818359375
                },
                {
                    "start": 671.98,
                    "end": 672.2,
                    "word": " math",
                    "probability": 0.87841796875
                },
                {
                    "start": 672.2,
                    "end": 672.32,
                    "word": " of",
                    "probability": 0.90185546875
                },
                {
                    "start": 672.32,
                    "end": 672.46,
                    "word": " this",
                    "probability": 0.88037109375
                },
                {
                    "start": 672.46,
                    "end": 672.6,
                    "word": " thing",
                    "probability": 0.81201171875
                },
                {
                    "start": 672.6,
                    "end": 672.62,
                    "word": ",",
                    "probability": 0.75439453125
                },
                {
                    "start": 672.62,
                    "end": 672.72,
                    "word": " I",
                    "probability": 0.99755859375
                },
                {
                    "start": 672.72,
                    "end": 672.86,
                    "word": " have",
                    "probability": 0.87939453125
                },
                {
                    "start": 672.86,
                    "end": 672.96,
                    "word": " to",
                    "probability": 0.9072265625
                },
                {
                    "start": 672.96,
                    "end": 673.14,
                    "word": " like",
                    "probability": 0.77099609375
                },
                {
                    "start": 673.14,
                    "end": 673.76,
                    "word": " model",
                    "probability": 0.712890625
                },
                {
                    "start": 673.76,
                    "end": 674.02,
                    "word": " the",
                    "probability": 0.81591796875
                }
            ]
        },
        {
            "id": 92,
            "seek": 53908,
            "start": 674.02,
            "end": 679.3,
            "text": " distributions that these came from. And that's quite a hard problem. And especially in high",
            "tokens": [
                51548,
                37870,
                300,
                613,
                1361,
                490,
                13,
                400,
                300,
                311,
                1596,
                257,
                1152,
                1154,
                13,
                400,
                2318,
                294,
                1090,
                51812
            ],
            "temperature": 0.0,
            "avg_logprob": -0.21383304904390882,
            "compression_ratio": 1.7492354740061162,
            "no_speech_prob": 0.00669097900390625,
            "words": [
                {
                    "start": 674.02,
                    "end": 674.58,
                    "word": " distributions",
                    "probability": 0.80712890625
                },
                {
                    "start": 674.58,
                    "end": 674.82,
                    "word": " that",
                    "probability": 0.87939453125
                },
                {
                    "start": 674.82,
                    "end": 674.94,
                    "word": " these",
                    "probability": 0.72412109375
                },
                {
                    "start": 674.94,
                    "end": 675.22,
                    "word": " came",
                    "probability": 0.81298828125
                },
                {
                    "start": 675.22,
                    "end": 675.54,
                    "word": " from",
                    "probability": 0.81982421875
                },
                {
                    "start": 675.54,
                    "end": 675.78,
                    "word": ".",
                    "probability": 0.73974609375
                },
                {
                    "start": 675.78,
                    "end": 676.0,
                    "word": " And",
                    "probability": 0.8994140625
                },
                {
                    "start": 676.0,
                    "end": 676.18,
                    "word": " that's",
                    "probability": 0.93994140625
                },
                {
                    "start": 676.18,
                    "end": 676.36,
                    "word": " quite",
                    "probability": 0.8818359375
                },
                {
                    "start": 676.5,
                    "end": 676.7,
                    "word": " a hard",
                    "probability": 0.8427734375
                },
                {
                    "start": 676.7,
                    "end": 677.22,
                    "word": " problem",
                    "probability": 0.75146484375
                },
                {
                    "start": 677.22,
                    "end": 677.62,
                    "word": ".",
                    "probability": 0.7177734375
                },
                {
                    "start": 677.68,
                    "end": 678.04,
                    "word": " And",
                    "probability": 0.88720703125
                },
                {
                    "start": 678.04,
                    "end": 678.9,
                    "word": " especially",
                    "probability": 0.6748046875
                },
                {
                    "start": 678.9,
                    "end": 679.12,
                    "word": " in",
                    "probability": 0.89697265625
                },
                {
                    "start": 679.12,
                    "end": 679.3,
                    "word": " high",
                    "probability": 0.86376953125
                }
            ]
        },
        {
            "id": 93,
            "seek": 56580,
            "start": 679.3,
            "end": 685.32,
            "text": " dimensions. So things get kind of complicated. So the solution to this that I came up with is to",
            "tokens": [
                50364,
                12819,
                13,
                407,
                721,
                483,
                733,
                295,
                6179,
                13,
                407,
                264,
                3827,
                281,
                341,
                300,
                286,
                1361,
                493,
                365,
                307,
                281,
                50660
            ],
            "temperature": 0.0,
            "avg_logprob": -0.23232323954803774,
            "compression_ratio": 1.7156862745098038,
            "no_speech_prob": 0.011505126953125,
            "words": [
                {
                    "start": 679.3,
                    "end": 679.72,
                    "word": " dimensions",
                    "probability": 0.70068359375
                },
                {
                    "start": 679.72,
                    "end": 680.16,
                    "word": ".",
                    "probability": 0.1739501953125
                },
                {
                    "start": 680.16,
                    "end": 680.58,
                    "word": " So",
                    "probability": 0.87841796875
                },
                {
                    "start": 680.58,
                    "end": 680.98,
                    "word": " things",
                    "probability": 0.64794921875
                },
                {
                    "start": 680.98,
                    "end": 681.12,
                    "word": " get",
                    "probability": 0.8857421875
                },
                {
                    "start": 681.12,
                    "end": 681.36,
                    "word": " kind",
                    "probability": 0.77734375
                },
                {
                    "start": 681.36,
                    "end": 681.44,
                    "word": " of",
                    "probability": 0.90625
                },
                {
                    "start": 681.44,
                    "end": 681.86,
                    "word": " complicated",
                    "probability": 0.8154296875
                },
                {
                    "start": 681.86,
                    "end": 682.28,
                    "word": ".",
                    "probability": 0.97412109375
                },
                {
                    "start": 682.28,
                    "end": 682.62,
                    "word": " So",
                    "probability": 0.2998046875
                },
                {
                    "start": 682.62,
                    "end": 682.78,
                    "word": " the",
                    "probability": 0.7900390625
                },
                {
                    "start": 682.78,
                    "end": 683.1,
                    "word": " solution",
                    "probability": 0.8427734375
                },
                {
                    "start": 683.1,
                    "end": 683.28,
                    "word": " to",
                    "probability": 0.90380859375
                },
                {
                    "start": 683.28,
                    "end": 683.5,
                    "word": " this",
                    "probability": 0.880859375
                },
                {
                    "start": 683.5,
                    "end": 683.72,
                    "word": " that",
                    "probability": 0.77099609375
                },
                {
                    "start": 683.72,
                    "end": 684.04,
                    "word": " I",
                    "probability": 0.99072265625
                },
                {
                    "start": 684.04,
                    "end": 684.24,
                    "word": " came",
                    "probability": 0.794921875
                },
                {
                    "start": 684.24,
                    "end": 684.4,
                    "word": " up",
                    "probability": 0.90673828125
                },
                {
                    "start": 684.4,
                    "end": 684.64,
                    "word": " with",
                    "probability": 0.81396484375
                },
                {
                    "start": 684.64,
                    "end": 685.16,
                    "word": " is",
                    "probability": 0.7939453125
                },
                {
                    "start": 685.16,
                    "end": 685.32,
                    "word": " to",
                    "probability": 0.90283203125
                }
            ]
        },
        {
            "id": 94,
            "seek": 56580,
            "start": 685.32,
            "end": 691.54,
            "text": " use a cover tree. Now, why do I want to use a cover tree? So the cover tree, basically,",
            "tokens": [
                50660,
                764,
                257,
                2060,
                4230,
                13,
                823,
                11,
                983,
                360,
                286,
                528,
                281,
                764,
                257,
                2060,
                4230,
                30,
                407,
                264,
                2060,
                4230,
                11,
                1936,
                11,
                50964
            ],
            "temperature": 0.0,
            "avg_logprob": -0.23232323954803774,
            "compression_ratio": 1.7156862745098038,
            "no_speech_prob": 0.011505126953125,
            "words": [
                {
                    "start": 685.32,
                    "end": 685.5,
                    "word": " use",
                    "probability": 0.822265625
                },
                {
                    "start": 685.64,
                    "end": 685.84,
                    "word": " a cover",
                    "probability": 0.65185546875
                },
                {
                    "start": 685.84,
                    "end": 686.08,
                    "word": " tree",
                    "probability": 0.759765625
                },
                {
                    "start": 686.08,
                    "end": 686.5,
                    "word": ".",
                    "probability": 0.98779296875
                },
                {
                    "start": 686.52,
                    "end": 687.0,
                    "word": " Now",
                    "probability": 0.8759765625
                },
                {
                    "start": 687.0,
                    "end": 687.12,
                    "word": ",",
                    "probability": 0.58935546875
                },
                {
                    "start": 687.12,
                    "end": 687.44,
                    "word": " why",
                    "probability": 0.81201171875
                },
                {
                    "start": 687.44,
                    "end": 687.64,
                    "word": " do",
                    "probability": 0.91796875
                },
                {
                    "start": 687.64,
                    "end": 687.76,
                    "word": " I",
                    "probability": 0.99462890625
                },
                {
                    "start": 687.76,
                    "end": 687.9,
                    "word": " want",
                    "probability": 0.82177734375
                },
                {
                    "start": 687.9,
                    "end": 687.98,
                    "word": " to",
                    "probability": 0.90966796875
                },
                {
                    "start": 687.98,
                    "end": 688.14,
                    "word": " use",
                    "probability": 0.82275390625
                },
                {
                    "start": 688.26,
                    "end": 688.44,
                    "word": " a cover",
                    "probability": 0.8056640625
                },
                {
                    "start": 688.44,
                    "end": 688.68,
                    "word": " tree",
                    "probability": 0.8271484375
                },
                {
                    "start": 688.68,
                    "end": 689.16,
                    "word": "?",
                    "probability": 0.8662109375
                },
                {
                    "start": 689.64,
                    "end": 689.94,
                    "word": " So",
                    "probability": 0.86181640625
                },
                {
                    "start": 689.94,
                    "end": 690.42,
                    "word": " the",
                    "probability": 0.6552734375
                },
                {
                    "start": 690.42,
                    "end": 690.64,
                    "word": " cover",
                    "probability": 0.89697265625
                },
                {
                    "start": 690.64,
                    "end": 690.84,
                    "word": " tree",
                    "probability": 0.83740234375
                },
                {
                    "start": 690.84,
                    "end": 690.98,
                    "word": ",",
                    "probability": 0.453857421875
                },
                {
                    "start": 690.98,
                    "end": 691.22,
                    "word": " basically",
                    "probability": 0.880859375
                },
                {
                    "start": 691.22,
                    "end": 691.54,
                    "word": ",",
                    "probability": 0.7255859375
                }
            ]
        },
        {
            "id": 95,
            "seek": 56580,
            "start": 692.12,
            "end": 696.42,
            "text": " the key takeaway for cover trees is it's a k-nearest neighbors data structure. There are",
            "tokens": [
                51016,
                264,
                2141,
                30681,
                337,
                2060,
                5852,
                307,
                309,
                311,
                257,
                350,
                12,
                716,
                17363,
                12512,
                1412,
                3877,
                13,
                821,
                366,
                51216
            ],
            "temperature": 0.0,
            "avg_logprob": -0.23232323954803774,
            "compression_ratio": 1.7156862745098038,
            "no_speech_prob": 0.011505126953125,
            "words": [
                {
                    "start": 692.12,
                    "end": 692.6,
                    "word": " the",
                    "probability": 0.8037109375
                },
                {
                    "start": 692.6,
                    "end": 692.96,
                    "word": " key",
                    "probability": 0.68408203125
                },
                {
                    "start": 692.96,
                    "end": 693.3,
                    "word": " takeaway",
                    "probability": 0.705078125
                },
                {
                    "start": 693.3,
                    "end": 693.52,
                    "word": " for",
                    "probability": 0.8916015625
                },
                {
                    "start": 693.52,
                    "end": 693.74,
                    "word": " cover",
                    "probability": 0.89208984375
                },
                {
                    "start": 693.74,
                    "end": 693.98,
                    "word": " trees",
                    "probability": 0.78173828125
                },
                {
                    "start": 693.98,
                    "end": 694.26,
                    "word": " is",
                    "probability": 0.87255859375
                },
                {
                    "start": 694.26,
                    "end": 694.78,
                    "word": " it's",
                    "probability": 0.820556640625
                },
                {
                    "start": 694.88,
                    "end": 695.02,
                    "word": " a k",
                    "probability": 0.490478515625
                },
                {
                    "start": 695.02,
                    "end": 695.2,
                    "word": "-nearest",
                    "probability": 0.8235677083333334
                },
                {
                    "start": 695.2,
                    "end": 695.42,
                    "word": " neighbors",
                    "probability": 0.381591796875
                },
                {
                    "start": 695.42,
                    "end": 695.66,
                    "word": " data",
                    "probability": 0.77099609375
                },
                {
                    "start": 695.66,
                    "end": 695.96,
                    "word": " structure",
                    "probability": 0.74755859375
                },
                {
                    "start": 695.96,
                    "end": 696.2,
                    "word": ".",
                    "probability": 0.85400390625
                },
                {
                    "start": 696.2,
                    "end": 696.34,
                    "word": " There",
                    "probability": 0.794921875
                },
                {
                    "start": 696.34,
                    "end": 696.42,
                    "word": " are",
                    "probability": 0.892578125
                }
            ]
        },
        {
            "id": 96,
            "seek": 56580,
            "start": 696.42,
            "end": 706.02,
            "text": " many like it. There's kd trees, there's b trees, there's kd trees, you know,",
            "tokens": [
                51216,
                867,
                411,
                309,
                13,
                821,
                311,
                350,
                67,
                5852,
                11,
                456,
                311,
                272,
                5852,
                11,
                456,
                311,
                350,
                67,
                5852,
                11,
                291,
                458,
                11,
                51616
            ],
            "temperature": 0.0,
            "avg_logprob": -0.23232323954803774,
            "compression_ratio": 1.7156862745098038,
            "no_speech_prob": 0.011505126953125,
            "words": [
                {
                    "start": 696.42,
                    "end": 696.62,
                    "word": " many",
                    "probability": 0.8232421875
                },
                {
                    "start": 696.62,
                    "end": 696.9,
                    "word": " like",
                    "probability": 0.86865234375
                },
                {
                    "start": 696.9,
                    "end": 697.06,
                    "word": " it",
                    "probability": 0.89990234375
                },
                {
                    "start": 697.06,
                    "end": 697.32,
                    "word": ".",
                    "probability": 0.8583984375
                },
                {
                    "start": 697.32,
                    "end": 697.64,
                    "word": " There's",
                    "probability": 0.88525390625
                },
                {
                    "start": 697.64,
                    "end": 697.92,
                    "word": " kd",
                    "probability": 0.59130859375
                },
                {
                    "start": 697.92,
                    "end": 698.32,
                    "word": " trees",
                    "probability": 0.55078125
                },
                {
                    "start": 698.32,
                    "end": 698.62,
                    "word": ",",
                    "probability": 0.78369140625
                },
                {
                    "start": 698.62,
                    "end": 699.26,
                    "word": " there's",
                    "probability": 0.852294921875
                },
                {
                    "start": 699.26,
                    "end": 699.46,
                    "word": " b",
                    "probability": 0.9033203125
                },
                {
                    "start": 699.46,
                    "end": 699.92,
                    "word": " trees",
                    "probability": 0.7490234375
                },
                {
                    "start": 699.92,
                    "end": 700.14,
                    "word": ",",
                    "probability": 0.90478515625
                },
                {
                    "start": 700.14,
                    "end": 700.74,
                    "word": " there's",
                    "probability": 0.90966796875
                },
                {
                    "start": 700.74,
                    "end": 702.16,
                    "word": " kd",
                    "probability": 0.910888671875
                },
                {
                    "start": 702.16,
                    "end": 702.5,
                    "word": " trees",
                    "probability": 0.83447265625
                },
                {
                    "start": 702.5,
                    "end": 703.84,
                    "word": ",",
                    "probability": 0.59326171875
                },
                {
                    "start": 703.84,
                    "end": 704.22,
                    "word": " you",
                    "probability": 0.49609375
                },
                {
                    "start": 704.22,
                    "end": 705.0,
                    "word": " know",
                    "probability": 0.8095703125
                },
                {
                    "start": 705.0,
                    "end": 706.02,
                    "word": ",",
                    "probability": 0.71923828125
                }
            ]
        },
        {
            "id": 97,
            "seek": 59108,
            "start": 706.02,
            "end": 714.2,
            "text": " there's k-means trees, a whole bunch. And, well, the cover tree has this wonderful thing where",
            "tokens": [
                50384,
                456,
                311,
                350,
                12,
                1398,
                599,
                5852,
                11,
                257,
                1379,
                3840,
                13,
                400,
                11,
                731,
                11,
                264,
                2060,
                4230,
                575,
                341,
                3715,
                551,
                689,
                50768
            ],
            "temperature": 0.0,
            "avg_logprob": -0.2995631605066279,
            "compression_ratio": 1.5756302521008403,
            "no_speech_prob": 0.002590179443359375,
            "words": [
                {
                    "start": 706.02,
                    "end": 706.96,
                    "word": " there's",
                    "probability": 0.212371826171875
                },
                {
                    "start": 706.96,
                    "end": 707.28,
                    "word": " k",
                    "probability": 0.21240234375
                },
                {
                    "start": 707.28,
                    "end": 707.52,
                    "word": "-means",
                    "probability": 0.93310546875
                },
                {
                    "start": 707.52,
                    "end": 707.8,
                    "word": " trees",
                    "probability": 0.78466796875
                },
                {
                    "start": 707.8,
                    "end": 708.3,
                    "word": ",",
                    "probability": 0.6337890625
                },
                {
                    "start": 708.54,
                    "end": 708.78,
                    "word": " a whole",
                    "probability": 0.822265625
                },
                {
                    "start": 708.78,
                    "end": 709.14,
                    "word": " bunch",
                    "probability": 0.8037109375
                },
                {
                    "start": 709.14,
                    "end": 709.7,
                    "word": ".",
                    "probability": 0.63037109375
                },
                {
                    "start": 710.14,
                    "end": 710.52,
                    "word": " And",
                    "probability": 0.5791015625
                },
                {
                    "start": 710.52,
                    "end": 710.6,
                    "word": ",",
                    "probability": 0.226318359375
                },
                {
                    "start": 710.6,
                    "end": 710.98,
                    "word": " well",
                    "probability": 0.76904296875
                },
                {
                    "start": 710.98,
                    "end": 711.64,
                    "word": ",",
                    "probability": 0.865234375
                },
                {
                    "start": 711.64,
                    "end": 711.8,
                    "word": " the",
                    "probability": 0.77197265625
                },
                {
                    "start": 711.8,
                    "end": 711.98,
                    "word": " cover",
                    "probability": 0.1431884765625
                },
                {
                    "start": 711.98,
                    "end": 712.24,
                    "word": " tree",
                    "probability": 0.6708984375
                },
                {
                    "start": 712.24,
                    "end": 713.16,
                    "word": " has",
                    "probability": 0.88720703125
                },
                {
                    "start": 713.16,
                    "end": 713.36,
                    "word": " this",
                    "probability": 0.88134765625
                },
                {
                    "start": 713.36,
                    "end": 713.64,
                    "word": " wonderful",
                    "probability": 0.7431640625
                },
                {
                    "start": 713.64,
                    "end": 713.92,
                    "word": " thing",
                    "probability": 0.81689453125
                },
                {
                    "start": 713.92,
                    "end": 714.2,
                    "word": " where",
                    "probability": 0.81591796875
                }
            ]
        },
        {
            "id": 98,
            "seek": 59108,
            "start": 714.94,
            "end": 722.06,
            "text": " in 2016, Mario Maggiorni and Wenjing Lau proved that it can arbitrarily well approximate the",
            "tokens": [
                50812,
                294,
                6549,
                11,
                9343,
                6395,
                7834,
                1865,
                72,
                293,
                23716,
                38863,
                47979,
                14617,
                300,
                309,
                393,
                19071,
                3289,
                731,
                30874,
                264,
                51160
            ],
            "temperature": 0.0,
            "avg_logprob": -0.2995631605066279,
            "compression_ratio": 1.5756302521008403,
            "no_speech_prob": 0.002590179443359375,
            "words": [
                {
                    "start": 714.94,
                    "end": 715.22,
                    "word": " in",
                    "probability": 0.6708984375
                },
                {
                    "start": 715.22,
                    "end": 716.0,
                    "word": " 2016",
                    "probability": 0.68701171875
                },
                {
                    "start": 716.0,
                    "end": 716.6,
                    "word": ",",
                    "probability": 0.568359375
                },
                {
                    "start": 716.6,
                    "end": 717.1,
                    "word": " Mario",
                    "probability": 0.84423828125
                },
                {
                    "start": 717.1,
                    "end": 717.64,
                    "word": " Maggiorni",
                    "probability": 0.6322021484375
                },
                {
                    "start": 717.64,
                    "end": 718.2,
                    "word": " and",
                    "probability": 0.87255859375
                },
                {
                    "start": 718.2,
                    "end": 718.5,
                    "word": " Wenjing",
                    "probability": 0.709228515625
                },
                {
                    "start": 718.5,
                    "end": 718.74,
                    "word": " Lau",
                    "probability": 0.36962890625
                },
                {
                    "start": 718.74,
                    "end": 719.66,
                    "word": " proved",
                    "probability": 0.375
                },
                {
                    "start": 719.66,
                    "end": 720.12,
                    "word": " that",
                    "probability": 0.87646484375
                },
                {
                    "start": 720.12,
                    "end": 720.34,
                    "word": " it",
                    "probability": 0.890625
                },
                {
                    "start": 720.34,
                    "end": 720.52,
                    "word": " can",
                    "probability": 0.8984375
                },
                {
                    "start": 720.52,
                    "end": 721.02,
                    "word": " arbitrarily",
                    "probability": 0.855224609375
                },
                {
                    "start": 721.02,
                    "end": 721.36,
                    "word": " well",
                    "probability": 0.82177734375
                },
                {
                    "start": 721.36,
                    "end": 721.8,
                    "word": " approximate",
                    "probability": 0.72265625
                },
                {
                    "start": 721.8,
                    "end": 722.06,
                    "word": " the",
                    "probability": 0.80859375
                }
            ]
        },
        {
            "id": 99,
            "seek": 59108,
            "start": 722.06,
            "end": 727.08,
            "text": " underlying distribution of the data. There's some caveats to that statement, and mathematically,",
            "tokens": [
                51160,
                14217,
                7316,
                295,
                264,
                1412,
                13,
                821,
                311,
                512,
                11730,
                1720,
                281,
                300,
                5629,
                11,
                293,
                44003,
                11,
                51408
            ],
            "temperature": 0.0,
            "avg_logprob": -0.2995631605066279,
            "compression_ratio": 1.5756302521008403,
            "no_speech_prob": 0.002590179443359375,
            "words": [
                {
                    "start": 722.06,
                    "end": 722.34,
                    "word": " underlying",
                    "probability": 0.8740234375
                },
                {
                    "start": 722.34,
                    "end": 722.76,
                    "word": " distribution",
                    "probability": 0.73828125
                },
                {
                    "start": 722.76,
                    "end": 722.98,
                    "word": " of",
                    "probability": 0.90771484375
                },
                {
                    "start": 722.98,
                    "end": 723.06,
                    "word": " the",
                    "probability": 0.79638671875
                },
                {
                    "start": 723.06,
                    "end": 723.26,
                    "word": " data",
                    "probability": 0.82666015625
                },
                {
                    "start": 723.26,
                    "end": 723.82,
                    "word": ".",
                    "probability": 0.98291015625
                },
                {
                    "start": 724.56,
                    "end": 725.12,
                    "word": " There's",
                    "probability": 0.86474609375
                },
                {
                    "start": 725.12,
                    "end": 725.3,
                    "word": " some",
                    "probability": 0.8271484375
                },
                {
                    "start": 725.3,
                    "end": 725.58,
                    "word": " caveats",
                    "probability": 0.870361328125
                },
                {
                    "start": 725.58,
                    "end": 725.72,
                    "word": " to",
                    "probability": 0.89892578125
                },
                {
                    "start": 725.72,
                    "end": 725.86,
                    "word": " that",
                    "probability": 0.87451171875
                },
                {
                    "start": 725.86,
                    "end": 726.18,
                    "word": " statement",
                    "probability": 0.841796875
                },
                {
                    "start": 726.18,
                    "end": 726.28,
                    "word": ",",
                    "probability": 0.343505859375
                },
                {
                    "start": 726.28,
                    "end": 726.34,
                    "word": " and",
                    "probability": 0.865234375
                },
                {
                    "start": 726.34,
                    "end": 726.82,
                    "word": " mathematically",
                    "probability": 0.755859375
                },
                {
                    "start": 726.82,
                    "end": 727.08,
                    "word": ",",
                    "probability": 0.439208984375
                }
            ]
        },
        {
            "id": 100,
            "seek": 59108,
            "start": 727.08,
            "end": 731.3,
            "text": " it will probably, to properly statement, will take it, take the whole page. But basically,",
            "tokens": [
                51408,
                309,
                486,
                1391,
                11,
                281,
                6108,
                5629,
                11,
                486,
                747,
                309,
                11,
                747,
                264,
                1379,
                3028,
                13,
                583,
                1936,
                11,
                51620
            ],
            "temperature": 0.0,
            "avg_logprob": -0.2995631605066279,
            "compression_ratio": 1.5756302521008403,
            "no_speech_prob": 0.002590179443359375,
            "words": [
                {
                    "start": 727.08,
                    "end": 727.08,
                    "word": " it",
                    "probability": 0.3046875
                },
                {
                    "start": 727.08,
                    "end": 727.24,
                    "word": " will",
                    "probability": 0.4697265625
                },
                {
                    "start": 727.24,
                    "end": 727.64,
                    "word": " probably",
                    "probability": 0.76611328125
                },
                {
                    "start": 727.64,
                    "end": 727.72,
                    "word": ",",
                    "probability": 0.428955078125
                },
                {
                    "start": 727.72,
                    "end": 727.88,
                    "word": " to",
                    "probability": 0.79345703125
                },
                {
                    "start": 727.88,
                    "end": 728.14,
                    "word": " properly",
                    "probability": 0.51318359375
                },
                {
                    "start": 728.14,
                    "end": 728.48,
                    "word": " statement",
                    "probability": 0.5986328125
                },
                {
                    "start": 728.48,
                    "end": 728.7,
                    "word": ",",
                    "probability": 0.70166015625
                },
                {
                    "start": 728.7,
                    "end": 728.84,
                    "word": " will",
                    "probability": 0.420654296875
                },
                {
                    "start": 728.84,
                    "end": 729.12,
                    "word": " take",
                    "probability": 0.80908203125
                },
                {
                    "start": 729.12,
                    "end": 729.26,
                    "word": " it",
                    "probability": 0.37255859375
                },
                {
                    "start": 729.26,
                    "end": 729.5,
                    "word": ",",
                    "probability": 0.83544921875
                },
                {
                    "start": 729.5,
                    "end": 729.9,
                    "word": " take",
                    "probability": 0.78125
                },
                {
                    "start": 729.9,
                    "end": 730.02,
                    "word": " the",
                    "probability": 0.783203125
                },
                {
                    "start": 730.02,
                    "end": 730.14,
                    "word": " whole",
                    "probability": 0.79736328125
                },
                {
                    "start": 730.14,
                    "end": 730.42,
                    "word": " page",
                    "probability": 0.79736328125
                },
                {
                    "start": 730.42,
                    "end": 730.52,
                    "word": ".",
                    "probability": 0.5087890625
                },
                {
                    "start": 730.52,
                    "end": 730.64,
                    "word": " But",
                    "probability": 0.896484375
                },
                {
                    "start": 730.64,
                    "end": 731.16,
                    "word": " basically",
                    "probability": 0.83251953125
                },
                {
                    "start": 731.16,
                    "end": 731.3,
                    "word": ",",
                    "probability": 0.5029296875
                }
            ]
        },
        {
            "id": 101,
            "seek": 61894,
            "start": 731.3,
            "end": 736.58,
            "text": " the whole concept of their proof is for a nice data set, and to know what a nice data set is,",
            "tokens": [
                50364,
                264,
                1379,
                3410,
                295,
                641,
                8177,
                307,
                337,
                257,
                1481,
                1412,
                992,
                11,
                293,
                281,
                458,
                437,
                257,
                1481,
                1412,
                992,
                307,
                11,
                50624
            ],
            "temperature": 0.0,
            "avg_logprob": -0.2124191769238176,
            "compression_ratio": 1.7557251908396947,
            "no_speech_prob": 0.233642578125,
            "words": [
                {
                    "start": 731.3,
                    "end": 731.38,
                    "word": " the",
                    "probability": 0.15673828125
                },
                {
                    "start": 731.38,
                    "end": 731.64,
                    "word": " whole",
                    "probability": 0.07781982421875
                },
                {
                    "start": 731.64,
                    "end": 732.06,
                    "word": " concept",
                    "probability": 0.80517578125
                },
                {
                    "start": 732.06,
                    "end": 732.22,
                    "word": " of",
                    "probability": 0.90673828125
                },
                {
                    "start": 732.22,
                    "end": 732.4,
                    "word": " their",
                    "probability": 0.8642578125
                },
                {
                    "start": 732.4,
                    "end": 732.84,
                    "word": " proof",
                    "probability": 0.814453125
                },
                {
                    "start": 732.84,
                    "end": 733.18,
                    "word": " is",
                    "probability": 0.8857421875
                },
                {
                    "start": 733.18,
                    "end": 733.82,
                    "word": " for",
                    "probability": 0.6220703125
                },
                {
                    "start": 733.94,
                    "end": 734.14,
                    "word": " a nice",
                    "probability": 0.82275390625
                },
                {
                    "start": 734.14,
                    "end": 734.34,
                    "word": " data",
                    "probability": 0.45947265625
                },
                {
                    "start": 734.34,
                    "end": 734.6,
                    "word": " set",
                    "probability": 0.8955078125
                },
                {
                    "start": 734.6,
                    "end": 734.9,
                    "word": ",",
                    "probability": 0.391357421875
                },
                {
                    "start": 734.9,
                    "end": 735.36,
                    "word": " and",
                    "probability": 0.74853515625
                },
                {
                    "start": 735.36,
                    "end": 735.5,
                    "word": " to",
                    "probability": 0.8984375
                },
                {
                    "start": 735.5,
                    "end": 735.62,
                    "word": " know",
                    "probability": 0.81494140625
                },
                {
                    "start": 735.62,
                    "end": 735.76,
                    "word": " what",
                    "probability": 0.87890625
                },
                {
                    "start": 735.86,
                    "end": 736.02,
                    "word": " a nice",
                    "probability": 0.8291015625
                },
                {
                    "start": 736.02,
                    "end": 736.22,
                    "word": " data",
                    "probability": 0.75732421875
                },
                {
                    "start": 736.22,
                    "end": 736.4,
                    "word": " set",
                    "probability": 0.88037109375
                },
                {
                    "start": 736.4,
                    "end": 736.56,
                    "word": " is",
                    "probability": 0.87744140625
                },
                {
                    "start": 736.56,
                    "end": 736.58,
                    "word": ",",
                    "probability": 0.64892578125
                }
            ]
        },
        {
            "id": 102,
            "seek": 61894,
            "start": 736.58,
            "end": 741.6,
            "text": " you need to know what a low dimensional manifold is. And you have to have enough data from that.",
            "tokens": [
                50624,
                291,
                643,
                281,
                458,
                437,
                257,
                2295,
                18795,
                47138,
                307,
                13,
                400,
                291,
                362,
                281,
                362,
                1547,
                1412,
                490,
                300,
                13,
                50852
            ],
            "temperature": 0.0,
            "avg_logprob": -0.2124191769238176,
            "compression_ratio": 1.7557251908396947,
            "no_speech_prob": 0.233642578125,
            "words": [
                {
                    "start": 736.58,
                    "end": 736.68,
                    "word": " you",
                    "probability": 0.88232421875
                },
                {
                    "start": 736.68,
                    "end": 736.88,
                    "word": " need",
                    "probability": 0.80029296875
                },
                {
                    "start": 736.88,
                    "end": 736.98,
                    "word": " to",
                    "probability": 0.90576171875
                },
                {
                    "start": 736.98,
                    "end": 737.12,
                    "word": " know",
                    "probability": 0.8125
                },
                {
                    "start": 737.12,
                    "end": 737.24,
                    "word": " what",
                    "probability": 0.88525390625
                },
                {
                    "start": 737.32,
                    "end": 737.42,
                    "word": " a low",
                    "probability": 0.8662109375
                },
                {
                    "start": 737.42,
                    "end": 737.72,
                    "word": " dimensional",
                    "probability": 0.57373046875
                },
                {
                    "start": 737.72,
                    "end": 738.04,
                    "word": " manifold",
                    "probability": 0.9453125
                },
                {
                    "start": 738.04,
                    "end": 738.46,
                    "word": " is",
                    "probability": 0.90380859375
                },
                {
                    "start": 738.46,
                    "end": 738.86,
                    "word": ".",
                    "probability": 0.51806640625
                },
                {
                    "start": 739.36,
                    "end": 739.76,
                    "word": " And",
                    "probability": 0.705078125
                },
                {
                    "start": 739.76,
                    "end": 739.9,
                    "word": " you",
                    "probability": 0.77734375
                },
                {
                    "start": 739.9,
                    "end": 740.02,
                    "word": " have",
                    "probability": 0.88134765625
                },
                {
                    "start": 740.02,
                    "end": 740.08,
                    "word": " to",
                    "probability": 0.89013671875
                },
                {
                    "start": 740.08,
                    "end": 740.26,
                    "word": " have",
                    "probability": 0.88330078125
                },
                {
                    "start": 740.26,
                    "end": 740.5,
                    "word": " enough",
                    "probability": 0.79296875
                },
                {
                    "start": 740.5,
                    "end": 740.72,
                    "word": " data",
                    "probability": 0.81494140625
                },
                {
                    "start": 740.72,
                    "end": 740.9,
                    "word": " from",
                    "probability": 0.80517578125
                },
                {
                    "start": 740.9,
                    "end": 741.58,
                    "word": " that",
                    "probability": 0.88232421875
                },
                {
                    "start": 741.58,
                    "end": 741.6,
                    "word": ".",
                    "probability": 0.720703125
                }
            ]
        },
        {
            "id": 103,
            "seek": 61894,
            "start": 742.0,
            "end": 748.94,
            "text": " But for a nice data set, which most of our data sets are pretty nice, sort of, kind of,",
            "tokens": [
                50904,
                583,
                337,
                257,
                1481,
                1412,
                992,
                11,
                597,
                881,
                295,
                527,
                1412,
                6352,
                366,
                1238,
                1481,
                11,
                1333,
                295,
                11,
                733,
                295,
                11,
                51192
            ],
            "temperature": 0.0,
            "avg_logprob": -0.2124191769238176,
            "compression_ratio": 1.7557251908396947,
            "no_speech_prob": 0.233642578125,
            "words": [
                {
                    "start": 742.0,
                    "end": 742.4,
                    "word": " But",
                    "probability": 0.8798828125
                },
                {
                    "start": 742.4,
                    "end": 742.52,
                    "word": " for",
                    "probability": 0.892578125
                },
                {
                    "start": 742.62,
                    "end": 742.76,
                    "word": " a nice",
                    "probability": 0.841796875
                },
                {
                    "start": 742.76,
                    "end": 742.9,
                    "word": " data",
                    "probability": 0.7734375
                },
                {
                    "start": 742.9,
                    "end": 743.12,
                    "word": " set",
                    "probability": 0.90283203125
                },
                {
                    "start": 743.12,
                    "end": 743.3,
                    "word": ",",
                    "probability": 0.93017578125
                },
                {
                    "start": 743.3,
                    "end": 743.54,
                    "word": " which",
                    "probability": 0.9072265625
                },
                {
                    "start": 743.54,
                    "end": 743.76,
                    "word": " most",
                    "probability": 0.787109375
                },
                {
                    "start": 743.76,
                    "end": 743.86,
                    "word": " of",
                    "probability": 0.9013671875
                },
                {
                    "start": 743.86,
                    "end": 743.98,
                    "word": " our",
                    "probability": 0.818359375
                },
                {
                    "start": 743.98,
                    "end": 744.14,
                    "word": " data",
                    "probability": 0.76025390625
                },
                {
                    "start": 744.14,
                    "end": 744.38,
                    "word": " sets",
                    "probability": 0.806640625
                },
                {
                    "start": 744.38,
                    "end": 744.86,
                    "word": " are",
                    "probability": 0.8935546875
                },
                {
                    "start": 744.86,
                    "end": 745.52,
                    "word": " pretty",
                    "probability": 0.82568359375
                },
                {
                    "start": 745.52,
                    "end": 745.92,
                    "word": " nice",
                    "probability": 0.81396484375
                },
                {
                    "start": 745.92,
                    "end": 746.2,
                    "word": ",",
                    "probability": 0.76611328125
                },
                {
                    "start": 746.2,
                    "end": 747.28,
                    "word": " sort",
                    "probability": 0.4365234375
                },
                {
                    "start": 747.28,
                    "end": 747.44,
                    "word": " of",
                    "probability": 0.90283203125
                },
                {
                    "start": 747.44,
                    "end": 747.6,
                    "word": ",",
                    "probability": 0.6416015625
                },
                {
                    "start": 747.6,
                    "end": 747.76,
                    "word": " kind",
                    "probability": 0.751953125
                },
                {
                    "start": 747.76,
                    "end": 748.0,
                    "word": " of",
                    "probability": 0.90625
                },
                {
                    "start": 748.0,
                    "end": 748.94,
                    "word": ",",
                    "probability": 0.60595703125
                }
            ]
        },
        {
            "id": 104,
            "seek": 61894,
            "start": 749.12,
            "end": 753.26,
            "text": " a cover tree will arbitrarily well approximate the underlying distribution of the data.",
            "tokens": [
                51236,
                257,
                2060,
                4230,
                486,
                19071,
                3289,
                731,
                30874,
                264,
                14217,
                7316,
                295,
                264,
                1412,
                13,
                51444
            ],
            "temperature": 0.0,
            "avg_logprob": -0.2124191769238176,
            "compression_ratio": 1.7557251908396947,
            "no_speech_prob": 0.233642578125,
            "words": [
                {
                    "start": 749.12,
                    "end": 749.26,
                    "word": " a cover",
                    "probability": 0.393798828125
                },
                {
                    "start": 749.26,
                    "end": 749.44,
                    "word": " tree",
                    "probability": 0.64599609375
                },
                {
                    "start": 749.44,
                    "end": 749.68,
                    "word": " will",
                    "probability": 0.82470703125
                },
                {
                    "start": 749.68,
                    "end": 750.34,
                    "word": " arbitrarily",
                    "probability": 0.819091796875
                },
                {
                    "start": 750.34,
                    "end": 750.62,
                    "word": " well",
                    "probability": 0.7333984375
                },
                {
                    "start": 750.62,
                    "end": 751.08,
                    "word": " approximate",
                    "probability": 0.7177734375
                },
                {
                    "start": 751.08,
                    "end": 751.42,
                    "word": " the",
                    "probability": 0.8173828125
                },
                {
                    "start": 751.42,
                    "end": 751.82,
                    "word": " underlying",
                    "probability": 0.91162109375
                },
                {
                    "start": 751.82,
                    "end": 752.28,
                    "word": " distribution",
                    "probability": 0.73681640625
                },
                {
                    "start": 752.28,
                    "end": 752.48,
                    "word": " of",
                    "probability": 0.90625
                },
                {
                    "start": 752.48,
                    "end": 752.6,
                    "word": " the",
                    "probability": 0.81591796875
                },
                {
                    "start": 752.6,
                    "end": 752.86,
                    "word": " data",
                    "probability": 0.810546875
                },
                {
                    "start": 752.86,
                    "end": 753.26,
                    "word": ".",
                    "probability": 0.98828125
                }
            ]
        },
        {
            "id": 105,
            "seek": 61894,
            "start": 754.46,
            "end": 759.16,
            "text": " There's some caveats, and you can go read their paper if you want on exactly what that means.",
            "tokens": [
                51524,
                821,
                311,
                512,
                11730,
                1720,
                11,
                293,
                291,
                393,
                352,
                1401,
                641,
                3035,
                498,
                291,
                528,
                322,
                2293,
                437,
                300,
                1355,
                13,
                51756
            ],
            "temperature": 0.0,
            "avg_logprob": -0.2124191769238176,
            "compression_ratio": 1.7557251908396947,
            "no_speech_prob": 0.233642578125,
            "words": [
                {
                    "start": 754.46,
                    "end": 754.86,
                    "word": " There's",
                    "probability": 0.890380859375
                },
                {
                    "start": 754.86,
                    "end": 755.0,
                    "word": " some",
                    "probability": 0.82421875
                },
                {
                    "start": 755.0,
                    "end": 755.36,
                    "word": " caveats",
                    "probability": 0.826904296875
                },
                {
                    "start": 755.36,
                    "end": 755.56,
                    "word": ",",
                    "probability": 0.57861328125
                },
                {
                    "start": 755.56,
                    "end": 755.74,
                    "word": " and",
                    "probability": 0.8916015625
                },
                {
                    "start": 755.74,
                    "end": 755.88,
                    "word": " you",
                    "probability": 0.8935546875
                },
                {
                    "start": 755.88,
                    "end": 756.0,
                    "word": " can",
                    "probability": 0.8955078125
                },
                {
                    "start": 756.0,
                    "end": 756.14,
                    "word": " go",
                    "probability": 0.90771484375
                },
                {
                    "start": 756.14,
                    "end": 756.38,
                    "word": " read",
                    "probability": 0.9052734375
                },
                {
                    "start": 756.38,
                    "end": 756.68,
                    "word": " their",
                    "probability": 0.9072265625
                },
                {
                    "start": 756.68,
                    "end": 756.98,
                    "word": " paper",
                    "probability": 0.8115234375
                },
                {
                    "start": 756.98,
                    "end": 757.14,
                    "word": " if",
                    "probability": 0.8369140625
                },
                {
                    "start": 757.14,
                    "end": 757.24,
                    "word": " you",
                    "probability": 0.89404296875
                },
                {
                    "start": 757.24,
                    "end": 757.44,
                    "word": " want",
                    "probability": 0.8203125
                },
                {
                    "start": 757.44,
                    "end": 757.82,
                    "word": " on",
                    "probability": 0.83740234375
                },
                {
                    "start": 757.82,
                    "end": 758.22,
                    "word": " exactly",
                    "probability": 0.7861328125
                },
                {
                    "start": 758.22,
                    "end": 758.42,
                    "word": " what",
                    "probability": 0.88916015625
                },
                {
                    "start": 758.42,
                    "end": 758.58,
                    "word": " that",
                    "probability": 0.8828125
                },
                {
                    "start": 758.58,
                    "end": 758.76,
                    "word": " means",
                    "probability": 0.8720703125
                },
                {
                    "start": 758.76,
                    "end": 759.16,
                    "word": ".",
                    "probability": 0.95458984375
                }
            ]
        },
        {
            "id": 106,
            "seek": 64736,
            "start": 759.16,
            "end": 764.14,
            "text": " But for us, what this means is, if I use a cover tree to build a model of my data for",
            "tokens": [
                50364,
                583,
                337,
                505,
                11,
                437,
                341,
                1355,
                307,
                11,
                498,
                286,
                764,
                257,
                2060,
                4230,
                281,
                1322,
                257,
                2316,
                295,
                452,
                1412,
                337,
                50612
            ],
            "temperature": 0.0,
            "avg_logprob": -0.22343750170298984,
            "compression_ratio": 1.9438202247191012,
            "no_speech_prob": 0.00806427001953125,
            "words": [
                {
                    "start": 759.16,
                    "end": 759.36,
                    "word": " But",
                    "probability": 0.469482421875
                },
                {
                    "start": 759.36,
                    "end": 759.6,
                    "word": " for",
                    "probability": 0.8515625
                },
                {
                    "start": 759.6,
                    "end": 759.86,
                    "word": " us",
                    "probability": 0.90380859375
                },
                {
                    "start": 759.86,
                    "end": 759.98,
                    "word": ",",
                    "probability": 0.6181640625
                },
                {
                    "start": 759.98,
                    "end": 760.08,
                    "word": " what",
                    "probability": 0.84814453125
                },
                {
                    "start": 760.08,
                    "end": 760.26,
                    "word": " this",
                    "probability": 0.87548828125
                },
                {
                    "start": 760.26,
                    "end": 760.42,
                    "word": " means",
                    "probability": 0.876953125
                },
                {
                    "start": 760.42,
                    "end": 760.6,
                    "word": " is",
                    "probability": 0.8935546875
                },
                {
                    "start": 760.6,
                    "end": 761.02,
                    "word": ",",
                    "probability": 0.333740234375
                },
                {
                    "start": 761.02,
                    "end": 761.1,
                    "word": " if",
                    "probability": 0.88720703125
                },
                {
                    "start": 761.1,
                    "end": 761.22,
                    "word": " I",
                    "probability": 0.99169921875
                },
                {
                    "start": 761.22,
                    "end": 761.4,
                    "word": " use",
                    "probability": 0.82958984375
                },
                {
                    "start": 761.56,
                    "end": 761.8,
                    "word": " a cover",
                    "probability": 0.537109375
                },
                {
                    "start": 761.8,
                    "end": 762.08,
                    "word": " tree",
                    "probability": 0.6298828125
                },
                {
                    "start": 762.08,
                    "end": 762.46,
                    "word": " to",
                    "probability": 0.884765625
                },
                {
                    "start": 762.46,
                    "end": 762.94,
                    "word": " build",
                    "probability": 0.873046875
                },
                {
                    "start": 763.16,
                    "end": 763.36,
                    "word": " a model",
                    "probability": 0.81494140625
                },
                {
                    "start": 763.36,
                    "end": 763.5,
                    "word": " of",
                    "probability": 0.8046875
                },
                {
                    "start": 763.5,
                    "end": 763.62,
                    "word": " my",
                    "probability": 0.8994140625
                },
                {
                    "start": 763.62,
                    "end": 763.86,
                    "word": " data",
                    "probability": 0.81201171875
                },
                {
                    "start": 763.86,
                    "end": 764.14,
                    "word": " for",
                    "probability": 0.84765625
                }
            ]
        },
        {
            "id": 107,
            "seek": 64736,
            "start": 764.14,
            "end": 768.68,
            "text": " k-nearest neighbors, I can take k-nearest neighbors, or I can just fiddle around with",
            "tokens": [
                50612,
                350,
                12,
                716,
                17363,
                12512,
                11,
                286,
                393,
                747,
                350,
                12,
                716,
                17363,
                12512,
                11,
                420,
                286,
                393,
                445,
                24553,
                2285,
                926,
                365,
                50848
            ],
            "temperature": 0.0,
            "avg_logprob": -0.22343750170298984,
            "compression_ratio": 1.9438202247191012,
            "no_speech_prob": 0.00806427001953125,
            "words": [
                {
                    "start": 764.14,
                    "end": 764.4,
                    "word": " k",
                    "probability": 0.65673828125
                },
                {
                    "start": 764.4,
                    "end": 764.58,
                    "word": "-nearest",
                    "probability": 0.7662760416666666
                },
                {
                    "start": 764.58,
                    "end": 764.86,
                    "word": " neighbors",
                    "probability": 0.583984375
                },
                {
                    "start": 764.86,
                    "end": 765.12,
                    "word": ",",
                    "probability": 0.91162109375
                },
                {
                    "start": 765.12,
                    "end": 765.38,
                    "word": " I",
                    "probability": 0.99462890625
                },
                {
                    "start": 765.38,
                    "end": 765.52,
                    "word": " can",
                    "probability": 0.89453125
                },
                {
                    "start": 765.52,
                    "end": 765.74,
                    "word": " take",
                    "probability": 0.80859375
                },
                {
                    "start": 765.74,
                    "end": 765.9,
                    "word": " k",
                    "probability": 0.96826171875
                },
                {
                    "start": 765.9,
                    "end": 766.12,
                    "word": "-nearest",
                    "probability": 0.8883463541666666
                },
                {
                    "start": 766.12,
                    "end": 766.46,
                    "word": " neighbors",
                    "probability": 0.8046875
                },
                {
                    "start": 766.46,
                    "end": 766.76,
                    "word": ",",
                    "probability": 0.5322265625
                },
                {
                    "start": 766.76,
                    "end": 767.14,
                    "word": " or",
                    "probability": 0.89990234375
                },
                {
                    "start": 767.14,
                    "end": 767.28,
                    "word": " I",
                    "probability": 0.9912109375
                },
                {
                    "start": 767.28,
                    "end": 767.4,
                    "word": " can",
                    "probability": 0.888671875
                },
                {
                    "start": 767.4,
                    "end": 767.6,
                    "word": " just",
                    "probability": 0.8798828125
                },
                {
                    "start": 767.6,
                    "end": 768.18,
                    "word": " fiddle",
                    "probability": 0.91748046875
                },
                {
                    "start": 768.18,
                    "end": 768.52,
                    "word": " around",
                    "probability": 0.8642578125
                },
                {
                    "start": 768.52,
                    "end": 768.68,
                    "word": " with",
                    "probability": 0.8115234375
                }
            ]
        },
        {
            "id": 108,
            "seek": 64736,
            "start": 768.68,
            "end": 772.5,
            "text": " the properties of the cover tree to, like, get information about my data set.",
            "tokens": [
                50848,
                264,
                7221,
                295,
                264,
                2060,
                4230,
                281,
                11,
                411,
                11,
                483,
                1589,
                466,
                452,
                1412,
                992,
                13,
                51016
            ],
            "temperature": 0.0,
            "avg_logprob": -0.22343750170298984,
            "compression_ratio": 1.9438202247191012,
            "no_speech_prob": 0.00806427001953125,
            "words": [
                {
                    "start": 768.68,
                    "end": 769.04,
                    "word": " the",
                    "probability": 0.8037109375
                },
                {
                    "start": 769.04,
                    "end": 769.36,
                    "word": " properties",
                    "probability": 0.77978515625
                },
                {
                    "start": 769.36,
                    "end": 769.5,
                    "word": " of",
                    "probability": 0.9091796875
                },
                {
                    "start": 769.5,
                    "end": 769.58,
                    "word": " the",
                    "probability": 0.80126953125
                },
                {
                    "start": 769.58,
                    "end": 769.72,
                    "word": " cover",
                    "probability": 0.791015625
                },
                {
                    "start": 769.72,
                    "end": 769.92,
                    "word": " tree",
                    "probability": 0.8310546875
                },
                {
                    "start": 769.92,
                    "end": 770.12,
                    "word": " to",
                    "probability": 0.8857421875
                },
                {
                    "start": 770.12,
                    "end": 770.14,
                    "word": ",",
                    "probability": 0.105712890625
                },
                {
                    "start": 770.14,
                    "end": 770.32,
                    "word": " like",
                    "probability": 0.859375
                },
                {
                    "start": 770.32,
                    "end": 770.46,
                    "word": ",",
                    "probability": 0.9775390625
                },
                {
                    "start": 770.46,
                    "end": 770.66,
                    "word": " get",
                    "probability": 0.8994140625
                },
                {
                    "start": 770.66,
                    "end": 771.22,
                    "word": " information",
                    "probability": 0.736328125
                },
                {
                    "start": 771.22,
                    "end": 771.6,
                    "word": " about",
                    "probability": 0.82373046875
                },
                {
                    "start": 771.6,
                    "end": 771.82,
                    "word": " my",
                    "probability": 0.9013671875
                },
                {
                    "start": 771.82,
                    "end": 771.98,
                    "word": " data",
                    "probability": 0.463134765625
                },
                {
                    "start": 771.98,
                    "end": 772.24,
                    "word": " set",
                    "probability": 0.82470703125
                },
                {
                    "start": 772.24,
                    "end": 772.5,
                    "word": ".",
                    "probability": 0.424072265625
                }
            ]
        },
        {
            "id": 109,
            "seek": 64736,
            "start": 773.42,
            "end": 777.44,
            "text": " So, I can tell things like the local dimensionality, I can tell things like how",
            "tokens": [
                51088,
                407,
                11,
                286,
                393,
                980,
                721,
                411,
                264,
                2654,
                10139,
                1860,
                11,
                286,
                393,
                980,
                721,
                411,
                577,
                51276
            ],
            "temperature": 0.0,
            "avg_logprob": -0.22343750170298984,
            "compression_ratio": 1.9438202247191012,
            "no_speech_prob": 0.00806427001953125,
            "words": [
                {
                    "start": 773.42,
                    "end": 773.82,
                    "word": " So",
                    "probability": 0.78466796875
                },
                {
                    "start": 773.82,
                    "end": 774.0,
                    "word": ",",
                    "probability": 0.3837890625
                },
                {
                    "start": 774.0,
                    "end": 774.08,
                    "word": " I",
                    "probability": 0.99853515625
                },
                {
                    "start": 774.08,
                    "end": 774.2,
                    "word": " can",
                    "probability": 0.89599609375
                },
                {
                    "start": 774.2,
                    "end": 774.36,
                    "word": " tell",
                    "probability": 0.8173828125
                },
                {
                    "start": 774.36,
                    "end": 774.64,
                    "word": " things",
                    "probability": 0.73095703125
                },
                {
                    "start": 774.64,
                    "end": 774.84,
                    "word": " like",
                    "probability": 0.828125
                },
                {
                    "start": 774.84,
                    "end": 775.02,
                    "word": " the",
                    "probability": 0.798828125
                },
                {
                    "start": 775.02,
                    "end": 775.22,
                    "word": " local",
                    "probability": 0.77587890625
                },
                {
                    "start": 775.22,
                    "end": 775.98,
                    "word": " dimensionality",
                    "probability": 0.929931640625
                },
                {
                    "start": 775.98,
                    "end": 776.26,
                    "word": ",",
                    "probability": 0.64599609375
                },
                {
                    "start": 776.26,
                    "end": 776.32,
                    "word": " I",
                    "probability": 0.97900390625
                },
                {
                    "start": 776.32,
                    "end": 776.42,
                    "word": " can",
                    "probability": 0.89599609375
                },
                {
                    "start": 776.42,
                    "end": 776.58,
                    "word": " tell",
                    "probability": 0.82373046875
                },
                {
                    "start": 776.58,
                    "end": 776.92,
                    "word": " things",
                    "probability": 0.67236328125
                },
                {
                    "start": 776.92,
                    "end": 777.24,
                    "word": " like",
                    "probability": 0.86083984375
                },
                {
                    "start": 777.24,
                    "end": 777.44,
                    "word": " how",
                    "probability": 0.8916015625
                }
            ]
        },
        {
            "id": 110,
            "seek": 64736,
            "start": 777.44,
            "end": 781.88,
            "text": " things glue together, and I can tell, like, sort of, the shape of my data fairly well. I can tell",
            "tokens": [
                51276,
                721,
                8998,
                1214,
                11,
                293,
                286,
                393,
                980,
                11,
                411,
                11,
                1333,
                295,
                11,
                264,
                3909,
                295,
                452,
                1412,
                6457,
                731,
                13,
                286,
                393,
                980,
                51500
            ],
            "temperature": 0.0,
            "avg_logprob": -0.22343750170298984,
            "compression_ratio": 1.9438202247191012,
            "no_speech_prob": 0.00806427001953125,
            "words": [
                {
                    "start": 777.44,
                    "end": 777.66,
                    "word": " things",
                    "probability": 0.72265625
                },
                {
                    "start": 777.66,
                    "end": 777.84,
                    "word": " glue",
                    "probability": 0.5634765625
                },
                {
                    "start": 777.84,
                    "end": 778.32,
                    "word": " together",
                    "probability": 0.7783203125
                },
                {
                    "start": 778.32,
                    "end": 778.82,
                    "word": ",",
                    "probability": 0.861328125
                },
                {
                    "start": 778.82,
                    "end": 779.18,
                    "word": " and",
                    "probability": 0.857421875
                },
                {
                    "start": 779.18,
                    "end": 779.28,
                    "word": " I",
                    "probability": 0.98974609375
                },
                {
                    "start": 779.28,
                    "end": 779.4,
                    "word": " can",
                    "probability": 0.89892578125
                },
                {
                    "start": 779.4,
                    "end": 779.62,
                    "word": " tell",
                    "probability": 0.83349609375
                },
                {
                    "start": 779.62,
                    "end": 779.72,
                    "word": ",",
                    "probability": 0.595703125
                },
                {
                    "start": 779.72,
                    "end": 779.88,
                    "word": " like",
                    "probability": 0.81982421875
                },
                {
                    "start": 779.88,
                    "end": 779.94,
                    "word": ",",
                    "probability": 0.98193359375
                },
                {
                    "start": 779.94,
                    "end": 780.1,
                    "word": " sort",
                    "probability": 0.814453125
                },
                {
                    "start": 780.1,
                    "end": 780.16,
                    "word": " of",
                    "probability": 0.908203125
                },
                {
                    "start": 780.16,
                    "end": 780.18,
                    "word": ",",
                    "probability": 0.1048583984375
                },
                {
                    "start": 780.18,
                    "end": 780.32,
                    "word": " the",
                    "probability": 0.8134765625
                },
                {
                    "start": 780.32,
                    "end": 780.6,
                    "word": " shape",
                    "probability": 0.80908203125
                },
                {
                    "start": 780.6,
                    "end": 780.72,
                    "word": " of",
                    "probability": 0.90283203125
                },
                {
                    "start": 780.72,
                    "end": 780.84,
                    "word": " my",
                    "probability": 0.8984375
                },
                {
                    "start": 780.84,
                    "end": 781.04,
                    "word": " data",
                    "probability": 0.81494140625
                },
                {
                    "start": 781.04,
                    "end": 781.28,
                    "word": " fairly",
                    "probability": 0.66845703125
                },
                {
                    "start": 781.28,
                    "end": 781.48,
                    "word": " well",
                    "probability": 0.880859375
                },
                {
                    "start": 781.48,
                    "end": 781.56,
                    "word": ".",
                    "probability": 0.818359375
                },
                {
                    "start": 781.56,
                    "end": 781.62,
                    "word": " I",
                    "probability": 0.99755859375
                },
                {
                    "start": 781.62,
                    "end": 781.72,
                    "word": " can",
                    "probability": 0.89892578125
                },
                {
                    "start": 781.72,
                    "end": 781.88,
                    "word": " tell",
                    "probability": 0.80908203125
                }
            ]
        },
        {
            "id": 111,
            "seek": 64736,
            "start": 781.88,
            "end": 787.58,
            "text": " clusters and things. I can infer what clusters are. Let's go build a very simple cover tree",
            "tokens": [
                51500,
                23313,
                293,
                721,
                13,
                286,
                393,
                13596,
                437,
                23313,
                366,
                13,
                961,
                311,
                352,
                1322,
                257,
                588,
                2199,
                2060,
                4230,
                51784
            ],
            "temperature": 0.0,
            "avg_logprob": -0.22343750170298984,
            "compression_ratio": 1.9438202247191012,
            "no_speech_prob": 0.00806427001953125,
            "words": [
                {
                    "start": 781.88,
                    "end": 782.32,
                    "word": " clusters",
                    "probability": 0.7353515625
                },
                {
                    "start": 782.32,
                    "end": 782.56,
                    "word": " and",
                    "probability": 0.810546875
                },
                {
                    "start": 782.56,
                    "end": 782.76,
                    "word": " things",
                    "probability": 0.69384765625
                },
                {
                    "start": 782.76,
                    "end": 782.98,
                    "word": ".",
                    "probability": 0.358154296875
                },
                {
                    "start": 782.98,
                    "end": 783.06,
                    "word": " I",
                    "probability": 0.9765625
                },
                {
                    "start": 783.06,
                    "end": 783.38,
                    "word": " can",
                    "probability": 0.90087890625
                },
                {
                    "start": 783.38,
                    "end": 784.34,
                    "word": " infer",
                    "probability": 0.91015625
                },
                {
                    "start": 784.34,
                    "end": 784.68,
                    "word": " what",
                    "probability": 0.8818359375
                },
                {
                    "start": 784.68,
                    "end": 785.0,
                    "word": " clusters",
                    "probability": 0.77197265625
                },
                {
                    "start": 785.0,
                    "end": 785.2,
                    "word": " are",
                    "probability": 0.8935546875
                },
                {
                    "start": 785.2,
                    "end": 785.6,
                    "word": ".",
                    "probability": 0.9931640625
                },
                {
                    "start": 785.68,
                    "end": 785.98,
                    "word": " Let's",
                    "probability": 0.931884765625
                },
                {
                    "start": 785.98,
                    "end": 786.16,
                    "word": " go",
                    "probability": 0.90869140625
                },
                {
                    "start": 786.16,
                    "end": 786.56,
                    "word": " build",
                    "probability": 0.8359375
                },
                {
                    "start": 786.74,
                    "end": 786.92,
                    "word": " a very",
                    "probability": 0.72119140625
                },
                {
                    "start": 786.92,
                    "end": 787.18,
                    "word": " simple",
                    "probability": 0.9130859375
                },
                {
                    "start": 787.18,
                    "end": 787.4,
                    "word": " cover",
                    "probability": 0.845703125
                },
                {
                    "start": 787.4,
                    "end": 787.58,
                    "word": " tree",
                    "probability": 0.83544921875
                }
            ]
        },
        {
            "id": 112,
            "seek": 67320,
            "start": 787.58,
            "end": 793.06,
            "text": " in two-dimensional data. So, here's a, you know, infinity sign, a bow tie, whatever you want to",
            "tokens": [
                50364,
                294,
                732,
                12,
                18759,
                1412,
                13,
                407,
                11,
                510,
                311,
                257,
                11,
                291,
                458,
                11,
                13202,
                1465,
                11,
                257,
                4503,
                7582,
                11,
                2035,
                291,
                528,
                281,
                50636
            ],
            "temperature": 0.0,
            "avg_logprob": -0.2141768362948565,
            "compression_ratio": 1.6340579710144927,
            "no_speech_prob": 0.0185394287109375,
            "words": [
                {
                    "start": 787.58,
                    "end": 787.72,
                    "word": " in",
                    "probability": 0.0743408203125
                },
                {
                    "start": 787.72,
                    "end": 787.82,
                    "word": " two",
                    "probability": 0.53955078125
                },
                {
                    "start": 787.82,
                    "end": 788.1,
                    "word": "-dimensional",
                    "probability": 0.6767578125
                },
                {
                    "start": 788.1,
                    "end": 788.34,
                    "word": " data",
                    "probability": 0.77197265625
                },
                {
                    "start": 788.34,
                    "end": 788.78,
                    "word": ".",
                    "probability": 0.88427734375
                },
                {
                    "start": 788.82,
                    "end": 788.96,
                    "word": " So",
                    "probability": 0.76953125
                },
                {
                    "start": 788.96,
                    "end": 789.06,
                    "word": ",",
                    "probability": 0.09600830078125
                },
                {
                    "start": 789.06,
                    "end": 789.34,
                    "word": " here's",
                    "probability": 0.871826171875
                },
                {
                    "start": 789.6,
                    "end": 789.6,
                    "word": " a,",
                    "probability": 0.291015625
                },
                {
                    "start": 789.6,
                    "end": 790.08,
                    "word": " you",
                    "probability": 0.82275390625
                },
                {
                    "start": 790.08,
                    "end": 790.14,
                    "word": " know",
                    "probability": 0.81640625
                },
                {
                    "start": 790.14,
                    "end": 790.18,
                    "word": ",",
                    "probability": 0.953125
                },
                {
                    "start": 790.18,
                    "end": 790.5,
                    "word": " infinity",
                    "probability": 0.7685546875
                },
                {
                    "start": 790.5,
                    "end": 790.98,
                    "word": " sign",
                    "probability": 0.78759765625
                },
                {
                    "start": 790.98,
                    "end": 791.38,
                    "word": ",",
                    "probability": 0.86669921875
                },
                {
                    "start": 791.98,
                    "end": 792.2,
                    "word": " a bow",
                    "probability": 0.943359375
                },
                {
                    "start": 792.2,
                    "end": 792.42,
                    "word": " tie",
                    "probability": 0.4921875
                },
                {
                    "start": 792.42,
                    "end": 792.5,
                    "word": ",",
                    "probability": 0.8720703125
                },
                {
                    "start": 792.5,
                    "end": 792.66,
                    "word": " whatever",
                    "probability": 0.80029296875
                },
                {
                    "start": 792.66,
                    "end": 792.88,
                    "word": " you",
                    "probability": 0.8984375
                },
                {
                    "start": 792.88,
                    "end": 792.98,
                    "word": " want",
                    "probability": 0.79736328125
                },
                {
                    "start": 792.98,
                    "end": 793.06,
                    "word": " to",
                    "probability": 0.90380859375
                }
            ]
        },
        {
            "id": 113,
            "seek": 67320,
            "start": 793.06,
            "end": 797.9,
            "text": " call it, and I start my cover tree by picking a point at random and then building a sphere.",
            "tokens": [
                50636,
                818,
                309,
                11,
                293,
                286,
                722,
                452,
                2060,
                4230,
                538,
                8867,
                257,
                935,
                412,
                4974,
                293,
                550,
                2390,
                257,
                16687,
                13,
                50856
            ],
            "temperature": 0.0,
            "avg_logprob": -0.2141768362948565,
            "compression_ratio": 1.6340579710144927,
            "no_speech_prob": 0.0185394287109375,
            "words": [
                {
                    "start": 793.06,
                    "end": 793.24,
                    "word": " call",
                    "probability": 0.82421875
                },
                {
                    "start": 793.24,
                    "end": 793.4,
                    "word": " it",
                    "probability": 0.9033203125
                },
                {
                    "start": 793.4,
                    "end": 793.54,
                    "word": ",",
                    "probability": 0.72021484375
                },
                {
                    "start": 793.54,
                    "end": 794.12,
                    "word": " and",
                    "probability": 0.896484375
                },
                {
                    "start": 794.12,
                    "end": 794.36,
                    "word": " I",
                    "probability": 0.9853515625
                },
                {
                    "start": 794.36,
                    "end": 794.66,
                    "word": " start",
                    "probability": 0.86279296875
                },
                {
                    "start": 794.66,
                    "end": 794.8,
                    "word": " my",
                    "probability": 0.9111328125
                },
                {
                    "start": 794.8,
                    "end": 794.96,
                    "word": " cover",
                    "probability": 0.58984375
                },
                {
                    "start": 794.96,
                    "end": 795.1,
                    "word": " tree",
                    "probability": 0.779296875
                },
                {
                    "start": 795.1,
                    "end": 795.3,
                    "word": " by",
                    "probability": 0.90771484375
                },
                {
                    "start": 795.3,
                    "end": 795.58,
                    "word": " picking",
                    "probability": 0.73681640625
                },
                {
                    "start": 795.78,
                    "end": 795.94,
                    "word": " a point",
                    "probability": 0.8505859375
                },
                {
                    "start": 795.94,
                    "end": 796.1,
                    "word": " at",
                    "probability": 0.90380859375
                },
                {
                    "start": 796.1,
                    "end": 796.32,
                    "word": " random",
                    "probability": 0.7353515625
                },
                {
                    "start": 796.32,
                    "end": 796.54,
                    "word": " and",
                    "probability": 0.72265625
                },
                {
                    "start": 796.54,
                    "end": 796.72,
                    "word": " then",
                    "probability": 0.716796875
                },
                {
                    "start": 796.72,
                    "end": 796.98,
                    "word": " building",
                    "probability": 0.8125
                },
                {
                    "start": 797.22,
                    "end": 797.46,
                    "word": " a sphere",
                    "probability": 0.75537109375
                },
                {
                    "start": 797.46,
                    "end": 797.9,
                    "word": ".",
                    "probability": 0.197998046875
                }
            ]
        },
        {
            "id": 114,
            "seek": 67320,
            "start": 798.08,
            "end": 803.36,
            "text": " You know, in this case, it's a circle that covers everything. So, this is the start. And now,",
            "tokens": [
                50888,
                509,
                458,
                11,
                294,
                341,
                1389,
                11,
                309,
                311,
                257,
                6329,
                300,
                10538,
                1203,
                13,
                407,
                11,
                341,
                307,
                264,
                722,
                13,
                400,
                586,
                11,
                51152
            ],
            "temperature": 0.0,
            "avg_logprob": -0.2141768362948565,
            "compression_ratio": 1.6340579710144927,
            "no_speech_prob": 0.0185394287109375,
            "words": [
                {
                    "start": 798.08,
                    "end": 798.26,
                    "word": " You",
                    "probability": 0.66259765625
                },
                {
                    "start": 798.26,
                    "end": 798.4,
                    "word": " know",
                    "probability": 0.81591796875
                },
                {
                    "start": 798.4,
                    "end": 798.5,
                    "word": ",",
                    "probability": 0.9365234375
                },
                {
                    "start": 798.5,
                    "end": 798.98,
                    "word": " in",
                    "probability": 0.89453125
                },
                {
                    "start": 798.98,
                    "end": 799.14,
                    "word": " this",
                    "probability": 0.8798828125
                },
                {
                    "start": 799.14,
                    "end": 799.36,
                    "word": " case",
                    "probability": 0.806640625
                },
                {
                    "start": 799.36,
                    "end": 799.46,
                    "word": ",",
                    "probability": 0.471923828125
                },
                {
                    "start": 799.46,
                    "end": 799.54,
                    "word": " it's",
                    "probability": 0.89501953125
                },
                {
                    "start": 799.64,
                    "end": 799.92,
                    "word": " a circle",
                    "probability": 0.90625
                },
                {
                    "start": 799.92,
                    "end": 800.44,
                    "word": " that",
                    "probability": 0.77294921875
                },
                {
                    "start": 800.44,
                    "end": 800.7,
                    "word": " covers",
                    "probability": 0.7890625
                },
                {
                    "start": 800.7,
                    "end": 801.06,
                    "word": " everything",
                    "probability": 0.8955078125
                },
                {
                    "start": 801.06,
                    "end": 801.34,
                    "word": ".",
                    "probability": 0.96044921875
                },
                {
                    "start": 801.34,
                    "end": 801.74,
                    "word": " So",
                    "probability": 0.89990234375
                },
                {
                    "start": 801.74,
                    "end": 801.78,
                    "word": ",",
                    "probability": 0.72998046875
                },
                {
                    "start": 801.78,
                    "end": 801.92,
                    "word": " this",
                    "probability": 0.8818359375
                },
                {
                    "start": 801.92,
                    "end": 802.04,
                    "word": " is",
                    "probability": 0.904296875
                },
                {
                    "start": 802.04,
                    "end": 802.18,
                    "word": " the",
                    "probability": 0.81640625
                },
                {
                    "start": 802.18,
                    "end": 802.44,
                    "word": " start",
                    "probability": 0.83837890625
                },
                {
                    "start": 802.44,
                    "end": 802.74,
                    "word": ".",
                    "probability": 0.4658203125
                },
                {
                    "start": 802.74,
                    "end": 803.08,
                    "word": " And",
                    "probability": 0.8349609375
                },
                {
                    "start": 803.08,
                    "end": 803.3,
                    "word": " now",
                    "probability": 0.84130859375
                },
                {
                    "start": 803.3,
                    "end": 803.36,
                    "word": ",",
                    "probability": 0.6357421875
                }
            ]
        },
        {
            "id": 115,
            "seek": 67320,
            "start": 803.36,
            "end": 807.44,
            "text": " this doesn't help me that much. I haven't split up my data geometrically to, like,",
            "tokens": [
                51152,
                341,
                1177,
                380,
                854,
                385,
                300,
                709,
                13,
                286,
                2378,
                380,
                7472,
                493,
                452,
                1412,
                12956,
                81,
                984,
                281,
                11,
                411,
                11,
                51352
            ],
            "temperature": 0.0,
            "avg_logprob": -0.2141768362948565,
            "compression_ratio": 1.6340579710144927,
            "no_speech_prob": 0.0185394287109375,
            "words": [
                {
                    "start": 803.36,
                    "end": 803.62,
                    "word": " this",
                    "probability": 0.8828125
                },
                {
                    "start": 803.62,
                    "end": 803.88,
                    "word": " doesn't",
                    "probability": 0.932861328125
                },
                {
                    "start": 803.88,
                    "end": 804.04,
                    "word": " help",
                    "probability": 0.81103515625
                },
                {
                    "start": 804.04,
                    "end": 804.14,
                    "word": " me",
                    "probability": 0.904296875
                },
                {
                    "start": 804.14,
                    "end": 804.28,
                    "word": " that",
                    "probability": 0.8857421875
                },
                {
                    "start": 804.28,
                    "end": 804.44,
                    "word": " much",
                    "probability": 0.80712890625
                },
                {
                    "start": 804.44,
                    "end": 804.54,
                    "word": ".",
                    "probability": 0.7841796875
                },
                {
                    "start": 804.54,
                    "end": 804.66,
                    "word": " I",
                    "probability": 0.99853515625
                },
                {
                    "start": 804.66,
                    "end": 805.0,
                    "word": " haven't",
                    "probability": 0.93701171875
                },
                {
                    "start": 805.0,
                    "end": 805.28,
                    "word": " split",
                    "probability": 0.79052734375
                },
                {
                    "start": 805.28,
                    "end": 805.42,
                    "word": " up",
                    "probability": 0.89404296875
                },
                {
                    "start": 805.42,
                    "end": 805.58,
                    "word": " my",
                    "probability": 0.8984375
                },
                {
                    "start": 805.58,
                    "end": 805.8,
                    "word": " data",
                    "probability": 0.81640625
                },
                {
                    "start": 805.8,
                    "end": 806.48,
                    "word": " geometrically",
                    "probability": 0.9137369791666666
                },
                {
                    "start": 806.48,
                    "end": 806.78,
                    "word": " to",
                    "probability": 0.8798828125
                },
                {
                    "start": 806.78,
                    "end": 806.92,
                    "word": ",",
                    "probability": 0.654296875
                },
                {
                    "start": 806.92,
                    "end": 807.1,
                    "word": " like",
                    "probability": 0.8662109375
                },
                {
                    "start": 807.1,
                    "end": 807.44,
                    "word": ",",
                    "probability": 0.98583984375
                }
            ]
        },
        {
            "id": 116,
            "seek": 67320,
            "start": 807.44,
            "end": 813.42,
            "text": " divide and conquer the k-nearest neighbors structure, because the naive way would take",
            "tokens": [
                51352,
                9845,
                293,
                24136,
                264,
                350,
                12,
                716,
                17363,
                12512,
                3877,
                11,
                570,
                264,
                29052,
                636,
                576,
                747,
                51660
            ],
            "temperature": 0.0,
            "avg_logprob": -0.2141768362948565,
            "compression_ratio": 1.6340579710144927,
            "no_speech_prob": 0.0185394287109375,
            "words": [
                {
                    "start": 807.44,
                    "end": 807.74,
                    "word": " divide",
                    "probability": 0.87353515625
                },
                {
                    "start": 807.74,
                    "end": 807.94,
                    "word": " and",
                    "probability": 0.89306640625
                },
                {
                    "start": 807.94,
                    "end": 808.3,
                    "word": " conquer",
                    "probability": 0.84033203125
                },
                {
                    "start": 808.3,
                    "end": 808.62,
                    "word": " the",
                    "probability": 0.81787109375
                },
                {
                    "start": 808.62,
                    "end": 808.88,
                    "word": " k",
                    "probability": 0.50341796875
                },
                {
                    "start": 808.88,
                    "end": 809.08,
                    "word": "-nearest",
                    "probability": 0.83837890625
                },
                {
                    "start": 809.08,
                    "end": 809.4,
                    "word": " neighbors",
                    "probability": 0.377197265625
                },
                {
                    "start": 809.4,
                    "end": 811.28,
                    "word": " structure",
                    "probability": 0.62109375
                },
                {
                    "start": 811.28,
                    "end": 811.44,
                    "word": ",",
                    "probability": 0.360107421875
                },
                {
                    "start": 811.44,
                    "end": 811.64,
                    "word": " because",
                    "probability": 0.828125
                },
                {
                    "start": 811.64,
                    "end": 812.14,
                    "word": " the",
                    "probability": 0.80126953125
                },
                {
                    "start": 812.14,
                    "end": 812.46,
                    "word": " naive",
                    "probability": 0.59326171875
                },
                {
                    "start": 812.46,
                    "end": 812.76,
                    "word": " way",
                    "probability": 0.88134765625
                },
                {
                    "start": 812.76,
                    "end": 813.12,
                    "word": " would",
                    "probability": 0.84326171875
                },
                {
                    "start": 813.12,
                    "end": 813.42,
                    "word": " take",
                    "probability": 0.82421875
                }
            ]
        },
        {
            "id": 117,
            "seek": 69974,
            "start": 813.96,
            "end": 818.88,
            "text": " a linear time, big O of n, but we want to divide and conquer, so it takes log of n.",
            "tokens": [
                50368,
                257,
                8213,
                565,
                11,
                955,
                422,
                295,
                297,
                11,
                457,
                321,
                528,
                281,
                9845,
                293,
                24136,
                11,
                370,
                309,
                2516,
                3565,
                295,
                297,
                13,
                50628
            ],
            "temperature": 0.0,
            "avg_logprob": -0.19616699474863708,
            "compression_ratio": 1.6221374045801527,
            "no_speech_prob": 0.0154266357421875,
            "words": [
                {
                    "start": 813.96,
                    "end": 814.32,
                    "word": " a linear",
                    "probability": 0.78759765625
                },
                {
                    "start": 814.32,
                    "end": 814.68,
                    "word": " time",
                    "probability": 0.75634765625
                },
                {
                    "start": 814.68,
                    "end": 814.8,
                    "word": ",",
                    "probability": 0.61279296875
                },
                {
                    "start": 814.8,
                    "end": 815.1,
                    "word": " big",
                    "probability": 0.52294921875
                },
                {
                    "start": 815.1,
                    "end": 815.3,
                    "word": " O",
                    "probability": 0.86767578125
                },
                {
                    "start": 815.3,
                    "end": 815.44,
                    "word": " of",
                    "probability": 0.251708984375
                },
                {
                    "start": 815.44,
                    "end": 815.6,
                    "word": " n",
                    "probability": 0.6181640625
                },
                {
                    "start": 815.6,
                    "end": 816.18,
                    "word": ",",
                    "probability": 0.346923828125
                },
                {
                    "start": 816.18,
                    "end": 816.46,
                    "word": " but",
                    "probability": 0.85595703125
                },
                {
                    "start": 816.46,
                    "end": 816.58,
                    "word": " we",
                    "probability": 0.8955078125
                },
                {
                    "start": 816.58,
                    "end": 816.74,
                    "word": " want",
                    "probability": 0.75830078125
                },
                {
                    "start": 816.74,
                    "end": 816.84,
                    "word": " to",
                    "probability": 0.83544921875
                },
                {
                    "start": 816.84,
                    "end": 817.12,
                    "word": " divide",
                    "probability": 0.86328125
                },
                {
                    "start": 817.12,
                    "end": 817.28,
                    "word": " and",
                    "probability": 0.87890625
                },
                {
                    "start": 817.28,
                    "end": 817.52,
                    "word": " conquer",
                    "probability": 0.853515625
                },
                {
                    "start": 817.52,
                    "end": 817.64,
                    "word": ",",
                    "probability": 0.42724609375
                },
                {
                    "start": 817.64,
                    "end": 817.7,
                    "word": " so",
                    "probability": 0.8837890625
                },
                {
                    "start": 817.7,
                    "end": 817.8,
                    "word": " it",
                    "probability": 0.833984375
                },
                {
                    "start": 817.8,
                    "end": 818.06,
                    "word": " takes",
                    "probability": 0.708984375
                },
                {
                    "start": 818.06,
                    "end": 818.3,
                    "word": " log",
                    "probability": 0.82080078125
                },
                {
                    "start": 818.3,
                    "end": 818.42,
                    "word": " of",
                    "probability": 0.7177734375
                },
                {
                    "start": 818.42,
                    "end": 818.56,
                    "word": " n",
                    "probability": 0.94873046875
                },
                {
                    "start": 818.56,
                    "end": 818.88,
                    "word": ".",
                    "probability": 0.8916015625
                }
            ]
        },
        {
            "id": 118,
            "seek": 69974,
            "start": 818.88,
            "end": 823.64,
            "text": " So, here's what I do. I shrink my sphere down, add another one, and I cover it again.",
            "tokens": [
                50628,
                407,
                11,
                510,
                311,
                437,
                286,
                360,
                13,
                286,
                23060,
                452,
                16687,
                760,
                11,
                909,
                1071,
                472,
                11,
                293,
                286,
                2060,
                309,
                797,
                13,
                50876
            ],
            "temperature": 0.0,
            "avg_logprob": -0.19616699474863708,
            "compression_ratio": 1.6221374045801527,
            "no_speech_prob": 0.0154266357421875,
            "words": [
                {
                    "start": 818.88,
                    "end": 819.0,
                    "word": " So",
                    "probability": 0.75830078125
                },
                {
                    "start": 819.0,
                    "end": 819.12,
                    "word": ",",
                    "probability": 0.1419677734375
                },
                {
                    "start": 819.12,
                    "end": 819.46,
                    "word": " here's",
                    "probability": 0.886474609375
                },
                {
                    "start": 819.46,
                    "end": 819.64,
                    "word": " what",
                    "probability": 0.884765625
                },
                {
                    "start": 819.64,
                    "end": 819.74,
                    "word": " I",
                    "probability": 0.990234375
                },
                {
                    "start": 819.74,
                    "end": 819.92,
                    "word": " do",
                    "probability": 0.90185546875
                },
                {
                    "start": 819.92,
                    "end": 820.12,
                    "word": ".",
                    "probability": 0.7138671875
                },
                {
                    "start": 820.12,
                    "end": 820.2,
                    "word": " I",
                    "probability": 0.982421875
                },
                {
                    "start": 820.2,
                    "end": 820.86,
                    "word": " shrink",
                    "probability": 0.89892578125
                },
                {
                    "start": 820.86,
                    "end": 821.04,
                    "word": " my",
                    "probability": 0.89599609375
                },
                {
                    "start": 821.04,
                    "end": 821.2,
                    "word": " sphere",
                    "probability": 0.84765625
                },
                {
                    "start": 821.2,
                    "end": 821.42,
                    "word": " down",
                    "probability": 0.80908203125
                },
                {
                    "start": 821.42,
                    "end": 821.58,
                    "word": ",",
                    "probability": 0.92138671875
                },
                {
                    "start": 821.58,
                    "end": 821.76,
                    "word": " add",
                    "probability": 0.78466796875
                },
                {
                    "start": 821.76,
                    "end": 821.96,
                    "word": " another",
                    "probability": 0.80615234375
                },
                {
                    "start": 821.96,
                    "end": 822.18,
                    "word": " one",
                    "probability": 0.82861328125
                },
                {
                    "start": 822.18,
                    "end": 822.26,
                    "word": ",",
                    "probability": 0.70263671875
                },
                {
                    "start": 822.26,
                    "end": 822.26,
                    "word": " and",
                    "probability": 0.689453125
                },
                {
                    "start": 822.26,
                    "end": 822.38,
                    "word": " I",
                    "probability": 0.76416015625
                },
                {
                    "start": 822.38,
                    "end": 822.9,
                    "word": " cover",
                    "probability": 0.86767578125
                },
                {
                    "start": 822.9,
                    "end": 823.06,
                    "word": " it",
                    "probability": 0.90283203125
                },
                {
                    "start": 823.06,
                    "end": 823.28,
                    "word": " again",
                    "probability": 0.9208984375
                },
                {
                    "start": 823.28,
                    "end": 823.64,
                    "word": ".",
                    "probability": 0.96240234375
                }
            ]
        },
        {
            "id": 119,
            "seek": 69974,
            "start": 823.76,
            "end": 829.38,
            "text": " You can kind of tell that this object is longer than it is wide. It's kind of,",
            "tokens": [
                50876,
                509,
                393,
                733,
                295,
                980,
                300,
                341,
                2657,
                307,
                2854,
                813,
                309,
                307,
                4874,
                13,
                467,
                311,
                733,
                295,
                11,
                51160
            ],
            "temperature": 0.0,
            "avg_logprob": -0.19616699474863708,
            "compression_ratio": 1.6221374045801527,
            "no_speech_prob": 0.0154266357421875,
            "words": [
                {
                    "start": 823.76,
                    "end": 823.92,
                    "word": " You",
                    "probability": 0.7060546875
                },
                {
                    "start": 823.92,
                    "end": 824.04,
                    "word": " can",
                    "probability": 0.89453125
                },
                {
                    "start": 824.04,
                    "end": 824.22,
                    "word": " kind",
                    "probability": 0.70166015625
                },
                {
                    "start": 824.22,
                    "end": 824.3,
                    "word": " of",
                    "probability": 0.90283203125
                },
                {
                    "start": 824.3,
                    "end": 824.5,
                    "word": " tell",
                    "probability": 0.76513671875
                },
                {
                    "start": 824.5,
                    "end": 824.74,
                    "word": " that",
                    "probability": 0.8173828125
                },
                {
                    "start": 824.74,
                    "end": 825.3,
                    "word": " this",
                    "probability": 0.6083984375
                },
                {
                    "start": 825.3,
                    "end": 826.26,
                    "word": " object",
                    "probability": 0.814453125
                },
                {
                    "start": 826.26,
                    "end": 826.5,
                    "word": " is",
                    "probability": 0.89892578125
                },
                {
                    "start": 826.5,
                    "end": 826.84,
                    "word": " longer",
                    "probability": 0.8369140625
                },
                {
                    "start": 826.84,
                    "end": 827.54,
                    "word": " than",
                    "probability": 0.86474609375
                },
                {
                    "start": 827.54,
                    "end": 827.66,
                    "word": " it",
                    "probability": 0.90478515625
                },
                {
                    "start": 827.66,
                    "end": 827.8,
                    "word": " is",
                    "probability": 0.908203125
                },
                {
                    "start": 827.8,
                    "end": 828.12,
                    "word": " wide",
                    "probability": 0.8447265625
                },
                {
                    "start": 828.12,
                    "end": 828.48,
                    "word": ".",
                    "probability": 0.56494140625
                },
                {
                    "start": 828.5,
                    "end": 828.64,
                    "word": " It's",
                    "probability": 0.7003173828125
                },
                {
                    "start": 828.64,
                    "end": 828.88,
                    "word": " kind",
                    "probability": 0.796875
                },
                {
                    "start": 828.88,
                    "end": 829.1,
                    "word": " of",
                    "probability": 0.9033203125
                },
                {
                    "start": 829.1,
                    "end": 829.38,
                    "word": ",",
                    "probability": 0.50341796875
                }
            ]
        },
        {
            "id": 120,
            "seek": 69974,
            "start": 829.38,
            "end": 833.46,
            "text": " if you really squint, it's kind of a one-dimensional thing in this direction.",
            "tokens": [
                51160,
                498,
                291,
                534,
                2339,
                686,
                11,
                309,
                311,
                733,
                295,
                257,
                472,
                12,
                18759,
                551,
                294,
                341,
                3513,
                13,
                51348
            ],
            "temperature": 0.0,
            "avg_logprob": -0.19616699474863708,
            "compression_ratio": 1.6221374045801527,
            "no_speech_prob": 0.0154266357421875,
            "words": [
                {
                    "start": 829.38,
                    "end": 829.6,
                    "word": " if",
                    "probability": 0.896484375
                },
                {
                    "start": 829.6,
                    "end": 829.72,
                    "word": " you",
                    "probability": 0.89453125
                },
                {
                    "start": 829.72,
                    "end": 830.02,
                    "word": " really",
                    "probability": 0.791015625
                },
                {
                    "start": 830.02,
                    "end": 830.56,
                    "word": " squint",
                    "probability": 0.9814453125
                },
                {
                    "start": 830.56,
                    "end": 830.72,
                    "word": ",",
                    "probability": 0.94775390625
                },
                {
                    "start": 830.72,
                    "end": 830.9,
                    "word": " it's",
                    "probability": 0.939208984375
                },
                {
                    "start": 830.9,
                    "end": 831.1,
                    "word": " kind",
                    "probability": 0.75390625
                },
                {
                    "start": 831.1,
                    "end": 831.24,
                    "word": " of",
                    "probability": 0.90625
                },
                {
                    "start": 831.48,
                    "end": 831.72,
                    "word": " a one",
                    "probability": 0.79638671875
                },
                {
                    "start": 831.72,
                    "end": 832.02,
                    "word": "-dimensional",
                    "probability": 0.846435546875
                },
                {
                    "start": 832.02,
                    "end": 832.4,
                    "word": " thing",
                    "probability": 0.80859375
                },
                {
                    "start": 832.4,
                    "end": 832.54,
                    "word": " in",
                    "probability": 0.884765625
                },
                {
                    "start": 832.54,
                    "end": 832.72,
                    "word": " this",
                    "probability": 0.87939453125
                },
                {
                    "start": 832.72,
                    "end": 833.1,
                    "word": " direction",
                    "probability": 0.91015625
                },
                {
                    "start": 833.1,
                    "end": 833.46,
                    "word": ".",
                    "probability": 0.98193359375
                }
            ]
        },
        {
            "id": 121,
            "seek": 69974,
            "start": 835.1,
            "end": 839.96,
            "text": " So, that's great. But now, if I go, okay, well, I'm going to split it up further, so really divide",
            "tokens": [
                51420,
                407,
                11,
                300,
                311,
                869,
                13,
                583,
                586,
                11,
                498,
                286,
                352,
                11,
                1392,
                11,
                731,
                11,
                286,
                478,
                516,
                281,
                7472,
                309,
                493,
                3052,
                11,
                370,
                534,
                9845,
                51688
            ],
            "temperature": 0.0,
            "avg_logprob": -0.19616699474863708,
            "compression_ratio": 1.6221374045801527,
            "no_speech_prob": 0.0154266357421875,
            "words": [
                {
                    "start": 835.1,
                    "end": 835.46,
                    "word": " So",
                    "probability": 0.82373046875
                },
                {
                    "start": 835.46,
                    "end": 835.46,
                    "word": ",",
                    "probability": 0.52978515625
                },
                {
                    "start": 835.46,
                    "end": 835.6,
                    "word": " that's",
                    "probability": 0.939697265625
                },
                {
                    "start": 835.6,
                    "end": 835.86,
                    "word": " great",
                    "probability": 0.8759765625
                },
                {
                    "start": 835.86,
                    "end": 836.08,
                    "word": ".",
                    "probability": 0.52001953125
                },
                {
                    "start": 836.08,
                    "end": 836.26,
                    "word": " But",
                    "probability": 0.8798828125
                },
                {
                    "start": 836.26,
                    "end": 836.46,
                    "word": " now",
                    "probability": 0.81787109375
                },
                {
                    "start": 836.46,
                    "end": 836.58,
                    "word": ",",
                    "probability": 0.462158203125
                },
                {
                    "start": 836.58,
                    "end": 836.68,
                    "word": " if",
                    "probability": 0.8408203125
                },
                {
                    "start": 836.68,
                    "end": 836.74,
                    "word": " I",
                    "probability": 0.9931640625
                },
                {
                    "start": 836.74,
                    "end": 836.92,
                    "word": " go",
                    "probability": 0.892578125
                },
                {
                    "start": 836.92,
                    "end": 837.08,
                    "word": ",",
                    "probability": 0.81591796875
                },
                {
                    "start": 837.08,
                    "end": 837.26,
                    "word": " okay",
                    "probability": 0.3642578125
                },
                {
                    "start": 837.26,
                    "end": 837.38,
                    "word": ",",
                    "probability": 0.88623046875
                },
                {
                    "start": 837.38,
                    "end": 837.52,
                    "word": " well",
                    "probability": 0.6767578125
                },
                {
                    "start": 837.52,
                    "end": 837.64,
                    "word": ",",
                    "probability": 0.84423828125
                },
                {
                    "start": 837.64,
                    "end": 837.78,
                    "word": " I'm",
                    "probability": 0.99609375
                },
                {
                    "start": 837.78,
                    "end": 837.86,
                    "word": " going",
                    "probability": 0.654296875
                },
                {
                    "start": 837.86,
                    "end": 838.16,
                    "word": " to",
                    "probability": 0.89794921875
                },
                {
                    "start": 838.16,
                    "end": 838.5,
                    "word": " split",
                    "probability": 0.86572265625
                },
                {
                    "start": 838.5,
                    "end": 838.6,
                    "word": " it",
                    "probability": 0.90185546875
                },
                {
                    "start": 838.6,
                    "end": 838.76,
                    "word": " up",
                    "probability": 0.90478515625
                },
                {
                    "start": 838.76,
                    "end": 839.1,
                    "word": " further",
                    "probability": 0.85595703125
                },
                {
                    "start": 839.1,
                    "end": 839.2,
                    "word": ",",
                    "probability": 0.485107421875
                },
                {
                    "start": 839.2,
                    "end": 839.3,
                    "word": " so",
                    "probability": 0.87646484375
                },
                {
                    "start": 839.3,
                    "end": 839.6,
                    "word": " really",
                    "probability": 0.1663818359375
                },
                {
                    "start": 839.6,
                    "end": 839.96,
                    "word": " divide",
                    "probability": 0.8876953125
                }
            ]
        },
        {
            "id": 122,
            "seek": 72878,
            "start": 839.96,
            "end": 844.92,
            "text": " and conquer, and I, so I've got these two nodes, and they're both children of the previous node on",
            "tokens": [
                50364,
                293,
                24136,
                11,
                293,
                286,
                11,
                370,
                286,
                600,
                658,
                613,
                732,
                13891,
                11,
                293,
                436,
                434,
                1293,
                2227,
                295,
                264,
                3894,
                9984,
                322,
                50608
            ],
            "temperature": 0.0,
            "avg_logprob": -0.19906495523265028,
            "compression_ratio": 1.84,
            "no_speech_prob": 0.1688232421875,
            "words": [
                {
                    "start": 839.96,
                    "end": 840.1,
                    "word": " and",
                    "probability": 0.1292724609375
                },
                {
                    "start": 840.1,
                    "end": 840.32,
                    "word": " conquer",
                    "probability": 0.72607421875
                },
                {
                    "start": 840.32,
                    "end": 840.48,
                    "word": ",",
                    "probability": 0.10113525390625
                },
                {
                    "start": 840.48,
                    "end": 840.52,
                    "word": " and",
                    "probability": 0.53955078125
                },
                {
                    "start": 840.52,
                    "end": 840.74,
                    "word": " I",
                    "probability": 0.7109375
                },
                {
                    "start": 840.74,
                    "end": 840.9,
                    "word": ",",
                    "probability": 0.1810302734375
                },
                {
                    "start": 840.9,
                    "end": 841.38,
                    "word": " so",
                    "probability": 0.8818359375
                },
                {
                    "start": 841.38,
                    "end": 841.68,
                    "word": " I've",
                    "probability": 0.91015625
                },
                {
                    "start": 841.68,
                    "end": 841.9,
                    "word": " got",
                    "probability": 0.90234375
                },
                {
                    "start": 841.9,
                    "end": 842.26,
                    "word": " these",
                    "probability": 0.7080078125
                },
                {
                    "start": 842.26,
                    "end": 842.48,
                    "word": " two",
                    "probability": 0.78857421875
                },
                {
                    "start": 842.48,
                    "end": 842.82,
                    "word": " nodes",
                    "probability": 0.80029296875
                },
                {
                    "start": 842.82,
                    "end": 843.08,
                    "word": ",",
                    "probability": 0.497802734375
                },
                {
                    "start": 843.08,
                    "end": 843.16,
                    "word": " and",
                    "probability": 0.68505859375
                },
                {
                    "start": 843.16,
                    "end": 843.34,
                    "word": " they're",
                    "probability": 0.841796875
                },
                {
                    "start": 843.34,
                    "end": 843.58,
                    "word": " both",
                    "probability": 0.8212890625
                },
                {
                    "start": 843.58,
                    "end": 843.96,
                    "word": " children",
                    "probability": 0.689453125
                },
                {
                    "start": 843.96,
                    "end": 844.16,
                    "word": " of",
                    "probability": 0.9013671875
                },
                {
                    "start": 844.16,
                    "end": 844.26,
                    "word": " the",
                    "probability": 0.82373046875
                },
                {
                    "start": 844.26,
                    "end": 844.56,
                    "word": " previous",
                    "probability": 0.755859375
                },
                {
                    "start": 844.56,
                    "end": 844.78,
                    "word": " node",
                    "probability": 0.7197265625
                },
                {
                    "start": 844.78,
                    "end": 844.92,
                    "word": " on",
                    "probability": 0.892578125
                }
            ]
        },
        {
            "id": 123,
            "seek": 72878,
            "start": 844.92,
            "end": 850.22,
            "text": " the previous slide, and I'm going to build their children. So, here's their children.",
            "tokens": [
                50608,
                264,
                3894,
                4137,
                11,
                293,
                286,
                478,
                516,
                281,
                1322,
                641,
                2227,
                13,
                407,
                11,
                510,
                311,
                641,
                2227,
                13,
                50856
            ],
            "temperature": 0.0,
            "avg_logprob": -0.19906495523265028,
            "compression_ratio": 1.84,
            "no_speech_prob": 0.1688232421875,
            "words": [
                {
                    "start": 844.92,
                    "end": 845.04,
                    "word": " the",
                    "probability": 0.7568359375
                },
                {
                    "start": 845.04,
                    "end": 845.24,
                    "word": " previous",
                    "probability": 0.76708984375
                },
                {
                    "start": 845.24,
                    "end": 845.62,
                    "word": " slide",
                    "probability": 0.8564453125
                },
                {
                    "start": 845.62,
                    "end": 845.78,
                    "word": ",",
                    "probability": 0.7451171875
                },
                {
                    "start": 845.78,
                    "end": 846.16,
                    "word": " and",
                    "probability": 0.8916015625
                },
                {
                    "start": 846.16,
                    "end": 846.36,
                    "word": " I'm",
                    "probability": 0.9912109375
                },
                {
                    "start": 846.36,
                    "end": 846.44,
                    "word": " going",
                    "probability": 0.638671875
                },
                {
                    "start": 846.44,
                    "end": 846.52,
                    "word": " to",
                    "probability": 0.9013671875
                },
                {
                    "start": 846.52,
                    "end": 846.78,
                    "word": " build",
                    "probability": 0.87841796875
                },
                {
                    "start": 846.78,
                    "end": 847.12,
                    "word": " their",
                    "probability": 0.89453125
                },
                {
                    "start": 847.12,
                    "end": 847.58,
                    "word": " children",
                    "probability": 0.7119140625
                },
                {
                    "start": 847.58,
                    "end": 848.02,
                    "word": ".",
                    "probability": 0.966796875
                },
                {
                    "start": 848.26,
                    "end": 848.7,
                    "word": " So",
                    "probability": 0.8759765625
                },
                {
                    "start": 848.7,
                    "end": 848.78,
                    "word": ",",
                    "probability": 0.055419921875
                },
                {
                    "start": 848.78,
                    "end": 849.06,
                    "word": " here's",
                    "probability": 0.76708984375
                },
                {
                    "start": 849.06,
                    "end": 849.34,
                    "word": " their",
                    "probability": 0.90283203125
                },
                {
                    "start": 849.34,
                    "end": 849.78,
                    "word": " children",
                    "probability": 0.70947265625
                },
                {
                    "start": 849.78,
                    "end": 850.22,
                    "word": ".",
                    "probability": 0.93896484375
                }
            ]
        },
        {
            "id": 124,
            "seek": 72878,
            "start": 850.88,
            "end": 856.76,
            "text": " So, I divide and conquer their stuff, and I can kind of tell that each of their children have",
            "tokens": [
                50904,
                407,
                11,
                286,
                9845,
                293,
                24136,
                641,
                1507,
                11,
                293,
                286,
                393,
                733,
                295,
                980,
                300,
                1184,
                295,
                641,
                2227,
                362,
                51220
            ],
            "temperature": 0.0,
            "avg_logprob": -0.19906495523265028,
            "compression_ratio": 1.84,
            "no_speech_prob": 0.1688232421875,
            "words": [
                {
                    "start": 850.88,
                    "end": 851.06,
                    "word": " So",
                    "probability": 0.50927734375
                },
                {
                    "start": 851.06,
                    "end": 851.1,
                    "word": ",",
                    "probability": 0.65625
                },
                {
                    "start": 851.1,
                    "end": 851.14,
                    "word": " I",
                    "probability": 0.982421875
                },
                {
                    "start": 851.14,
                    "end": 851.38,
                    "word": " divide",
                    "probability": 0.87890625
                },
                {
                    "start": 851.38,
                    "end": 851.52,
                    "word": " and",
                    "probability": 0.89208984375
                },
                {
                    "start": 851.52,
                    "end": 851.72,
                    "word": " conquer",
                    "probability": 0.865234375
                },
                {
                    "start": 851.72,
                    "end": 852.24,
                    "word": " their",
                    "probability": 0.390625
                },
                {
                    "start": 852.24,
                    "end": 852.96,
                    "word": " stuff",
                    "probability": 0.791015625
                },
                {
                    "start": 852.96,
                    "end": 853.24,
                    "word": ",",
                    "probability": 0.806640625
                },
                {
                    "start": 853.24,
                    "end": 853.76,
                    "word": " and",
                    "probability": 0.75927734375
                },
                {
                    "start": 853.76,
                    "end": 853.98,
                    "word": " I",
                    "probability": 0.9873046875
                },
                {
                    "start": 853.98,
                    "end": 854.18,
                    "word": " can",
                    "probability": 0.89404296875
                },
                {
                    "start": 854.18,
                    "end": 854.48,
                    "word": " kind",
                    "probability": 0.72119140625
                },
                {
                    "start": 854.48,
                    "end": 854.58,
                    "word": " of",
                    "probability": 0.90283203125
                },
                {
                    "start": 854.58,
                    "end": 854.8,
                    "word": " tell",
                    "probability": 0.82763671875
                },
                {
                    "start": 854.8,
                    "end": 855.04,
                    "word": " that",
                    "probability": 0.8623046875
                },
                {
                    "start": 855.04,
                    "end": 855.92,
                    "word": " each",
                    "probability": 0.7109375
                },
                {
                    "start": 855.92,
                    "end": 856.06,
                    "word": " of",
                    "probability": 0.90869140625
                },
                {
                    "start": 856.06,
                    "end": 856.2,
                    "word": " their",
                    "probability": 0.89892578125
                },
                {
                    "start": 856.2,
                    "end": 856.46,
                    "word": " children",
                    "probability": 0.71044921875
                },
                {
                    "start": 856.46,
                    "end": 856.76,
                    "word": " have",
                    "probability": 0.85888671875
                }
            ]
        },
        {
            "id": 125,
            "seek": 72878,
            "start": 856.76,
            "end": 863.58,
            "text": " four, and if I, if you squint a little bit, it's kind of square-shaped in this dimension.",
            "tokens": [
                51220,
                1451,
                11,
                293,
                498,
                286,
                11,
                498,
                291,
                2339,
                686,
                257,
                707,
                857,
                11,
                309,
                311,
                733,
                295,
                3732,
                12,
                23103,
                294,
                341,
                10139,
                13,
                51524
            ],
            "temperature": 0.0,
            "avg_logprob": -0.19906495523265028,
            "compression_ratio": 1.84,
            "no_speech_prob": 0.1688232421875,
            "words": [
                {
                    "start": 856.76,
                    "end": 857.6,
                    "word": " four",
                    "probability": 0.82421875
                },
                {
                    "start": 857.6,
                    "end": 858.06,
                    "word": ",",
                    "probability": 0.6474609375
                },
                {
                    "start": 858.06,
                    "end": 858.34,
                    "word": " and",
                    "probability": 0.89404296875
                },
                {
                    "start": 858.34,
                    "end": 858.56,
                    "word": " if",
                    "probability": 0.88134765625
                },
                {
                    "start": 858.56,
                    "end": 858.66,
                    "word": " I",
                    "probability": 0.52783203125
                },
                {
                    "start": 858.66,
                    "end": 858.94,
                    "word": ",",
                    "probability": 0.84619140625
                },
                {
                    "start": 858.94,
                    "end": 859.16,
                    "word": " if",
                    "probability": 0.90185546875
                },
                {
                    "start": 859.16,
                    "end": 859.26,
                    "word": " you",
                    "probability": 0.89111328125
                },
                {
                    "start": 859.26,
                    "end": 859.52,
                    "word": " squint",
                    "probability": 0.978515625
                },
                {
                    "start": 859.64,
                    "end": 859.74,
                    "word": " a little",
                    "probability": 0.80224609375
                },
                {
                    "start": 859.74,
                    "end": 859.88,
                    "word": " bit",
                    "probability": 0.88818359375
                },
                {
                    "start": 859.88,
                    "end": 859.92,
                    "word": ",",
                    "probability": 0.83251953125
                },
                {
                    "start": 859.92,
                    "end": 860.08,
                    "word": " it's",
                    "probability": 0.946533203125
                },
                {
                    "start": 860.08,
                    "end": 860.26,
                    "word": " kind",
                    "probability": 0.8203125
                },
                {
                    "start": 860.26,
                    "end": 860.54,
                    "word": " of",
                    "probability": 0.90869140625
                },
                {
                    "start": 860.54,
                    "end": 861.24,
                    "word": " square",
                    "probability": 0.8349609375
                },
                {
                    "start": 861.24,
                    "end": 861.62,
                    "word": "-shaped",
                    "probability": 0.65234375
                },
                {
                    "start": 861.62,
                    "end": 862.62,
                    "word": " in",
                    "probability": 0.78564453125
                },
                {
                    "start": 862.62,
                    "end": 862.78,
                    "word": " this",
                    "probability": 0.8779296875
                },
                {
                    "start": 862.78,
                    "end": 863.14,
                    "word": " dimension",
                    "probability": 0.90380859375
                },
                {
                    "start": 863.14,
                    "end": 863.58,
                    "word": ".",
                    "probability": 0.6220703125
                }
            ]
        },
        {
            "id": 126,
            "seek": 72878,
            "start": 863.74,
            "end": 869.0,
            "text": " At the scale, the top scale, it was too blurry. I could only see a line here. At the scale,",
            "tokens": [
                51556,
                1711,
                264,
                4373,
                11,
                264,
                1192,
                4373,
                11,
                309,
                390,
                886,
                37644,
                13,
                286,
                727,
                787,
                536,
                257,
                1622,
                510,
                13,
                1711,
                264,
                4373,
                11,
                51812
            ],
            "temperature": 0.0,
            "avg_logprob": -0.19906495523265028,
            "compression_ratio": 1.84,
            "no_speech_prob": 0.1688232421875,
            "words": [
                {
                    "start": 863.74,
                    "end": 864.06,
                    "word": " At",
                    "probability": 0.9140625
                },
                {
                    "start": 864.06,
                    "end": 864.2,
                    "word": " the",
                    "probability": 0.68603515625
                },
                {
                    "start": 864.2,
                    "end": 864.58,
                    "word": " scale",
                    "probability": 0.80517578125
                },
                {
                    "start": 864.58,
                    "end": 865.4,
                    "word": ",",
                    "probability": 0.91845703125
                },
                {
                    "start": 865.4,
                    "end": 865.66,
                    "word": " the",
                    "probability": 0.58154296875
                },
                {
                    "start": 865.66,
                    "end": 865.94,
                    "word": " top",
                    "probability": 0.822265625
                },
                {
                    "start": 865.94,
                    "end": 866.24,
                    "word": " scale",
                    "probability": 0.81396484375
                },
                {
                    "start": 866.24,
                    "end": 866.3,
                    "word": ",",
                    "probability": 0.67041015625
                },
                {
                    "start": 866.3,
                    "end": 866.44,
                    "word": " it",
                    "probability": 0.89501953125
                },
                {
                    "start": 866.44,
                    "end": 866.58,
                    "word": " was",
                    "probability": 0.8896484375
                },
                {
                    "start": 866.58,
                    "end": 866.86,
                    "word": " too",
                    "probability": 0.81298828125
                },
                {
                    "start": 866.86,
                    "end": 867.18,
                    "word": " blurry",
                    "probability": 0.78173828125
                },
                {
                    "start": 867.18,
                    "end": 867.36,
                    "word": ".",
                    "probability": 0.412353515625
                },
                {
                    "start": 867.36,
                    "end": 867.46,
                    "word": " I",
                    "probability": 0.9951171875
                },
                {
                    "start": 867.46,
                    "end": 867.58,
                    "word": " could",
                    "probability": 0.82177734375
                },
                {
                    "start": 867.58,
                    "end": 867.74,
                    "word": " only",
                    "probability": 0.861328125
                },
                {
                    "start": 867.74,
                    "end": 867.92,
                    "word": " see",
                    "probability": 0.81787109375
                },
                {
                    "start": 868.04,
                    "end": 868.2,
                    "word": " a line",
                    "probability": 0.88623046875
                },
                {
                    "start": 868.2,
                    "end": 868.44,
                    "word": " here",
                    "probability": 0.80712890625
                },
                {
                    "start": 868.44,
                    "end": 868.5,
                    "word": ".",
                    "probability": 0.84814453125
                },
                {
                    "start": 868.5,
                    "end": 868.62,
                    "word": " At",
                    "probability": 0.90869140625
                },
                {
                    "start": 868.62,
                    "end": 868.76,
                    "word": " the",
                    "probability": 0.77734375
                },
                {
                    "start": 868.76,
                    "end": 868.98,
                    "word": " scale",
                    "probability": 0.8251953125
                },
                {
                    "start": 868.98,
                    "end": 869.0,
                    "word": ",",
                    "probability": 0.86962890625
                }
            ]
        },
        {
            "id": 127,
            "seek": 75572,
            "start": 869.0,
            "end": 874.9,
            "text": " I can kind of see that it's up in, it's got a nice shape, square shape, and so there's four",
            "tokens": [
                50364,
                286,
                393,
                733,
                295,
                536,
                300,
                309,
                311,
                493,
                294,
                11,
                309,
                311,
                658,
                257,
                1481,
                3909,
                11,
                3732,
                3909,
                11,
                293,
                370,
                456,
                311,
                1451,
                50656
            ],
            "temperature": 0.0,
            "avg_logprob": -0.19440406976744187,
            "compression_ratio": 1.9831223628691983,
            "no_speech_prob": 0.00679779052734375,
            "words": [
                {
                    "start": 869.0,
                    "end": 869.14,
                    "word": " I",
                    "probability": 0.284423828125
                },
                {
                    "start": 869.14,
                    "end": 869.22,
                    "word": " can",
                    "probability": 0.64794921875
                },
                {
                    "start": 869.22,
                    "end": 869.42,
                    "word": " kind",
                    "probability": 0.328369140625
                },
                {
                    "start": 869.42,
                    "end": 869.48,
                    "word": " of",
                    "probability": 0.89306640625
                },
                {
                    "start": 869.48,
                    "end": 869.62,
                    "word": " see",
                    "probability": 0.8125
                },
                {
                    "start": 869.62,
                    "end": 869.78,
                    "word": " that",
                    "probability": 0.81005859375
                },
                {
                    "start": 869.78,
                    "end": 870.06,
                    "word": " it's",
                    "probability": 0.827392578125
                },
                {
                    "start": 870.06,
                    "end": 870.46,
                    "word": " up",
                    "probability": 0.357177734375
                },
                {
                    "start": 870.46,
                    "end": 870.64,
                    "word": " in",
                    "probability": 0.26318359375
                },
                {
                    "start": 870.64,
                    "end": 870.72,
                    "word": ",",
                    "probability": 0.146484375
                },
                {
                    "start": 870.72,
                    "end": 870.98,
                    "word": " it's",
                    "probability": 0.9296875
                },
                {
                    "start": 870.98,
                    "end": 871.12,
                    "word": " got",
                    "probability": 0.88720703125
                },
                {
                    "start": 871.22,
                    "end": 871.46,
                    "word": " a nice",
                    "probability": 0.8115234375
                },
                {
                    "start": 871.46,
                    "end": 871.74,
                    "word": " shape",
                    "probability": 0.828125
                },
                {
                    "start": 871.74,
                    "end": 872.26,
                    "word": ",",
                    "probability": 0.765625
                },
                {
                    "start": 872.26,
                    "end": 872.64,
                    "word": " square",
                    "probability": 0.78271484375
                },
                {
                    "start": 872.64,
                    "end": 872.92,
                    "word": " shape",
                    "probability": 0.7939453125
                },
                {
                    "start": 872.92,
                    "end": 873.36,
                    "word": ",",
                    "probability": 0.1114501953125
                },
                {
                    "start": 873.36,
                    "end": 874.08,
                    "word": " and",
                    "probability": 0.6748046875
                },
                {
                    "start": 874.08,
                    "end": 874.28,
                    "word": " so",
                    "probability": 0.8896484375
                },
                {
                    "start": 874.28,
                    "end": 874.58,
                    "word": " there's",
                    "probability": 0.890869140625
                },
                {
                    "start": 874.58,
                    "end": 874.9,
                    "word": " four",
                    "probability": 0.78125
                }
            ]
        },
        {
            "id": 128,
            "seek": 75572,
            "start": 874.9,
            "end": 880.12,
            "text": " children, and the reason, you know, that that dimensionality is relevant to the number of",
            "tokens": [
                50656,
                2227,
                11,
                293,
                264,
                1778,
                11,
                291,
                458,
                11,
                300,
                300,
                10139,
                1860,
                307,
                7340,
                281,
                264,
                1230,
                295,
                50916
            ],
            "temperature": 0.0,
            "avg_logprob": -0.19440406976744187,
            "compression_ratio": 1.9831223628691983,
            "no_speech_prob": 0.00679779052734375,
            "words": [
                {
                    "start": 874.9,
                    "end": 875.28,
                    "word": " children",
                    "probability": 0.68505859375
                },
                {
                    "start": 875.28,
                    "end": 875.46,
                    "word": ",",
                    "probability": 0.22705078125
                },
                {
                    "start": 875.46,
                    "end": 875.48,
                    "word": " and",
                    "probability": 0.86962890625
                },
                {
                    "start": 875.48,
                    "end": 875.6,
                    "word": " the",
                    "probability": 0.77587890625
                },
                {
                    "start": 875.6,
                    "end": 875.82,
                    "word": " reason",
                    "probability": 0.9482421875
                },
                {
                    "start": 875.82,
                    "end": 876.16,
                    "word": ",",
                    "probability": 0.60009765625
                },
                {
                    "start": 876.16,
                    "end": 876.3,
                    "word": " you",
                    "probability": 0.8359375
                },
                {
                    "start": 876.3,
                    "end": 876.44,
                    "word": " know",
                    "probability": 0.81982421875
                },
                {
                    "start": 876.44,
                    "end": 876.66,
                    "word": ",",
                    "probability": 0.91552734375
                },
                {
                    "start": 876.66,
                    "end": 876.88,
                    "word": " that",
                    "probability": 0.81591796875
                },
                {
                    "start": 876.88,
                    "end": 877.12,
                    "word": " that",
                    "probability": 0.344970703125
                },
                {
                    "start": 877.12,
                    "end": 878.28,
                    "word": " dimensionality",
                    "probability": 0.776611328125
                },
                {
                    "start": 878.28,
                    "end": 878.64,
                    "word": " is",
                    "probability": 0.80712890625
                },
                {
                    "start": 878.64,
                    "end": 879.56,
                    "word": " relevant",
                    "probability": 0.86669921875
                },
                {
                    "start": 879.56,
                    "end": 879.78,
                    "word": " to",
                    "probability": 0.89501953125
                },
                {
                    "start": 879.78,
                    "end": 879.88,
                    "word": " the",
                    "probability": 0.80224609375
                },
                {
                    "start": 879.88,
                    "end": 880.0,
                    "word": " number",
                    "probability": 0.8896484375
                },
                {
                    "start": 880.0,
                    "end": 880.12,
                    "word": " of",
                    "probability": 0.90234375
                }
            ]
        },
        {
            "id": 129,
            "seek": 75572,
            "start": 880.12,
            "end": 885.86,
            "text": " children you have. And then I can go further, and you know, it's got some more, it's kind of split",
            "tokens": [
                50916,
                2227,
                291,
                362,
                13,
                400,
                550,
                286,
                393,
                352,
                3052,
                11,
                293,
                291,
                458,
                11,
                309,
                311,
                658,
                512,
                544,
                11,
                309,
                311,
                733,
                295,
                7472,
                51208
            ],
            "temperature": 0.0,
            "avg_logprob": -0.19440406976744187,
            "compression_ratio": 1.9831223628691983,
            "no_speech_prob": 0.00679779052734375,
            "words": [
                {
                    "start": 880.12,
                    "end": 880.38,
                    "word": " children",
                    "probability": 0.7060546875
                },
                {
                    "start": 880.38,
                    "end": 880.6,
                    "word": " you",
                    "probability": 0.8896484375
                },
                {
                    "start": 880.6,
                    "end": 880.96,
                    "word": " have",
                    "probability": 0.88232421875
                },
                {
                    "start": 880.96,
                    "end": 881.36,
                    "word": ".",
                    "probability": 0.7216796875
                },
                {
                    "start": 881.54,
                    "end": 881.94,
                    "word": " And",
                    "probability": 0.86083984375
                },
                {
                    "start": 881.94,
                    "end": 882.06,
                    "word": " then",
                    "probability": 0.7236328125
                },
                {
                    "start": 882.06,
                    "end": 882.2,
                    "word": " I",
                    "probability": 0.9765625
                },
                {
                    "start": 882.2,
                    "end": 882.34,
                    "word": " can",
                    "probability": 0.89990234375
                },
                {
                    "start": 882.34,
                    "end": 882.5,
                    "word": " go",
                    "probability": 0.9013671875
                },
                {
                    "start": 882.5,
                    "end": 882.88,
                    "word": " further",
                    "probability": 0.84814453125
                },
                {
                    "start": 882.88,
                    "end": 883.06,
                    "word": ",",
                    "probability": 0.476318359375
                },
                {
                    "start": 883.06,
                    "end": 883.22,
                    "word": " and",
                    "probability": 0.88818359375
                },
                {
                    "start": 883.22,
                    "end": 883.68,
                    "word": " you",
                    "probability": 0.66064453125
                },
                {
                    "start": 883.68,
                    "end": 883.8,
                    "word": " know",
                    "probability": 0.81201171875
                },
                {
                    "start": 883.8,
                    "end": 883.88,
                    "word": ",",
                    "probability": 0.76708984375
                },
                {
                    "start": 883.88,
                    "end": 883.98,
                    "word": " it's",
                    "probability": 0.909423828125
                },
                {
                    "start": 883.98,
                    "end": 884.1,
                    "word": " got",
                    "probability": 0.869140625
                },
                {
                    "start": 884.1,
                    "end": 884.26,
                    "word": " some",
                    "probability": 0.7705078125
                },
                {
                    "start": 884.26,
                    "end": 884.44,
                    "word": " more",
                    "probability": 0.84521484375
                },
                {
                    "start": 884.44,
                    "end": 884.56,
                    "word": ",",
                    "probability": 0.77880859375
                },
                {
                    "start": 884.56,
                    "end": 885.32,
                    "word": " it's",
                    "probability": 0.94775390625
                },
                {
                    "start": 885.32,
                    "end": 885.52,
                    "word": " kind",
                    "probability": 0.705078125
                },
                {
                    "start": 885.52,
                    "end": 885.64,
                    "word": " of",
                    "probability": 0.90625
                },
                {
                    "start": 885.64,
                    "end": 885.86,
                    "word": " split",
                    "probability": 0.8603515625
                }
            ]
        },
        {
            "id": 130,
            "seek": 75572,
            "start": 885.86,
            "end": 890.68,
            "text": " up nicely, and I've got like a one-dimensional object here again, because all these children",
            "tokens": [
                51208,
                493,
                9594,
                11,
                293,
                286,
                600,
                658,
                411,
                257,
                472,
                12,
                18759,
                2657,
                510,
                797,
                11,
                570,
                439,
                613,
                2227,
                51448
            ],
            "temperature": 0.0,
            "avg_logprob": -0.19440406976744187,
            "compression_ratio": 1.9831223628691983,
            "no_speech_prob": 0.00679779052734375,
            "words": [
                {
                    "start": 885.86,
                    "end": 886.06,
                    "word": " up",
                    "probability": 0.9052734375
                },
                {
                    "start": 886.06,
                    "end": 886.44,
                    "word": " nicely",
                    "probability": 0.8583984375
                },
                {
                    "start": 886.44,
                    "end": 886.5,
                    "word": ",",
                    "probability": 0.78125
                },
                {
                    "start": 886.5,
                    "end": 886.6,
                    "word": " and",
                    "probability": 0.8974609375
                },
                {
                    "start": 886.6,
                    "end": 886.74,
                    "word": " I've",
                    "probability": 0.969970703125
                },
                {
                    "start": 886.74,
                    "end": 886.94,
                    "word": " got",
                    "probability": 0.892578125
                },
                {
                    "start": 886.94,
                    "end": 887.18,
                    "word": " like",
                    "probability": 0.732421875
                },
                {
                    "start": 887.6,
                    "end": 887.84,
                    "word": " a one",
                    "probability": 0.8115234375
                },
                {
                    "start": 887.84,
                    "end": 888.18,
                    "word": "-dimensional",
                    "probability": 0.7734375
                },
                {
                    "start": 888.18,
                    "end": 888.86,
                    "word": " object",
                    "probability": 0.900390625
                },
                {
                    "start": 888.86,
                    "end": 889.26,
                    "word": " here",
                    "probability": 0.8046875
                },
                {
                    "start": 889.26,
                    "end": 889.56,
                    "word": " again",
                    "probability": 0.86962890625
                },
                {
                    "start": 889.56,
                    "end": 889.58,
                    "word": ",",
                    "probability": 0.75
                },
                {
                    "start": 889.58,
                    "end": 889.8,
                    "word": " because",
                    "probability": 0.8330078125
                },
                {
                    "start": 889.8,
                    "end": 890.02,
                    "word": " all",
                    "probability": 0.43505859375
                },
                {
                    "start": 890.02,
                    "end": 890.36,
                    "word": " these",
                    "probability": 0.68896484375
                },
                {
                    "start": 890.36,
                    "end": 890.68,
                    "word": " children",
                    "probability": 0.68603515625
                }
            ]
        },
        {
            "id": 131,
            "seek": 75572,
            "start": 890.68,
            "end": 895.94,
            "text": " are kind of one-dimensional-ish, and I can split up some more, and you can see that it's kind of",
            "tokens": [
                51448,
                366,
                733,
                295,
                472,
                12,
                18759,
                12,
                742,
                11,
                293,
                286,
                393,
                7472,
                493,
                512,
                544,
                11,
                293,
                291,
                393,
                536,
                300,
                309,
                311,
                733,
                295,
                51708
            ],
            "temperature": 0.0,
            "avg_logprob": -0.19440406976744187,
            "compression_ratio": 1.9831223628691983,
            "no_speech_prob": 0.00679779052734375,
            "words": [
                {
                    "start": 890.68,
                    "end": 890.88,
                    "word": " are",
                    "probability": 0.8974609375
                },
                {
                    "start": 890.88,
                    "end": 891.06,
                    "word": " kind",
                    "probability": 0.7890625
                },
                {
                    "start": 891.06,
                    "end": 891.18,
                    "word": " of",
                    "probability": 0.90673828125
                },
                {
                    "start": 891.18,
                    "end": 891.46,
                    "word": " one",
                    "probability": 0.80224609375
                },
                {
                    "start": 891.46,
                    "end": 891.82,
                    "word": "-dimensional",
                    "probability": 0.80810546875
                },
                {
                    "start": 891.82,
                    "end": 892.08,
                    "word": "-ish",
                    "probability": 0.78369140625
                },
                {
                    "start": 892.08,
                    "end": 892.44,
                    "word": ",",
                    "probability": 0.677734375
                },
                {
                    "start": 892.44,
                    "end": 893.04,
                    "word": " and",
                    "probability": 0.89501953125
                },
                {
                    "start": 893.04,
                    "end": 893.18,
                    "word": " I",
                    "probability": 0.9931640625
                },
                {
                    "start": 893.18,
                    "end": 893.3,
                    "word": " can",
                    "probability": 0.88818359375
                },
                {
                    "start": 893.3,
                    "end": 893.5,
                    "word": " split",
                    "probability": 0.8251953125
                },
                {
                    "start": 893.5,
                    "end": 893.64,
                    "word": " up",
                    "probability": 0.84521484375
                },
                {
                    "start": 893.64,
                    "end": 893.78,
                    "word": " some",
                    "probability": 0.81005859375
                },
                {
                    "start": 893.78,
                    "end": 894.1,
                    "word": " more",
                    "probability": 0.87744140625
                },
                {
                    "start": 894.1,
                    "end": 894.24,
                    "word": ",",
                    "probability": 0.88671875
                },
                {
                    "start": 894.24,
                    "end": 894.66,
                    "word": " and",
                    "probability": 0.8955078125
                },
                {
                    "start": 894.66,
                    "end": 894.86,
                    "word": " you",
                    "probability": 0.89501953125
                },
                {
                    "start": 894.86,
                    "end": 894.98,
                    "word": " can",
                    "probability": 0.8984375
                },
                {
                    "start": 894.98,
                    "end": 895.3,
                    "word": " see",
                    "probability": 0.81396484375
                },
                {
                    "start": 895.3,
                    "end": 895.46,
                    "word": " that",
                    "probability": 0.869140625
                },
                {
                    "start": 895.46,
                    "end": 895.6,
                    "word": " it's",
                    "probability": 0.9140625
                },
                {
                    "start": 895.6,
                    "end": 895.78,
                    "word": " kind",
                    "probability": 0.814453125
                },
                {
                    "start": 895.78,
                    "end": 895.94,
                    "word": " of",
                    "probability": 0.9013671875
                }
            ]
        },
        {
            "id": 132,
            "seek": 78176,
            "start": 895.94,
            "end": 900.98,
            "text": " conforming to your data nicely, and at every single step of the way, I can make some inferences about",
            "tokens": [
                50380,
                18975,
                278,
                281,
                428,
                1412,
                9594,
                11,
                293,
                412,
                633,
                2167,
                1823,
                295,
                264,
                636,
                11,
                286,
                393,
                652,
                512,
                13596,
                2667,
                466,
                50616
            ],
            "temperature": 0.0,
            "avg_logprob": -0.20329483128760173,
            "compression_ratio": 1.6255506607929515,
            "no_speech_prob": 0.08758544921875,
            "words": [
                {
                    "start": 895.94,
                    "end": 896.94,
                    "word": " conforming",
                    "probability": 0.4800682067871094
                },
                {
                    "start": 896.94,
                    "end": 897.02,
                    "word": " to",
                    "probability": 0.88818359375
                },
                {
                    "start": 897.02,
                    "end": 897.18,
                    "word": " your",
                    "probability": 0.76904296875
                },
                {
                    "start": 897.18,
                    "end": 897.36,
                    "word": " data",
                    "probability": 0.8203125
                },
                {
                    "start": 897.36,
                    "end": 897.76,
                    "word": " nicely",
                    "probability": 0.701171875
                },
                {
                    "start": 897.76,
                    "end": 897.98,
                    "word": ",",
                    "probability": 0.18017578125
                },
                {
                    "start": 897.98,
                    "end": 898.02,
                    "word": " and",
                    "probability": 0.80859375
                },
                {
                    "start": 898.02,
                    "end": 898.74,
                    "word": " at",
                    "probability": 0.60498046875
                },
                {
                    "start": 898.74,
                    "end": 898.94,
                    "word": " every",
                    "probability": 0.724609375
                },
                {
                    "start": 898.94,
                    "end": 899.18,
                    "word": " single",
                    "probability": 0.54541015625
                },
                {
                    "start": 899.18,
                    "end": 899.36,
                    "word": " step",
                    "probability": 0.79296875
                },
                {
                    "start": 899.36,
                    "end": 899.48,
                    "word": " of",
                    "probability": 0.8896484375
                },
                {
                    "start": 899.48,
                    "end": 899.52,
                    "word": " the",
                    "probability": 0.8125
                },
                {
                    "start": 899.52,
                    "end": 899.66,
                    "word": " way",
                    "probability": 0.89306640625
                },
                {
                    "start": 899.66,
                    "end": 899.76,
                    "word": ",",
                    "probability": 0.311767578125
                },
                {
                    "start": 899.76,
                    "end": 899.82,
                    "word": " I",
                    "probability": 0.9716796875
                },
                {
                    "start": 899.82,
                    "end": 899.94,
                    "word": " can",
                    "probability": 0.88623046875
                },
                {
                    "start": 899.94,
                    "end": 900.1,
                    "word": " make",
                    "probability": 0.8896484375
                },
                {
                    "start": 900.1,
                    "end": 900.28,
                    "word": " some",
                    "probability": 0.7666015625
                },
                {
                    "start": 900.28,
                    "end": 900.68,
                    "word": " inferences",
                    "probability": 0.89208984375
                },
                {
                    "start": 900.68,
                    "end": 900.98,
                    "word": " about",
                    "probability": 0.78466796875
                }
            ]
        },
        {
            "id": 133,
            "seek": 78176,
            "start": 900.98,
            "end": 906.38,
            "text": " what the shape of my data is. They're very cheap, because it's basically counting.",
            "tokens": [
                50616,
                437,
                264,
                3909,
                295,
                452,
                1412,
                307,
                13,
                814,
                434,
                588,
                7084,
                11,
                570,
                309,
                311,
                1936,
                13251,
                13,
                50860
            ],
            "temperature": 0.0,
            "avg_logprob": -0.20329483128760173,
            "compression_ratio": 1.6255506607929515,
            "no_speech_prob": 0.08758544921875,
            "words": [
                {
                    "start": 900.98,
                    "end": 901.32,
                    "word": " what",
                    "probability": 0.8193359375
                },
                {
                    "start": 901.32,
                    "end": 901.82,
                    "word": " the",
                    "probability": 0.73681640625
                },
                {
                    "start": 901.82,
                    "end": 902.58,
                    "word": " shape",
                    "probability": 0.81689453125
                },
                {
                    "start": 902.58,
                    "end": 902.7,
                    "word": " of",
                    "probability": 0.9091796875
                },
                {
                    "start": 902.7,
                    "end": 902.84,
                    "word": " my",
                    "probability": 0.89697265625
                },
                {
                    "start": 902.84,
                    "end": 903.08,
                    "word": " data",
                    "probability": 0.82373046875
                },
                {
                    "start": 903.08,
                    "end": 903.36,
                    "word": " is",
                    "probability": 0.896484375
                },
                {
                    "start": 903.36,
                    "end": 903.48,
                    "word": ".",
                    "probability": 0.28271484375
                },
                {
                    "start": 903.48,
                    "end": 903.82,
                    "word": " They're",
                    "probability": 0.65576171875
                },
                {
                    "start": 903.82,
                    "end": 904.1,
                    "word": " very",
                    "probability": 0.724609375
                },
                {
                    "start": 904.1,
                    "end": 904.5,
                    "word": " cheap",
                    "probability": 0.8359375
                },
                {
                    "start": 904.5,
                    "end": 904.62,
                    "word": ",",
                    "probability": 0.3994140625
                },
                {
                    "start": 904.62,
                    "end": 904.84,
                    "word": " because",
                    "probability": 0.82275390625
                },
                {
                    "start": 904.84,
                    "end": 905.04,
                    "word": " it's",
                    "probability": 0.93505859375
                },
                {
                    "start": 905.04,
                    "end": 905.42,
                    "word": " basically",
                    "probability": 0.83203125
                },
                {
                    "start": 905.42,
                    "end": 905.86,
                    "word": " counting",
                    "probability": 0.64794921875
                },
                {
                    "start": 905.86,
                    "end": 906.38,
                    "word": ".",
                    "probability": 0.9638671875
                }
            ]
        },
        {
            "id": 134,
            "seek": 78176,
            "start": 906.98,
            "end": 913.74,
            "text": " So how do we build a probability distribution? So on slide two, we built a very simple",
            "tokens": [
                50916,
                407,
                577,
                360,
                321,
                1322,
                257,
                8482,
                7316,
                30,
                407,
                322,
                4137,
                732,
                11,
                321,
                3094,
                257,
                588,
                2199,
                51248
            ],
            "temperature": 0.0,
            "avg_logprob": -0.20329483128760173,
            "compression_ratio": 1.6255506607929515,
            "no_speech_prob": 0.08758544921875,
            "words": [
                {
                    "start": 906.98,
                    "end": 907.34,
                    "word": " So",
                    "probability": 0.806640625
                },
                {
                    "start": 907.34,
                    "end": 907.58,
                    "word": " how",
                    "probability": 0.382080078125
                },
                {
                    "start": 907.58,
                    "end": 907.72,
                    "word": " do",
                    "probability": 0.8935546875
                },
                {
                    "start": 907.72,
                    "end": 907.84,
                    "word": " we",
                    "probability": 0.9033203125
                },
                {
                    "start": 907.84,
                    "end": 908.14,
                    "word": " build",
                    "probability": 0.86669921875
                },
                {
                    "start": 908.72,
                    "end": 909.04,
                    "word": " a probability",
                    "probability": 0.86083984375
                },
                {
                    "start": 909.04,
                    "end": 909.62,
                    "word": " distribution",
                    "probability": 0.71044921875
                },
                {
                    "start": 909.62,
                    "end": 910.14,
                    "word": "?",
                    "probability": 0.9736328125
                },
                {
                    "start": 910.2,
                    "end": 910.46,
                    "word": " So",
                    "probability": 0.495361328125
                },
                {
                    "start": 910.46,
                    "end": 911.64,
                    "word": " on",
                    "probability": 0.61279296875
                },
                {
                    "start": 911.64,
                    "end": 911.9,
                    "word": " slide",
                    "probability": 0.82861328125
                },
                {
                    "start": 911.9,
                    "end": 912.22,
                    "word": " two",
                    "probability": 0.474853515625
                },
                {
                    "start": 912.22,
                    "end": 912.46,
                    "word": ",",
                    "probability": 0.892578125
                },
                {
                    "start": 912.46,
                    "end": 912.7,
                    "word": " we",
                    "probability": 0.9033203125
                },
                {
                    "start": 912.7,
                    "end": 912.96,
                    "word": " built",
                    "probability": 0.90283203125
                },
                {
                    "start": 913.2,
                    "end": 913.36,
                    "word": " a very",
                    "probability": 0.74267578125
                },
                {
                    "start": 913.36,
                    "end": 913.74,
                    "word": " simple",
                    "probability": 0.91455078125
                }
            ]
        },
        {
            "id": 135,
            "seek": 78176,
            "start": 913.74,
            "end": 921.98,
            "text": " probability distribution, where I had a discrete space, and now I've got a tree, and essentially,",
            "tokens": [
                51276,
                8482,
                7316,
                11,
                689,
                286,
                632,
                257,
                27706,
                1901,
                11,
                293,
                586,
                286,
                600,
                658,
                257,
                4230,
                11,
                293,
                4476,
                11,
                51656
            ],
            "temperature": 0.0,
            "avg_logprob": -0.20329483128760173,
            "compression_ratio": 1.6255506607929515,
            "no_speech_prob": 0.08758544921875,
            "words": [
                {
                    "start": 913.74,
                    "end": 914.66,
                    "word": " probability",
                    "probability": 0.86376953125
                },
                {
                    "start": 914.66,
                    "end": 915.28,
                    "word": " distribution",
                    "probability": 0.75341796875
                },
                {
                    "start": 915.28,
                    "end": 915.94,
                    "word": ",",
                    "probability": 0.444580078125
                },
                {
                    "start": 915.94,
                    "end": 916.38,
                    "word": " where",
                    "probability": 0.9033203125
                },
                {
                    "start": 916.38,
                    "end": 916.66,
                    "word": " I",
                    "probability": 0.99609375
                },
                {
                    "start": 916.66,
                    "end": 917.32,
                    "word": " had",
                    "probability": 0.81005859375
                },
                {
                    "start": 918.22,
                    "end": 918.56,
                    "word": " a discrete",
                    "probability": 0.85302734375
                },
                {
                    "start": 918.56,
                    "end": 918.98,
                    "word": " space",
                    "probability": 0.81884765625
                },
                {
                    "start": 918.98,
                    "end": 919.22,
                    "word": ",",
                    "probability": 0.46142578125
                },
                {
                    "start": 919.22,
                    "end": 919.4,
                    "word": " and",
                    "probability": 0.8896484375
                },
                {
                    "start": 919.4,
                    "end": 919.66,
                    "word": " now",
                    "probability": 0.8857421875
                },
                {
                    "start": 919.66,
                    "end": 919.92,
                    "word": " I've",
                    "probability": 0.93798828125
                },
                {
                    "start": 919.92,
                    "end": 919.98,
                    "word": " got",
                    "probability": 0.88720703125
                },
                {
                    "start": 920.18,
                    "end": 920.44,
                    "word": " a tree",
                    "probability": 0.77783203125
                },
                {
                    "start": 920.44,
                    "end": 920.86,
                    "word": ",",
                    "probability": 0.337158203125
                },
                {
                    "start": 920.86,
                    "end": 920.96,
                    "word": " and",
                    "probability": 0.89208984375
                },
                {
                    "start": 920.96,
                    "end": 921.4,
                    "word": " essentially",
                    "probability": 0.64990234375
                },
                {
                    "start": 921.4,
                    "end": 921.98,
                    "word": ",",
                    "probability": 0.7265625
                }
            ]
        },
        {
            "id": 136,
            "seek": 80724,
            "start": 921.98,
            "end": 926.54,
            "text": " when I have a tree, at every single node, I can either go left, or right, or down, or you know,",
            "tokens": [
                50364,
                562,
                286,
                362,
                257,
                4230,
                11,
                412,
                633,
                2167,
                9984,
                11,
                286,
                393,
                2139,
                352,
                1411,
                11,
                420,
                558,
                11,
                420,
                760,
                11,
                420,
                291,
                458,
                11,
                50592
            ],
            "temperature": 0.0,
            "avg_logprob": -0.21116504333551647,
            "compression_ratio": 1.6712962962962963,
            "no_speech_prob": 0.0014781951904296875,
            "words": [
                {
                    "start": 921.98,
                    "end": 922.12,
                    "word": " when",
                    "probability": 0.1845703125
                },
                {
                    "start": 922.12,
                    "end": 922.18,
                    "word": " I",
                    "probability": 0.9404296875
                },
                {
                    "start": 922.18,
                    "end": 922.3,
                    "word": " have",
                    "probability": 0.880859375
                },
                {
                    "start": 922.42,
                    "end": 922.72,
                    "word": " a tree",
                    "probability": 0.798828125
                },
                {
                    "start": 922.72,
                    "end": 923.38,
                    "word": ",",
                    "probability": 0.2158203125
                },
                {
                    "start": 923.38,
                    "end": 923.56,
                    "word": " at",
                    "probability": 0.771484375
                },
                {
                    "start": 923.56,
                    "end": 923.78,
                    "word": " every",
                    "probability": 0.73828125
                },
                {
                    "start": 923.78,
                    "end": 924.04,
                    "word": " single",
                    "probability": 0.8408203125
                },
                {
                    "start": 924.04,
                    "end": 924.36,
                    "word": " node",
                    "probability": 0.79248046875
                },
                {
                    "start": 924.36,
                    "end": 924.48,
                    "word": ",",
                    "probability": 0.53662109375
                },
                {
                    "start": 924.48,
                    "end": 924.6,
                    "word": " I",
                    "probability": 0.990234375
                },
                {
                    "start": 924.6,
                    "end": 924.72,
                    "word": " can",
                    "probability": 0.89306640625
                },
                {
                    "start": 924.72,
                    "end": 924.86,
                    "word": " either",
                    "probability": 0.83251953125
                },
                {
                    "start": 924.86,
                    "end": 925.02,
                    "word": " go",
                    "probability": 0.89208984375
                },
                {
                    "start": 925.02,
                    "end": 925.22,
                    "word": " left",
                    "probability": 0.84765625
                },
                {
                    "start": 925.22,
                    "end": 925.34,
                    "word": ",",
                    "probability": 0.2073974609375
                },
                {
                    "start": 925.34,
                    "end": 925.42,
                    "word": " or",
                    "probability": 0.71044921875
                },
                {
                    "start": 925.42,
                    "end": 925.64,
                    "word": " right",
                    "probability": 0.87939453125
                },
                {
                    "start": 925.64,
                    "end": 925.72,
                    "word": ",",
                    "probability": 0.919921875
                },
                {
                    "start": 925.72,
                    "end": 925.8,
                    "word": " or",
                    "probability": 0.8916015625
                },
                {
                    "start": 925.8,
                    "end": 926.1,
                    "word": " down",
                    "probability": 0.80859375
                },
                {
                    "start": 926.1,
                    "end": 926.2,
                    "word": ",",
                    "probability": 0.6396484375
                },
                {
                    "start": 926.2,
                    "end": 926.26,
                    "word": " or",
                    "probability": 0.79833984375
                },
                {
                    "start": 926.26,
                    "end": 926.34,
                    "word": " you",
                    "probability": 0.335693359375
                },
                {
                    "start": 926.34,
                    "end": 926.48,
                    "word": " know",
                    "probability": 0.82421875
                },
                {
                    "start": 926.48,
                    "end": 926.54,
                    "word": ",",
                    "probability": 0.72412109375
                }
            ]
        },
        {
            "id": 137,
            "seek": 80724,
            "start": 926.54,
            "end": 933.66,
            "text": " in this case, my number of children I have is variable, but I will be at a parent node,",
            "tokens": [
                50592,
                294,
                341,
                1389,
                11,
                452,
                1230,
                295,
                2227,
                286,
                362,
                307,
                7006,
                11,
                457,
                286,
                486,
                312,
                412,
                257,
                2596,
                9984,
                11,
                50944
            ],
            "temperature": 0.0,
            "avg_logprob": -0.21116504333551647,
            "compression_ratio": 1.6712962962962963,
            "no_speech_prob": 0.0014781951904296875,
            "words": [
                {
                    "start": 926.54,
                    "end": 926.76,
                    "word": " in",
                    "probability": 0.892578125
                },
                {
                    "start": 926.76,
                    "end": 926.9,
                    "word": " this",
                    "probability": 0.87841796875
                },
                {
                    "start": 926.9,
                    "end": 927.08,
                    "word": " case",
                    "probability": 0.81201171875
                },
                {
                    "start": 927.08,
                    "end": 927.18,
                    "word": ",",
                    "probability": 0.6005859375
                },
                {
                    "start": 927.18,
                    "end": 927.28,
                    "word": " my",
                    "probability": 0.86083984375
                },
                {
                    "start": 927.28,
                    "end": 927.56,
                    "word": " number",
                    "probability": 0.87060546875
                },
                {
                    "start": 927.56,
                    "end": 927.7,
                    "word": " of",
                    "probability": 0.90576171875
                },
                {
                    "start": 927.7,
                    "end": 927.92,
                    "word": " children",
                    "probability": 0.69384765625
                },
                {
                    "start": 927.92,
                    "end": 928.1,
                    "word": " I",
                    "probability": 0.9501953125
                },
                {
                    "start": 928.1,
                    "end": 928.28,
                    "word": " have",
                    "probability": 0.87744140625
                },
                {
                    "start": 928.28,
                    "end": 928.48,
                    "word": " is",
                    "probability": 0.88720703125
                },
                {
                    "start": 928.48,
                    "end": 930.3,
                    "word": " variable",
                    "probability": 0.53125
                },
                {
                    "start": 930.3,
                    "end": 931.46,
                    "word": ",",
                    "probability": 0.4130859375
                },
                {
                    "start": 931.46,
                    "end": 931.54,
                    "word": " but",
                    "probability": 0.82421875
                },
                {
                    "start": 931.54,
                    "end": 931.66,
                    "word": " I",
                    "probability": 0.91650390625
                },
                {
                    "start": 931.66,
                    "end": 932.32,
                    "word": " will",
                    "probability": 0.471923828125
                },
                {
                    "start": 932.32,
                    "end": 932.62,
                    "word": " be",
                    "probability": 0.80615234375
                },
                {
                    "start": 932.62,
                    "end": 932.78,
                    "word": " at",
                    "probability": 0.454345703125
                },
                {
                    "start": 933.14,
                    "end": 933.34,
                    "word": " a parent",
                    "probability": 0.76220703125
                },
                {
                    "start": 933.34,
                    "end": 933.56,
                    "word": " node",
                    "probability": 0.77197265625
                },
                {
                    "start": 933.56,
                    "end": 933.66,
                    "word": ",",
                    "probability": 0.5361328125
                }
            ]
        },
        {
            "id": 138,
            "seek": 80724,
            "start": 933.66,
            "end": 940.7,
            "text": " and I have to go to one of its children. So I always have discrete choices, a discrete choice",
            "tokens": [
                50944,
                293,
                286,
                362,
                281,
                352,
                281,
                472,
                295,
                1080,
                2227,
                13,
                407,
                286,
                1009,
                362,
                27706,
                7994,
                11,
                257,
                27706,
                3922,
                51296
            ],
            "temperature": 0.0,
            "avg_logprob": -0.21116504333551647,
            "compression_ratio": 1.6712962962962963,
            "no_speech_prob": 0.0014781951904296875,
            "words": [
                {
                    "start": 933.66,
                    "end": 933.78,
                    "word": " and",
                    "probability": 0.888671875
                },
                {
                    "start": 933.78,
                    "end": 933.88,
                    "word": " I",
                    "probability": 0.99755859375
                },
                {
                    "start": 933.88,
                    "end": 934.0,
                    "word": " have",
                    "probability": 0.8720703125
                },
                {
                    "start": 934.0,
                    "end": 934.1,
                    "word": " to",
                    "probability": 0.90380859375
                },
                {
                    "start": 934.1,
                    "end": 934.22,
                    "word": " go",
                    "probability": 0.89990234375
                },
                {
                    "start": 934.22,
                    "end": 934.38,
                    "word": " to",
                    "probability": 0.8974609375
                },
                {
                    "start": 934.38,
                    "end": 934.52,
                    "word": " one",
                    "probability": 0.81494140625
                },
                {
                    "start": 934.52,
                    "end": 934.62,
                    "word": " of",
                    "probability": 0.90625
                },
                {
                    "start": 934.62,
                    "end": 934.74,
                    "word": " its",
                    "probability": 0.775390625
                },
                {
                    "start": 934.74,
                    "end": 935.08,
                    "word": " children",
                    "probability": 0.6943359375
                },
                {
                    "start": 935.08,
                    "end": 935.32,
                    "word": ".",
                    "probability": 0.7919921875
                },
                {
                    "start": 935.32,
                    "end": 935.46,
                    "word": " So",
                    "probability": 0.88720703125
                },
                {
                    "start": 935.46,
                    "end": 935.62,
                    "word": " I",
                    "probability": 0.849609375
                },
                {
                    "start": 935.62,
                    "end": 935.88,
                    "word": " always",
                    "probability": 0.748046875
                },
                {
                    "start": 935.88,
                    "end": 937.44,
                    "word": " have",
                    "probability": 0.8720703125
                },
                {
                    "start": 937.44,
                    "end": 939.06,
                    "word": " discrete",
                    "probability": 0.4111328125
                },
                {
                    "start": 939.06,
                    "end": 939.5,
                    "word": " choices",
                    "probability": 0.83935546875
                },
                {
                    "start": 939.5,
                    "end": 939.96,
                    "word": ",",
                    "probability": 0.82958984375
                },
                {
                    "start": 940.14,
                    "end": 940.4,
                    "word": " a discrete",
                    "probability": 0.75732421875
                },
                {
                    "start": 940.4,
                    "end": 940.7,
                    "word": " choice",
                    "probability": 0.7998046875
                }
            ]
        },
        {
            "id": 139,
            "seek": 80724,
            "start": 940.7,
            "end": 947.46,
            "text": " of where to go. So how does that apply? So here is a very, like, simple cover tree,",
            "tokens": [
                51296,
                295,
                689,
                281,
                352,
                13,
                407,
                577,
                775,
                300,
                3079,
                30,
                407,
                510,
                307,
                257,
                588,
                11,
                411,
                11,
                2199,
                2060,
                4230,
                11,
                51632
            ],
            "temperature": 0.0,
            "avg_logprob": -0.21116504333551647,
            "compression_ratio": 1.6712962962962963,
            "no_speech_prob": 0.0014781951904296875,
            "words": [
                {
                    "start": 940.7,
                    "end": 940.8,
                    "word": " of",
                    "probability": 0.89599609375
                },
                {
                    "start": 940.8,
                    "end": 940.92,
                    "word": " where",
                    "probability": 0.89599609375
                },
                {
                    "start": 940.92,
                    "end": 941.02,
                    "word": " to",
                    "probability": 0.90966796875
                },
                {
                    "start": 941.02,
                    "end": 941.26,
                    "word": " go",
                    "probability": 0.904296875
                },
                {
                    "start": 941.26,
                    "end": 941.62,
                    "word": ".",
                    "probability": 0.98046875
                },
                {
                    "start": 942.32,
                    "end": 942.68,
                    "word": " So",
                    "probability": 0.8974609375
                },
                {
                    "start": 942.68,
                    "end": 942.86,
                    "word": " how",
                    "probability": 0.8564453125
                },
                {
                    "start": 942.86,
                    "end": 943.02,
                    "word": " does",
                    "probability": 0.919921875
                },
                {
                    "start": 943.02,
                    "end": 943.16,
                    "word": " that",
                    "probability": 0.8818359375
                },
                {
                    "start": 943.16,
                    "end": 943.44,
                    "word": " apply",
                    "probability": 0.8671875
                },
                {
                    "start": 943.44,
                    "end": 943.8,
                    "word": "?",
                    "probability": 0.990234375
                },
                {
                    "start": 944.32,
                    "end": 944.6,
                    "word": " So",
                    "probability": 0.87890625
                },
                {
                    "start": 944.6,
                    "end": 945.0,
                    "word": " here",
                    "probability": 0.796875
                },
                {
                    "start": 945.0,
                    "end": 945.4,
                    "word": " is",
                    "probability": 0.90283203125
                },
                {
                    "start": 945.62,
                    "end": 945.9,
                    "word": " a very",
                    "probability": 0.7216796875
                },
                {
                    "start": 945.9,
                    "end": 946.06,
                    "word": ",",
                    "probability": 0.16650390625
                },
                {
                    "start": 946.06,
                    "end": 946.26,
                    "word": " like",
                    "probability": 0.853515625
                },
                {
                    "start": 946.26,
                    "end": 946.38,
                    "word": ",",
                    "probability": 0.64599609375
                },
                {
                    "start": 946.38,
                    "end": 946.7,
                    "word": " simple",
                    "probability": 0.859375
                },
                {
                    "start": 946.7,
                    "end": 947.08,
                    "word": " cover",
                    "probability": 0.703125
                },
                {
                    "start": 947.08,
                    "end": 947.28,
                    "word": " tree",
                    "probability": 0.80615234375
                },
                {
                    "start": 947.28,
                    "end": 947.46,
                    "word": ",",
                    "probability": 0.5654296875
                }
            ]
        },
        {
            "id": 140,
            "seek": 83074,
            "start": 947.46,
            "end": 954.84,
            "text": " and I've colored it by the total, like, the population of this node.",
            "tokens": [
                50388,
                293,
                286,
                600,
                14332,
                309,
                538,
                264,
                3217,
                11,
                411,
                11,
                264,
                4415,
                295,
                341,
                9984,
                13,
                50700
            ],
            "temperature": 0.0,
            "avg_logprob": -0.24266826923076923,
            "compression_ratio": 1.8222222222222222,
            "no_speech_prob": 0.0009112358093261719,
            "words": [
                {
                    "start": 947.46,
                    "end": 948.44,
                    "word": " and",
                    "probability": 0.395263671875
                },
                {
                    "start": 948.44,
                    "end": 949.08,
                    "word": " I've",
                    "probability": 0.916748046875
                },
                {
                    "start": 949.08,
                    "end": 949.3,
                    "word": " colored",
                    "probability": 0.7431640625
                },
                {
                    "start": 949.3,
                    "end": 949.62,
                    "word": " it",
                    "probability": 0.90087890625
                },
                {
                    "start": 949.62,
                    "end": 950.2,
                    "word": " by",
                    "probability": 0.8828125
                },
                {
                    "start": 950.2,
                    "end": 951.16,
                    "word": " the",
                    "probability": 0.79736328125
                },
                {
                    "start": 951.16,
                    "end": 951.86,
                    "word": " total",
                    "probability": 0.48095703125
                },
                {
                    "start": 951.86,
                    "end": 952.16,
                    "word": ",",
                    "probability": 0.424560546875
                },
                {
                    "start": 952.16,
                    "end": 952.32,
                    "word": " like",
                    "probability": 0.2314453125
                },
                {
                    "start": 952.32,
                    "end": 952.46,
                    "word": ",",
                    "probability": 0.1455078125
                },
                {
                    "start": 952.46,
                    "end": 952.58,
                    "word": " the",
                    "probability": 0.765625
                },
                {
                    "start": 952.58,
                    "end": 953.22,
                    "word": " population",
                    "probability": 0.85498046875
                },
                {
                    "start": 953.22,
                    "end": 953.68,
                    "word": " of",
                    "probability": 0.90380859375
                },
                {
                    "start": 953.68,
                    "end": 953.9,
                    "word": " this",
                    "probability": 0.884765625
                },
                {
                    "start": 953.9,
                    "end": 954.26,
                    "word": " node",
                    "probability": 0.791015625
                },
                {
                    "start": 954.26,
                    "end": 954.84,
                    "word": ".",
                    "probability": 0.59912109375
                }
            ]
        },
        {
            "id": 141,
            "seek": 83074,
            "start": 955.9,
            "end": 962.52,
            "text": " So this color here is the relative population of this node. So about 66% of the data",
            "tokens": [
                50780,
                407,
                341,
                2017,
                510,
                307,
                264,
                4972,
                4415,
                295,
                341,
                9984,
                13,
                407,
                466,
                21126,
                4,
                295,
                264,
                1412,
                51120
            ],
            "temperature": 0.0,
            "avg_logprob": -0.24266826923076923,
            "compression_ratio": 1.8222222222222222,
            "no_speech_prob": 0.0009112358093261719,
            "words": [
                {
                    "start": 955.9,
                    "end": 956.1,
                    "word": " So",
                    "probability": 0.73193359375
                },
                {
                    "start": 956.1,
                    "end": 956.66,
                    "word": " this",
                    "probability": 0.61376953125
                },
                {
                    "start": 956.66,
                    "end": 956.9,
                    "word": " color",
                    "probability": 0.8359375
                },
                {
                    "start": 956.9,
                    "end": 957.16,
                    "word": " here",
                    "probability": 0.79931640625
                },
                {
                    "start": 957.16,
                    "end": 957.32,
                    "word": " is",
                    "probability": 0.8896484375
                },
                {
                    "start": 957.32,
                    "end": 957.44,
                    "word": " the",
                    "probability": 0.81103515625
                },
                {
                    "start": 957.44,
                    "end": 958.04,
                    "word": " relative",
                    "probability": 0.70654296875
                },
                {
                    "start": 958.04,
                    "end": 958.54,
                    "word": " population",
                    "probability": 0.86181640625
                },
                {
                    "start": 958.54,
                    "end": 958.82,
                    "word": " of",
                    "probability": 0.90625
                },
                {
                    "start": 958.82,
                    "end": 958.98,
                    "word": " this",
                    "probability": 0.8828125
                },
                {
                    "start": 958.98,
                    "end": 959.22,
                    "word": " node",
                    "probability": 0.79248046875
                },
                {
                    "start": 959.22,
                    "end": 959.8,
                    "word": ".",
                    "probability": 0.91748046875
                },
                {
                    "start": 959.86,
                    "end": 960.0,
                    "word": " So",
                    "probability": 0.880859375
                },
                {
                    "start": 960.0,
                    "end": 960.86,
                    "word": " about",
                    "probability": 0.73828125
                },
                {
                    "start": 960.86,
                    "end": 961.48,
                    "word": " 66",
                    "probability": 0.86669921875
                },
                {
                    "start": 961.48,
                    "end": 961.72,
                    "word": "%",
                    "probability": 0.94287109375
                },
                {
                    "start": 961.72,
                    "end": 961.9,
                    "word": " of",
                    "probability": 0.90966796875
                },
                {
                    "start": 961.9,
                    "end": 962.0,
                    "word": " the",
                    "probability": 0.81640625
                },
                {
                    "start": 962.0,
                    "end": 962.52,
                    "word": " data",
                    "probability": 0.81396484375
                }
            ]
        },
        {
            "id": 142,
            "seek": 83074,
            "start": 963.4,
            "end": 970.96,
            "text": " is underneath this node. About 33% of the data is underneath this node. And you can see that",
            "tokens": [
                51168,
                307,
                7223,
                341,
                9984,
                13,
                7769,
                11816,
                4,
                295,
                264,
                1412,
                307,
                7223,
                341,
                9984,
                13,
                400,
                291,
                393,
                536,
                300,
                51532
            ],
            "temperature": 0.0,
            "avg_logprob": -0.24266826923076923,
            "compression_ratio": 1.8222222222222222,
            "no_speech_prob": 0.0009112358093261719,
            "words": [
                {
                    "start": 963.4,
                    "end": 963.98,
                    "word": " is",
                    "probability": 0.88720703125
                },
                {
                    "start": 963.98,
                    "end": 964.48,
                    "word": " underneath",
                    "probability": 0.85595703125
                },
                {
                    "start": 964.48,
                    "end": 965.1,
                    "word": " this",
                    "probability": 0.8828125
                },
                {
                    "start": 965.1,
                    "end": 965.42,
                    "word": " node",
                    "probability": 0.79833984375
                },
                {
                    "start": 965.42,
                    "end": 966.0,
                    "word": ".",
                    "probability": 0.69921875
                },
                {
                    "start": 966.5,
                    "end": 967.04,
                    "word": " About",
                    "probability": 0.830078125
                },
                {
                    "start": 967.04,
                    "end": 968.06,
                    "word": " 33",
                    "probability": 0.85546875
                },
                {
                    "start": 968.06,
                    "end": 968.28,
                    "word": "%",
                    "probability": 0.99267578125
                },
                {
                    "start": 968.28,
                    "end": 968.4,
                    "word": " of",
                    "probability": 0.83935546875
                },
                {
                    "start": 968.4,
                    "end": 968.46,
                    "word": " the",
                    "probability": 0.7451171875
                },
                {
                    "start": 968.46,
                    "end": 968.6,
                    "word": " data",
                    "probability": 0.822265625
                },
                {
                    "start": 968.6,
                    "end": 968.82,
                    "word": " is",
                    "probability": 0.91259765625
                },
                {
                    "start": 968.82,
                    "end": 969.12,
                    "word": " underneath",
                    "probability": 0.85107421875
                },
                {
                    "start": 969.12,
                    "end": 969.44,
                    "word": " this",
                    "probability": 0.8837890625
                },
                {
                    "start": 969.44,
                    "end": 969.68,
                    "word": " node",
                    "probability": 0.81494140625
                },
                {
                    "start": 969.68,
                    "end": 970.22,
                    "word": ".",
                    "probability": 0.79345703125
                },
                {
                    "start": 970.22,
                    "end": 970.44,
                    "word": " And",
                    "probability": 0.88232421875
                },
                {
                    "start": 970.44,
                    "end": 970.54,
                    "word": " you",
                    "probability": 0.89892578125
                },
                {
                    "start": 970.54,
                    "end": 970.66,
                    "word": " can",
                    "probability": 0.8984375
                },
                {
                    "start": 970.66,
                    "end": 970.82,
                    "word": " see",
                    "probability": 0.81103515625
                },
                {
                    "start": 970.82,
                    "end": 970.96,
                    "word": " that",
                    "probability": 0.8369140625
                }
            ]
        },
        {
            "id": 143,
            "seek": 85862,
            "start": 970.96,
            "end": 977.92,
            "text": " colors are kind of changing. I get a kind of a picture of how dense my data is, but that doesn't",
            "tokens": [
                50364,
                4577,
                366,
                733,
                295,
                4473,
                13,
                286,
                483,
                257,
                733,
                295,
                257,
                3036,
                295,
                577,
                18011,
                452,
                1412,
                307,
                11,
                457,
                300,
                1177,
                380,
                50708
            ],
            "temperature": 0.0,
            "avg_logprob": -0.23395647960049765,
            "compression_ratio": 1.8254716981132075,
            "no_speech_prob": 0.115966796875,
            "words": [
                {
                    "start": 970.96,
                    "end": 971.34,
                    "word": " colors",
                    "probability": 0.1011962890625
                },
                {
                    "start": 971.34,
                    "end": 971.54,
                    "word": " are",
                    "probability": 0.3046875
                },
                {
                    "start": 971.54,
                    "end": 971.7,
                    "word": " kind",
                    "probability": 0.60693359375
                },
                {
                    "start": 971.7,
                    "end": 971.78,
                    "word": " of",
                    "probability": 0.908203125
                },
                {
                    "start": 971.78,
                    "end": 972.2,
                    "word": " changing",
                    "probability": 0.464599609375
                },
                {
                    "start": 972.2,
                    "end": 972.62,
                    "word": ".",
                    "probability": 0.2117919921875
                },
                {
                    "start": 972.62,
                    "end": 972.7,
                    "word": " I",
                    "probability": 0.92431640625
                },
                {
                    "start": 972.7,
                    "end": 972.88,
                    "word": " get",
                    "probability": 0.87841796875
                },
                {
                    "start": 973.08,
                    "end": 973.82,
                    "word": " a kind",
                    "probability": 0.492919921875
                },
                {
                    "start": 973.82,
                    "end": 973.9,
                    "word": " of",
                    "probability": 0.9091796875
                },
                {
                    "start": 974.08,
                    "end": 974.32,
                    "word": " a picture",
                    "probability": 0.76806640625
                },
                {
                    "start": 974.32,
                    "end": 974.64,
                    "word": " of",
                    "probability": 0.90185546875
                },
                {
                    "start": 974.64,
                    "end": 975.08,
                    "word": " how",
                    "probability": 0.9091796875
                },
                {
                    "start": 975.08,
                    "end": 975.78,
                    "word": " dense",
                    "probability": 0.81787109375
                },
                {
                    "start": 975.78,
                    "end": 976.02,
                    "word": " my",
                    "probability": 0.90478515625
                },
                {
                    "start": 976.02,
                    "end": 976.28,
                    "word": " data",
                    "probability": 0.83203125
                },
                {
                    "start": 976.28,
                    "end": 976.62,
                    "word": " is",
                    "probability": 0.90380859375
                },
                {
                    "start": 976.62,
                    "end": 977.0,
                    "word": ",",
                    "probability": 0.271484375
                },
                {
                    "start": 977.0,
                    "end": 977.52,
                    "word": " but",
                    "probability": 0.8369140625
                },
                {
                    "start": 977.52,
                    "end": 977.62,
                    "word": " that",
                    "probability": 0.8447265625
                },
                {
                    "start": 977.62,
                    "end": 977.92,
                    "word": " doesn't",
                    "probability": 0.9375
                }
            ]
        },
        {
            "id": 144,
            "seek": 85862,
            "start": 977.92,
            "end": 982.68,
            "text": " help me, because I need to make choices. It doesn't help me make the choices at all steps.",
            "tokens": [
                50708,
                854,
                385,
                11,
                570,
                286,
                643,
                281,
                652,
                7994,
                13,
                467,
                1177,
                380,
                854,
                385,
                652,
                264,
                7994,
                412,
                439,
                4439,
                13,
                50952
            ],
            "temperature": 0.0,
            "avg_logprob": -0.23395647960049765,
            "compression_ratio": 1.8254716981132075,
            "no_speech_prob": 0.115966796875,
            "words": [
                {
                    "start": 977.92,
                    "end": 978.18,
                    "word": " help",
                    "probability": 0.80322265625
                },
                {
                    "start": 978.18,
                    "end": 978.32,
                    "word": " me",
                    "probability": 0.896484375
                },
                {
                    "start": 978.32,
                    "end": 978.38,
                    "word": ",",
                    "probability": 0.09332275390625
                },
                {
                    "start": 978.38,
                    "end": 978.5,
                    "word": " because",
                    "probability": 0.82763671875
                },
                {
                    "start": 978.5,
                    "end": 978.6,
                    "word": " I",
                    "probability": 0.62109375
                },
                {
                    "start": 978.6,
                    "end": 978.76,
                    "word": " need",
                    "probability": 0.82177734375
                },
                {
                    "start": 978.76,
                    "end": 978.88,
                    "word": " to",
                    "probability": 0.89306640625
                },
                {
                    "start": 978.88,
                    "end": 979.1,
                    "word": " make",
                    "probability": 0.8759765625
                },
                {
                    "start": 979.1,
                    "end": 979.72,
                    "word": " choices",
                    "probability": 0.48388671875
                },
                {
                    "start": 979.72,
                    "end": 979.82,
                    "word": ".",
                    "probability": 0.48095703125
                },
                {
                    "start": 979.82,
                    "end": 980.14,
                    "word": " It",
                    "probability": 0.85888671875
                },
                {
                    "start": 980.14,
                    "end": 980.34,
                    "word": " doesn't",
                    "probability": 0.93603515625
                },
                {
                    "start": 980.34,
                    "end": 980.48,
                    "word": " help",
                    "probability": 0.79931640625
                },
                {
                    "start": 980.48,
                    "end": 980.6,
                    "word": " me",
                    "probability": 0.89794921875
                },
                {
                    "start": 980.6,
                    "end": 980.9,
                    "word": " make",
                    "probability": 0.78125
                },
                {
                    "start": 980.9,
                    "end": 981.08,
                    "word": " the",
                    "probability": 0.7802734375
                },
                {
                    "start": 981.08,
                    "end": 981.42,
                    "word": " choices",
                    "probability": 0.83544921875
                },
                {
                    "start": 981.42,
                    "end": 981.62,
                    "word": " at",
                    "probability": 0.849609375
                },
                {
                    "start": 981.62,
                    "end": 981.78,
                    "word": " all",
                    "probability": 0.8876953125
                },
                {
                    "start": 981.78,
                    "end": 982.24,
                    "word": " steps",
                    "probability": 0.362548828125
                },
                {
                    "start": 982.24,
                    "end": 982.68,
                    "word": ".",
                    "probability": 0.76611328125
                }
            ]
        },
        {
            "id": 145,
            "seek": 85862,
            "start": 982.8,
            "end": 990.82,
            "text": " So if I go here, this helps me make my choices. So here, I've got 66% of my data over here, and I've",
            "tokens": [
                50952,
                407,
                498,
                286,
                352,
                510,
                11,
                341,
                3665,
                385,
                652,
                452,
                7994,
                13,
                407,
                510,
                11,
                286,
                600,
                658,
                21126,
                4,
                295,
                452,
                1412,
                670,
                510,
                11,
                293,
                286,
                600,
                51352
            ],
            "temperature": 0.0,
            "avg_logprob": -0.23395647960049765,
            "compression_ratio": 1.8254716981132075,
            "no_speech_prob": 0.115966796875,
            "words": [
                {
                    "start": 982.8,
                    "end": 983.04,
                    "word": " So",
                    "probability": 0.84619140625
                },
                {
                    "start": 983.04,
                    "end": 983.44,
                    "word": " if",
                    "probability": 0.67822265625
                },
                {
                    "start": 983.44,
                    "end": 983.58,
                    "word": " I",
                    "probability": 0.95166015625
                },
                {
                    "start": 983.58,
                    "end": 983.78,
                    "word": " go",
                    "probability": 0.529296875
                },
                {
                    "start": 983.78,
                    "end": 984.32,
                    "word": " here",
                    "probability": 0.81103515625
                },
                {
                    "start": 984.32,
                    "end": 984.4,
                    "word": ",",
                    "probability": 0.8037109375
                },
                {
                    "start": 984.4,
                    "end": 984.94,
                    "word": " this",
                    "probability": 0.87646484375
                },
                {
                    "start": 984.94,
                    "end": 985.22,
                    "word": " helps",
                    "probability": 0.73486328125
                },
                {
                    "start": 985.22,
                    "end": 985.38,
                    "word": " me",
                    "probability": 0.900390625
                },
                {
                    "start": 985.38,
                    "end": 985.56,
                    "word": " make",
                    "probability": 0.86865234375
                },
                {
                    "start": 985.56,
                    "end": 985.72,
                    "word": " my",
                    "probability": 0.90283203125
                },
                {
                    "start": 985.72,
                    "end": 986.06,
                    "word": " choices",
                    "probability": 0.8486328125
                },
                {
                    "start": 986.06,
                    "end": 986.5,
                    "word": ".",
                    "probability": 0.943359375
                },
                {
                    "start": 987.04,
                    "end": 987.26,
                    "word": " So",
                    "probability": 0.81201171875
                },
                {
                    "start": 987.26,
                    "end": 987.56,
                    "word": " here",
                    "probability": 0.783203125
                },
                {
                    "start": 987.56,
                    "end": 987.66,
                    "word": ",",
                    "probability": 0.258056640625
                },
                {
                    "start": 987.66,
                    "end": 987.78,
                    "word": " I've",
                    "probability": 0.960205078125
                },
                {
                    "start": 987.78,
                    "end": 987.86,
                    "word": " got",
                    "probability": 0.8896484375
                },
                {
                    "start": 987.86,
                    "end": 988.24,
                    "word": " 66",
                    "probability": 0.83349609375
                },
                {
                    "start": 988.24,
                    "end": 988.96,
                    "word": "%",
                    "probability": 0.50439453125
                },
                {
                    "start": 988.96,
                    "end": 989.08,
                    "word": " of",
                    "probability": 0.89111328125
                },
                {
                    "start": 989.08,
                    "end": 989.2,
                    "word": " my",
                    "probability": 0.9013671875
                },
                {
                    "start": 989.2,
                    "end": 989.42,
                    "word": " data",
                    "probability": 0.8173828125
                },
                {
                    "start": 989.42,
                    "end": 989.7,
                    "word": " over",
                    "probability": 0.80810546875
                },
                {
                    "start": 989.7,
                    "end": 989.96,
                    "word": " here",
                    "probability": 0.8115234375
                },
                {
                    "start": 989.96,
                    "end": 990.36,
                    "word": ",",
                    "probability": 0.63623046875
                },
                {
                    "start": 990.36,
                    "end": 990.62,
                    "word": " and",
                    "probability": 0.89501953125
                },
                {
                    "start": 990.62,
                    "end": 990.82,
                    "word": " I've",
                    "probability": 0.966552734375
                }
            ]
        },
        {
            "id": 146,
            "seek": 85862,
            "start": 990.82,
            "end": 998.84,
            "text": " got 33% of my data over here. And then once I've taken... I have to take one of those two choices.",
            "tokens": [
                51352,
                658,
                11816,
                4,
                295,
                452,
                1412,
                670,
                510,
                13,
                400,
                550,
                1564,
                286,
                600,
                2726,
                485,
                286,
                362,
                281,
                747,
                472,
                295,
                729,
                732,
                7994,
                13,
                51756
            ],
            "temperature": 0.0,
            "avg_logprob": -0.23395647960049765,
            "compression_ratio": 1.8254716981132075,
            "no_speech_prob": 0.115966796875,
            "words": [
                {
                    "start": 990.82,
                    "end": 991.24,
                    "word": " got",
                    "probability": 0.88525390625
                },
                {
                    "start": 991.24,
                    "end": 993.86,
                    "word": " 33",
                    "probability": 0.74462890625
                },
                {
                    "start": 993.86,
                    "end": 994.22,
                    "word": "%",
                    "probability": 0.97412109375
                },
                {
                    "start": 994.22,
                    "end": 994.36,
                    "word": " of",
                    "probability": 0.90673828125
                },
                {
                    "start": 994.36,
                    "end": 994.48,
                    "word": " my",
                    "probability": 0.9013671875
                },
                {
                    "start": 994.48,
                    "end": 994.72,
                    "word": " data",
                    "probability": 0.8134765625
                },
                {
                    "start": 994.72,
                    "end": 995.18,
                    "word": " over",
                    "probability": 0.89306640625
                },
                {
                    "start": 995.18,
                    "end": 995.48,
                    "word": " here",
                    "probability": 0.81494140625
                },
                {
                    "start": 995.48,
                    "end": 995.84,
                    "word": ".",
                    "probability": 0.57275390625
                },
                {
                    "start": 995.84,
                    "end": 996.06,
                    "word": " And",
                    "probability": 0.73876953125
                },
                {
                    "start": 996.06,
                    "end": 996.2,
                    "word": " then",
                    "probability": 0.70263671875
                },
                {
                    "start": 996.2,
                    "end": 996.4,
                    "word": " once",
                    "probability": 0.529296875
                },
                {
                    "start": 996.4,
                    "end": 996.58,
                    "word": " I've",
                    "probability": 0.783203125
                },
                {
                    "start": 996.58,
                    "end": 997.06,
                    "word": " taken...",
                    "probability": 0.236572265625
                },
                {
                    "start": 997.06,
                    "end": 997.36,
                    "word": " I",
                    "probability": 0.654296875
                },
                {
                    "start": 997.36,
                    "end": 997.56,
                    "word": " have",
                    "probability": 0.87890625
                },
                {
                    "start": 997.56,
                    "end": 997.66,
                    "word": " to",
                    "probability": 0.9033203125
                },
                {
                    "start": 997.66,
                    "end": 997.84,
                    "word": " take",
                    "probability": 0.79931640625
                },
                {
                    "start": 997.84,
                    "end": 998.06,
                    "word": " one",
                    "probability": 0.82080078125
                },
                {
                    "start": 998.06,
                    "end": 998.16,
                    "word": " of",
                    "probability": 0.90673828125
                },
                {
                    "start": 998.16,
                    "end": 998.28,
                    "word": " those",
                    "probability": 0.84130859375
                },
                {
                    "start": 998.28,
                    "end": 998.42,
                    "word": " two",
                    "probability": 0.810546875
                },
                {
                    "start": 998.42,
                    "end": 998.72,
                    "word": " choices",
                    "probability": 0.85205078125
                },
                {
                    "start": 998.72,
                    "end": 998.84,
                    "word": ".",
                    "probability": 0.9208984375
                }
            ]
        },
        {
            "id": 147,
            "seek": 88318,
            "start": 998.84,
            "end": 1005.04,
            "text": " If I take this choice, I'm now at this node, and I have two choices again. And the relative",
            "tokens": [
                50364,
                759,
                286,
                747,
                341,
                3922,
                11,
                286,
                478,
                586,
                412,
                341,
                9984,
                11,
                293,
                286,
                362,
                732,
                7994,
                797,
                13,
                400,
                264,
                4972,
                50676
            ],
            "temperature": 0.0,
            "avg_logprob": -0.1839192770421505,
            "compression_ratio": 1.9175257731958764,
            "no_speech_prob": 0.003429412841796875,
            "words": [
                {
                    "start": 998.84,
                    "end": 998.98,
                    "word": " If",
                    "probability": 0.313232421875
                },
                {
                    "start": 998.98,
                    "end": 999.1,
                    "word": " I",
                    "probability": 0.9755859375
                },
                {
                    "start": 999.1,
                    "end": 999.32,
                    "word": " take",
                    "probability": 0.80126953125
                },
                {
                    "start": 999.32,
                    "end": 999.6,
                    "word": " this",
                    "probability": 0.88134765625
                },
                {
                    "start": 999.6,
                    "end": 1000.0,
                    "word": " choice",
                    "probability": 0.787109375
                },
                {
                    "start": 1000.0,
                    "end": 1000.46,
                    "word": ",",
                    "probability": 0.67578125
                },
                {
                    "start": 1000.46,
                    "end": 1000.88,
                    "word": " I'm",
                    "probability": 0.939208984375
                },
                {
                    "start": 1000.88,
                    "end": 1001.18,
                    "word": " now",
                    "probability": 0.890625
                },
                {
                    "start": 1001.18,
                    "end": 1001.54,
                    "word": " at",
                    "probability": 0.9052734375
                },
                {
                    "start": 1001.54,
                    "end": 1001.74,
                    "word": " this",
                    "probability": 0.88134765625
                },
                {
                    "start": 1001.74,
                    "end": 1002.0,
                    "word": " node",
                    "probability": 0.8037109375
                },
                {
                    "start": 1002.0,
                    "end": 1002.4,
                    "word": ",",
                    "probability": 0.36083984375
                },
                {
                    "start": 1002.4,
                    "end": 1003.02,
                    "word": " and",
                    "probability": 0.87939453125
                },
                {
                    "start": 1003.02,
                    "end": 1003.18,
                    "word": " I",
                    "probability": 0.99072265625
                },
                {
                    "start": 1003.18,
                    "end": 1003.32,
                    "word": " have",
                    "probability": 0.671875
                },
                {
                    "start": 1003.32,
                    "end": 1003.5,
                    "word": " two",
                    "probability": 0.7783203125
                },
                {
                    "start": 1003.5,
                    "end": 1003.8,
                    "word": " choices",
                    "probability": 0.8583984375
                },
                {
                    "start": 1003.8,
                    "end": 1004.14,
                    "word": " again",
                    "probability": 0.8818359375
                },
                {
                    "start": 1004.14,
                    "end": 1004.38,
                    "word": ".",
                    "probability": 0.501953125
                },
                {
                    "start": 1004.38,
                    "end": 1004.5,
                    "word": " And",
                    "probability": 0.77197265625
                },
                {
                    "start": 1004.5,
                    "end": 1004.68,
                    "word": " the",
                    "probability": 0.81005859375
                },
                {
                    "start": 1004.68,
                    "end": 1005.04,
                    "word": " relative",
                    "probability": 0.86865234375
                }
            ]
        },
        {
            "id": 148,
            "seek": 88318,
            "start": 1005.04,
            "end": 1011.34,
            "text": " probability of taking this choice versus this one... Well, this is basically a 0.1 probability",
            "tokens": [
                50676,
                8482,
                295,
                1940,
                341,
                3922,
                5717,
                341,
                472,
                485,
                1042,
                11,
                341,
                307,
                1936,
                257,
                1958,
                13,
                16,
                8482,
                50988
            ],
            "temperature": 0.0,
            "avg_logprob": -0.1839192770421505,
            "compression_ratio": 1.9175257731958764,
            "no_speech_prob": 0.003429412841796875,
            "words": [
                {
                    "start": 1005.04,
                    "end": 1005.72,
                    "word": " probability",
                    "probability": 0.896484375
                },
                {
                    "start": 1005.72,
                    "end": 1005.94,
                    "word": " of",
                    "probability": 0.89794921875
                },
                {
                    "start": 1005.94,
                    "end": 1006.12,
                    "word": " taking",
                    "probability": 0.77001953125
                },
                {
                    "start": 1006.12,
                    "end": 1006.34,
                    "word": " this",
                    "probability": 0.88037109375
                },
                {
                    "start": 1006.34,
                    "end": 1006.72,
                    "word": " choice",
                    "probability": 0.82568359375
                },
                {
                    "start": 1006.72,
                    "end": 1007.12,
                    "word": " versus",
                    "probability": 0.70166015625
                },
                {
                    "start": 1007.12,
                    "end": 1007.4,
                    "word": " this",
                    "probability": 0.88232421875
                },
                {
                    "start": 1007.4,
                    "end": 1008.2,
                    "word": " one...",
                    "probability": 0.4383697509765625
                },
                {
                    "start": 1008.2,
                    "end": 1008.26,
                    "word": " Well",
                    "probability": 0.247314453125
                },
                {
                    "start": 1008.26,
                    "end": 1008.3,
                    "word": ",",
                    "probability": 0.697265625
                },
                {
                    "start": 1008.3,
                    "end": 1008.46,
                    "word": " this",
                    "probability": 0.8759765625
                },
                {
                    "start": 1008.46,
                    "end": 1008.78,
                    "word": " is",
                    "probability": 0.8935546875
                },
                {
                    "start": 1008.78,
                    "end": 1009.34,
                    "word": " basically",
                    "probability": 0.88818359375
                },
                {
                    "start": 1009.62,
                    "end": 1010.3,
                    "word": " a 0",
                    "probability": 0.75244140625
                },
                {
                    "start": 1010.3,
                    "end": 1010.68,
                    "word": ".1",
                    "probability": 0.994384765625
                },
                {
                    "start": 1010.68,
                    "end": 1011.34,
                    "word": " probability",
                    "probability": 0.92138671875
                }
            ]
        },
        {
            "id": 149,
            "seek": 88318,
            "start": 1011.34,
            "end": 1018.44,
            "text": " of taking this path, and this is a 0.9 probability of taking this path. And once I get to this node,",
            "tokens": [
                50988,
                295,
                1940,
                341,
                3100,
                11,
                293,
                341,
                307,
                257,
                1958,
                13,
                24,
                8482,
                295,
                1940,
                341,
                3100,
                13,
                400,
                1564,
                286,
                483,
                281,
                341,
                9984,
                11,
                51344
            ],
            "temperature": 0.0,
            "avg_logprob": -0.1839192770421505,
            "compression_ratio": 1.9175257731958764,
            "no_speech_prob": 0.003429412841796875,
            "words": [
                {
                    "start": 1011.34,
                    "end": 1011.54,
                    "word": " of",
                    "probability": 0.88818359375
                },
                {
                    "start": 1011.54,
                    "end": 1011.78,
                    "word": " taking",
                    "probability": 0.79541015625
                },
                {
                    "start": 1011.78,
                    "end": 1011.96,
                    "word": " this",
                    "probability": 0.8701171875
                },
                {
                    "start": 1011.96,
                    "end": 1012.96,
                    "word": " path",
                    "probability": 0.79736328125
                },
                {
                    "start": 1012.96,
                    "end": 1013.24,
                    "word": ",",
                    "probability": 0.6083984375
                },
                {
                    "start": 1013.24,
                    "end": 1013.68,
                    "word": " and",
                    "probability": 0.87353515625
                },
                {
                    "start": 1013.68,
                    "end": 1013.88,
                    "word": " this",
                    "probability": 0.88037109375
                },
                {
                    "start": 1013.88,
                    "end": 1014.08,
                    "word": " is",
                    "probability": 0.8994140625
                },
                {
                    "start": 1014.38,
                    "end": 1014.66,
                    "word": " a 0",
                    "probability": 0.98486328125
                },
                {
                    "start": 1014.66,
                    "end": 1015.04,
                    "word": ".9",
                    "probability": 0.997802734375
                },
                {
                    "start": 1015.04,
                    "end": 1015.54,
                    "word": " probability",
                    "probability": 0.9140625
                },
                {
                    "start": 1015.54,
                    "end": 1015.74,
                    "word": " of",
                    "probability": 0.90380859375
                },
                {
                    "start": 1015.74,
                    "end": 1015.98,
                    "word": " taking",
                    "probability": 0.80029296875
                },
                {
                    "start": 1015.98,
                    "end": 1016.24,
                    "word": " this",
                    "probability": 0.8818359375
                },
                {
                    "start": 1016.24,
                    "end": 1016.58,
                    "word": " path",
                    "probability": 0.85498046875
                },
                {
                    "start": 1016.58,
                    "end": 1017.1,
                    "word": ".",
                    "probability": 0.96435546875
                },
                {
                    "start": 1017.16,
                    "end": 1017.32,
                    "word": " And",
                    "probability": 0.7275390625
                },
                {
                    "start": 1017.32,
                    "end": 1017.46,
                    "word": " once",
                    "probability": 0.81494140625
                },
                {
                    "start": 1017.46,
                    "end": 1017.56,
                    "word": " I",
                    "probability": 0.98193359375
                },
                {
                    "start": 1017.56,
                    "end": 1017.66,
                    "word": " get",
                    "probability": 0.90234375
                },
                {
                    "start": 1017.66,
                    "end": 1017.78,
                    "word": " to",
                    "probability": 0.9033203125
                },
                {
                    "start": 1017.78,
                    "end": 1017.96,
                    "word": " this",
                    "probability": 0.88232421875
                },
                {
                    "start": 1017.96,
                    "end": 1018.18,
                    "word": " node",
                    "probability": 0.76513671875
                },
                {
                    "start": 1018.18,
                    "end": 1018.44,
                    "word": ",",
                    "probability": 0.9423828125
                }
            ]
        },
        {
            "id": 150,
            "seek": 88318,
            "start": 1018.44,
            "end": 1023.4,
            "text": " I have a probability of one of going straight down, because I have no other choices.",
            "tokens": [
                51344,
                286,
                362,
                257,
                8482,
                295,
                472,
                295,
                516,
                2997,
                760,
                11,
                570,
                286,
                362,
                572,
                661,
                7994,
                13,
                51592
            ],
            "temperature": 0.0,
            "avg_logprob": -0.1839192770421505,
            "compression_ratio": 1.9175257731958764,
            "no_speech_prob": 0.003429412841796875,
            "words": [
                {
                    "start": 1018.44,
                    "end": 1018.94,
                    "word": " I",
                    "probability": 0.99658203125
                },
                {
                    "start": 1018.94,
                    "end": 1019.2,
                    "word": " have",
                    "probability": 0.8798828125
                },
                {
                    "start": 1019.54,
                    "end": 1020.52,
                    "word": " a probability",
                    "probability": 0.88916015625
                },
                {
                    "start": 1020.52,
                    "end": 1020.82,
                    "word": " of",
                    "probability": 0.90234375
                },
                {
                    "start": 1020.82,
                    "end": 1021.12,
                    "word": " one",
                    "probability": 0.3671875
                },
                {
                    "start": 1021.12,
                    "end": 1021.26,
                    "word": " of",
                    "probability": 0.6240234375
                },
                {
                    "start": 1021.26,
                    "end": 1021.42,
                    "word": " going",
                    "probability": 0.8525390625
                },
                {
                    "start": 1021.42,
                    "end": 1021.64,
                    "word": " straight",
                    "probability": 0.82666015625
                },
                {
                    "start": 1021.64,
                    "end": 1021.84,
                    "word": " down",
                    "probability": 0.81298828125
                },
                {
                    "start": 1021.84,
                    "end": 1021.9,
                    "word": ",",
                    "probability": 0.29443359375
                },
                {
                    "start": 1021.9,
                    "end": 1022.04,
                    "word": " because",
                    "probability": 0.83544921875
                },
                {
                    "start": 1022.04,
                    "end": 1022.14,
                    "word": " I",
                    "probability": 0.9970703125
                },
                {
                    "start": 1022.14,
                    "end": 1022.26,
                    "word": " have",
                    "probability": 0.87890625
                },
                {
                    "start": 1022.26,
                    "end": 1022.42,
                    "word": " no",
                    "probability": 0.89599609375
                },
                {
                    "start": 1022.42,
                    "end": 1022.56,
                    "word": " other",
                    "probability": 0.8154296875
                },
                {
                    "start": 1022.56,
                    "end": 1022.88,
                    "word": " choices",
                    "probability": 0.8427734375
                },
                {
                    "start": 1022.88,
                    "end": 1023.4,
                    "word": ".",
                    "probability": 0.94482421875
                }
            ]
        },
        {
            "id": 151,
            "seek": 90704,
            "start": 1023.4,
            "end": 1029.2,
            "text": " And every single point, I have a discrete distribution, like we had on page one, which",
            "tokens": [
                50364,
                400,
                633,
                2167,
                935,
                11,
                286,
                362,
                257,
                27706,
                7316,
                11,
                411,
                321,
                632,
                322,
                3028,
                472,
                11,
                597,
                50668
            ],
            "temperature": 0.0,
            "avg_logprob": -0.20991847680314726,
            "compression_ratio": 1.6227272727272728,
            "no_speech_prob": 0.03515625,
            "words": [
                {
                    "start": 1023.4,
                    "end": 1023.68,
                    "word": " And",
                    "probability": 0.2200927734375
                },
                {
                    "start": 1023.68,
                    "end": 1023.98,
                    "word": " every",
                    "probability": 0.5341796875
                },
                {
                    "start": 1023.98,
                    "end": 1024.24,
                    "word": " single",
                    "probability": 0.87109375
                },
                {
                    "start": 1024.24,
                    "end": 1024.76,
                    "word": " point",
                    "probability": 0.8818359375
                },
                {
                    "start": 1024.76,
                    "end": 1024.88,
                    "word": ",",
                    "probability": 0.313232421875
                },
                {
                    "start": 1024.88,
                    "end": 1024.96,
                    "word": " I",
                    "probability": 0.98779296875
                },
                {
                    "start": 1024.96,
                    "end": 1025.14,
                    "word": " have",
                    "probability": 0.8720703125
                },
                {
                    "start": 1025.24,
                    "end": 1025.58,
                    "word": " a discrete",
                    "probability": 0.8076171875
                },
                {
                    "start": 1025.58,
                    "end": 1026.1,
                    "word": " distribution",
                    "probability": 0.7275390625
                },
                {
                    "start": 1026.1,
                    "end": 1026.58,
                    "word": ",",
                    "probability": 0.357421875
                },
                {
                    "start": 1026.58,
                    "end": 1026.74,
                    "word": " like",
                    "probability": 0.84423828125
                },
                {
                    "start": 1026.74,
                    "end": 1026.9,
                    "word": " we",
                    "probability": 0.8896484375
                },
                {
                    "start": 1026.9,
                    "end": 1027.08,
                    "word": " had",
                    "probability": 0.79345703125
                },
                {
                    "start": 1027.08,
                    "end": 1027.24,
                    "word": " on",
                    "probability": 0.90087890625
                },
                {
                    "start": 1027.24,
                    "end": 1028.06,
                    "word": " page",
                    "probability": 0.6474609375
                },
                {
                    "start": 1028.06,
                    "end": 1028.38,
                    "word": " one",
                    "probability": 0.57373046875
                },
                {
                    "start": 1028.38,
                    "end": 1028.6,
                    "word": ",",
                    "probability": 0.9130859375
                },
                {
                    "start": 1028.6,
                    "end": 1029.2,
                    "word": " which",
                    "probability": 0.91357421875
                }
            ]
        },
        {
            "id": 152,
            "seek": 90704,
            "start": 1029.2,
            "end": 1035.84,
            "text": " tells me where I'm going. And because my tree is geometrically motivated, these choices,",
            "tokens": [
                50668,
                5112,
                385,
                689,
                286,
                478,
                516,
                13,
                400,
                570,
                452,
                4230,
                307,
                12956,
                81,
                984,
                14515,
                11,
                613,
                7994,
                11,
                50976
            ],
            "temperature": 0.0,
            "avg_logprob": -0.20991847680314726,
            "compression_ratio": 1.6227272727272728,
            "no_speech_prob": 0.03515625,
            "words": [
                {
                    "start": 1029.2,
                    "end": 1029.9,
                    "word": " tells",
                    "probability": 0.75390625
                },
                {
                    "start": 1029.9,
                    "end": 1030.26,
                    "word": " me",
                    "probability": 0.9091796875
                },
                {
                    "start": 1030.26,
                    "end": 1030.9,
                    "word": " where",
                    "probability": 0.90771484375
                },
                {
                    "start": 1030.9,
                    "end": 1031.08,
                    "word": " I'm",
                    "probability": 0.98681640625
                },
                {
                    "start": 1031.08,
                    "end": 1031.38,
                    "word": " going",
                    "probability": 0.85302734375
                },
                {
                    "start": 1031.38,
                    "end": 1031.84,
                    "word": ".",
                    "probability": 0.9091796875
                },
                {
                    "start": 1031.84,
                    "end": 1031.94,
                    "word": " And",
                    "probability": 0.8623046875
                },
                {
                    "start": 1031.94,
                    "end": 1032.36,
                    "word": " because",
                    "probability": 0.8212890625
                },
                {
                    "start": 1032.36,
                    "end": 1032.62,
                    "word": " my",
                    "probability": 0.90771484375
                },
                {
                    "start": 1032.62,
                    "end": 1033.08,
                    "word": " tree",
                    "probability": 0.83935546875
                },
                {
                    "start": 1033.08,
                    "end": 1033.32,
                    "word": " is",
                    "probability": 0.9052734375
                },
                {
                    "start": 1033.32,
                    "end": 1034.12,
                    "word": " geometrically",
                    "probability": 0.89306640625
                },
                {
                    "start": 1034.12,
                    "end": 1034.66,
                    "word": " motivated",
                    "probability": 0.81396484375
                },
                {
                    "start": 1034.66,
                    "end": 1034.94,
                    "word": ",",
                    "probability": 0.9384765625
                },
                {
                    "start": 1034.94,
                    "end": 1035.16,
                    "word": " these",
                    "probability": 0.6748046875
                },
                {
                    "start": 1035.16,
                    "end": 1035.6,
                    "word": " choices",
                    "probability": 0.87060546875
                },
                {
                    "start": 1035.6,
                    "end": 1035.84,
                    "word": ",",
                    "probability": 0.74853515625
                }
            ]
        },
        {
            "id": 153,
            "seek": 90704,
            "start": 1035.84,
            "end": 1042.68,
            "text": " which were all discrete and easy to compute, and easy to count, and easy to keep track of,",
            "tokens": [
                50976,
                597,
                645,
                439,
                27706,
                293,
                1858,
                281,
                14722,
                11,
                293,
                1858,
                281,
                1207,
                11,
                293,
                1858,
                281,
                1066,
                2837,
                295,
                11,
                51212
            ],
            "temperature": 0.0,
            "avg_logprob": -0.20991847680314726,
            "compression_ratio": 1.6227272727272728,
            "no_speech_prob": 0.03515625,
            "words": [
                {
                    "start": 1035.84,
                    "end": 1035.98,
                    "word": " which",
                    "probability": 0.91552734375
                },
                {
                    "start": 1035.98,
                    "end": 1036.12,
                    "word": " were",
                    "probability": 0.5048828125
                },
                {
                    "start": 1036.12,
                    "end": 1036.3,
                    "word": " all",
                    "probability": 0.8896484375
                },
                {
                    "start": 1036.3,
                    "end": 1036.74,
                    "word": " discrete",
                    "probability": 0.7109375
                },
                {
                    "start": 1036.74,
                    "end": 1036.96,
                    "word": " and",
                    "probability": 0.783203125
                },
                {
                    "start": 1036.96,
                    "end": 1037.26,
                    "word": " easy",
                    "probability": 0.81396484375
                },
                {
                    "start": 1037.26,
                    "end": 1037.44,
                    "word": " to",
                    "probability": 0.91162109375
                },
                {
                    "start": 1037.44,
                    "end": 1037.84,
                    "word": " compute",
                    "probability": 0.8818359375
                },
                {
                    "start": 1037.84,
                    "end": 1038.22,
                    "word": ",",
                    "probability": 0.70947265625
                },
                {
                    "start": 1038.22,
                    "end": 1038.36,
                    "word": " and",
                    "probability": 0.58251953125
                },
                {
                    "start": 1038.36,
                    "end": 1038.66,
                    "word": " easy",
                    "probability": 0.818359375
                },
                {
                    "start": 1038.66,
                    "end": 1038.84,
                    "word": " to",
                    "probability": 0.90380859375
                },
                {
                    "start": 1038.84,
                    "end": 1039.16,
                    "word": " count",
                    "probability": 0.7880859375
                },
                {
                    "start": 1039.16,
                    "end": 1039.26,
                    "word": ",",
                    "probability": 0.29443359375
                },
                {
                    "start": 1039.26,
                    "end": 1039.34,
                    "word": " and",
                    "probability": 0.85107421875
                },
                {
                    "start": 1039.34,
                    "end": 1039.56,
                    "word": " easy",
                    "probability": 0.82470703125
                },
                {
                    "start": 1039.56,
                    "end": 1039.76,
                    "word": " to",
                    "probability": 0.88818359375
                },
                {
                    "start": 1039.76,
                    "end": 1039.9,
                    "word": " keep",
                    "probability": 0.591796875
                },
                {
                    "start": 1039.9,
                    "end": 1040.22,
                    "word": " track",
                    "probability": 0.85009765625
                },
                {
                    "start": 1040.22,
                    "end": 1040.5,
                    "word": " of",
                    "probability": 0.884765625
                },
                {
                    "start": 1040.5,
                    "end": 1042.68,
                    "word": ",",
                    "probability": 0.62890625
                }
            ]
        },
        {
            "id": 154,
            "seek": 90704,
            "start": 1042.68,
            "end": 1047.26,
            "text": " now I can compute the KL divergence, because it's trivial. It's all... it's just counting,",
            "tokens": [
                51328,
                586,
                286,
                393,
                14722,
                264,
                47991,
                47387,
                11,
                570,
                309,
                311,
                26703,
                13,
                467,
                311,
                439,
                485,
                309,
                311,
                445,
                13251,
                11,
                51548
            ],
            "temperature": 0.0,
            "avg_logprob": -0.20991847680314726,
            "compression_ratio": 1.6227272727272728,
            "no_speech_prob": 0.03515625,
            "words": [
                {
                    "start": 1042.68,
                    "end": 1042.98,
                    "word": " now",
                    "probability": 0.77734375
                },
                {
                    "start": 1042.98,
                    "end": 1043.34,
                    "word": " I",
                    "probability": 0.83154296875
                },
                {
                    "start": 1043.34,
                    "end": 1043.5,
                    "word": " can",
                    "probability": 0.88916015625
                },
                {
                    "start": 1043.5,
                    "end": 1043.78,
                    "word": " compute",
                    "probability": 0.869140625
                },
                {
                    "start": 1043.78,
                    "end": 1043.92,
                    "word": " the",
                    "probability": 0.806640625
                },
                {
                    "start": 1043.92,
                    "end": 1044.12,
                    "word": " KL",
                    "probability": 0.9833984375
                },
                {
                    "start": 1044.12,
                    "end": 1044.46,
                    "word": " divergence",
                    "probability": 0.68603515625
                },
                {
                    "start": 1044.46,
                    "end": 1044.66,
                    "word": ",",
                    "probability": 0.258056640625
                },
                {
                    "start": 1044.66,
                    "end": 1044.82,
                    "word": " because",
                    "probability": 0.8271484375
                },
                {
                    "start": 1044.82,
                    "end": 1045.36,
                    "word": " it's",
                    "probability": 0.95068359375
                },
                {
                    "start": 1045.36,
                    "end": 1045.68,
                    "word": " trivial",
                    "probability": 0.82958984375
                },
                {
                    "start": 1045.68,
                    "end": 1046.1,
                    "word": ".",
                    "probability": 0.84423828125
                },
                {
                    "start": 1046.1,
                    "end": 1046.32,
                    "word": " It's",
                    "probability": 0.945556640625
                },
                {
                    "start": 1046.32,
                    "end": 1046.6,
                    "word": " all...",
                    "probability": 0.20135498046875
                },
                {
                    "start": 1046.6,
                    "end": 1046.66,
                    "word": " it's",
                    "probability": 0.726318359375
                },
                {
                    "start": 1046.66,
                    "end": 1046.8,
                    "word": " just",
                    "probability": 0.8857421875
                },
                {
                    "start": 1046.8,
                    "end": 1047.14,
                    "word": " counting",
                    "probability": 0.76904296875
                },
                {
                    "start": 1047.14,
                    "end": 1047.26,
                    "word": ",",
                    "probability": 0.28466796875
                }
            ]
        },
        {
            "id": 155,
            "seek": 93398,
            "start": 1047.26,
            "end": 1054.52,
            "text": " taking logs, and summing things up. So this is my prior distribution. As I said,",
            "tokens": [
                50364,
                1940,
                20820,
                11,
                293,
                2408,
                2810,
                721,
                493,
                13,
                407,
                341,
                307,
                452,
                4059,
                7316,
                13,
                1018,
                286,
                848,
                11,
                50720
            ],
            "temperature": 0.0,
            "avg_logprob": -0.17920919066789198,
            "compression_ratio": 1.5782608695652174,
            "no_speech_prob": 0.12939453125,
            "words": [
                {
                    "start": 1047.26,
                    "end": 1047.54,
                    "word": " taking",
                    "probability": 0.07769775390625
                },
                {
                    "start": 1047.54,
                    "end": 1047.9,
                    "word": " logs",
                    "probability": 0.7626953125
                },
                {
                    "start": 1047.9,
                    "end": 1048.04,
                    "word": ",",
                    "probability": 0.03662109375
                },
                {
                    "start": 1048.04,
                    "end": 1048.06,
                    "word": " and",
                    "probability": 0.828125
                },
                {
                    "start": 1048.06,
                    "end": 1049.02,
                    "word": " summing",
                    "probability": 0.8369140625
                },
                {
                    "start": 1049.02,
                    "end": 1049.18,
                    "word": " things",
                    "probability": 0.71435546875
                },
                {
                    "start": 1049.18,
                    "end": 1049.4,
                    "word": " up",
                    "probability": 0.90234375
                },
                {
                    "start": 1049.4,
                    "end": 1049.88,
                    "word": ".",
                    "probability": 0.9091796875
                },
                {
                    "start": 1050.96,
                    "end": 1051.44,
                    "word": " So",
                    "probability": 0.71826171875
                },
                {
                    "start": 1051.44,
                    "end": 1052.08,
                    "word": " this",
                    "probability": 0.6962890625
                },
                {
                    "start": 1052.08,
                    "end": 1052.26,
                    "word": " is",
                    "probability": 0.90625
                },
                {
                    "start": 1052.26,
                    "end": 1052.6,
                    "word": " my",
                    "probability": 0.9130859375
                },
                {
                    "start": 1052.6,
                    "end": 1053.4,
                    "word": " prior",
                    "probability": 0.8486328125
                },
                {
                    "start": 1053.4,
                    "end": 1053.88,
                    "word": " distribution",
                    "probability": 0.73193359375
                },
                {
                    "start": 1053.88,
                    "end": 1053.96,
                    "word": ".",
                    "probability": 0.6474609375
                },
                {
                    "start": 1053.96,
                    "end": 1054.06,
                    "word": " As",
                    "probability": 0.86572265625
                },
                {
                    "start": 1054.06,
                    "end": 1054.16,
                    "word": " I",
                    "probability": 0.93505859375
                },
                {
                    "start": 1054.16,
                    "end": 1054.34,
                    "word": " said",
                    "probability": 0.86865234375
                },
                {
                    "start": 1054.34,
                    "end": 1054.52,
                    "word": ",",
                    "probability": 0.619140625
                }
            ]
        },
        {
            "id": 156,
            "seek": 93398,
            "start": 1054.52,
            "end": 1059.5,
            "text": " we're going to use some Bayesian statistics. And so here is my prior. How... this is built",
            "tokens": [
                50720,
                321,
                434,
                516,
                281,
                764,
                512,
                7840,
                42434,
                12523,
                13,
                400,
                370,
                510,
                307,
                452,
                4059,
                13,
                1012,
                485,
                341,
                307,
                3094,
                50972
            ],
            "temperature": 0.0,
            "avg_logprob": -0.17920919066789198,
            "compression_ratio": 1.5782608695652174,
            "no_speech_prob": 0.12939453125,
            "words": [
                {
                    "start": 1054.52,
                    "end": 1054.68,
                    "word": " we're",
                    "probability": 0.906005859375
                },
                {
                    "start": 1054.68,
                    "end": 1054.74,
                    "word": " going",
                    "probability": 0.79931640625
                },
                {
                    "start": 1054.74,
                    "end": 1054.78,
                    "word": " to",
                    "probability": 0.90869140625
                },
                {
                    "start": 1054.78,
                    "end": 1054.96,
                    "word": " use",
                    "probability": 0.822265625
                },
                {
                    "start": 1054.96,
                    "end": 1055.14,
                    "word": " some",
                    "probability": 0.71484375
                },
                {
                    "start": 1055.14,
                    "end": 1055.44,
                    "word": " Bayesian",
                    "probability": 0.95166015625
                },
                {
                    "start": 1055.44,
                    "end": 1055.92,
                    "word": " statistics",
                    "probability": 0.880859375
                },
                {
                    "start": 1055.92,
                    "end": 1056.22,
                    "word": ".",
                    "probability": 0.62744140625
                },
                {
                    "start": 1056.22,
                    "end": 1056.22,
                    "word": " And",
                    "probability": 0.3994140625
                },
                {
                    "start": 1056.22,
                    "end": 1056.56,
                    "word": " so",
                    "probability": 0.86572265625
                },
                {
                    "start": 1056.56,
                    "end": 1056.82,
                    "word": " here",
                    "probability": 0.76611328125
                },
                {
                    "start": 1056.82,
                    "end": 1056.98,
                    "word": " is",
                    "probability": 0.8330078125
                },
                {
                    "start": 1056.98,
                    "end": 1057.14,
                    "word": " my",
                    "probability": 0.90771484375
                },
                {
                    "start": 1057.14,
                    "end": 1057.54,
                    "word": " prior",
                    "probability": 0.84130859375
                },
                {
                    "start": 1057.54,
                    "end": 1057.92,
                    "word": ".",
                    "probability": 0.7353515625
                },
                {
                    "start": 1058.28,
                    "end": 1058.76,
                    "word": " How...",
                    "probability": 0.3247222900390625
                },
                {
                    "start": 1058.76,
                    "end": 1059.14,
                    "word": " this",
                    "probability": 0.3837890625
                },
                {
                    "start": 1059.14,
                    "end": 1059.3,
                    "word": " is",
                    "probability": 0.900390625
                },
                {
                    "start": 1059.3,
                    "end": 1059.5,
                    "word": " built",
                    "probability": 0.92724609375
                }
            ]
        },
        {
            "id": 157,
            "seek": 93398,
            "start": 1059.5,
            "end": 1066.9,
            "text": " with my training set. How do I include my test set? So when I see a piece of data coming over",
            "tokens": [
                50972,
                365,
                452,
                3097,
                992,
                13,
                1012,
                360,
                286,
                4090,
                452,
                1500,
                992,
                30,
                407,
                562,
                286,
                536,
                257,
                2522,
                295,
                1412,
                1348,
                670,
                51344
            ],
            "temperature": 0.0,
            "avg_logprob": -0.17920919066789198,
            "compression_ratio": 1.5782608695652174,
            "no_speech_prob": 0.12939453125,
            "words": [
                {
                    "start": 1059.5,
                    "end": 1059.64,
                    "word": " with",
                    "probability": 0.347412109375
                },
                {
                    "start": 1059.64,
                    "end": 1059.8,
                    "word": " my",
                    "probability": 0.9150390625
                },
                {
                    "start": 1059.8,
                    "end": 1060.06,
                    "word": " training",
                    "probability": 0.76904296875
                },
                {
                    "start": 1060.06,
                    "end": 1060.28,
                    "word": " set",
                    "probability": 0.86083984375
                },
                {
                    "start": 1060.28,
                    "end": 1060.5,
                    "word": ".",
                    "probability": 0.90478515625
                },
                {
                    "start": 1060.5,
                    "end": 1060.8,
                    "word": " How",
                    "probability": 0.884765625
                },
                {
                    "start": 1060.8,
                    "end": 1060.96,
                    "word": " do",
                    "probability": 0.9033203125
                },
                {
                    "start": 1060.96,
                    "end": 1061.1,
                    "word": " I",
                    "probability": 0.97216796875
                },
                {
                    "start": 1061.1,
                    "end": 1061.72,
                    "word": " include",
                    "probability": 0.75390625
                },
                {
                    "start": 1061.72,
                    "end": 1062.24,
                    "word": " my",
                    "probability": 0.89697265625
                },
                {
                    "start": 1062.24,
                    "end": 1062.46,
                    "word": " test",
                    "probability": 0.82421875
                },
                {
                    "start": 1062.46,
                    "end": 1062.6,
                    "word": " set",
                    "probability": 0.86376953125
                },
                {
                    "start": 1062.6,
                    "end": 1063.08,
                    "word": "?",
                    "probability": 0.93408203125
                },
                {
                    "start": 1064.12,
                    "end": 1064.6,
                    "word": " So",
                    "probability": 0.85498046875
                },
                {
                    "start": 1064.6,
                    "end": 1065.04,
                    "word": " when",
                    "probability": 0.76513671875
                },
                {
                    "start": 1065.04,
                    "end": 1065.22,
                    "word": " I",
                    "probability": 0.9833984375
                },
                {
                    "start": 1065.22,
                    "end": 1065.54,
                    "word": " see",
                    "probability": 0.81396484375
                },
                {
                    "start": 1065.74,
                    "end": 1065.94,
                    "word": " a piece",
                    "probability": 0.794921875
                },
                {
                    "start": 1065.94,
                    "end": 1066.1,
                    "word": " of",
                    "probability": 0.90966796875
                },
                {
                    "start": 1066.1,
                    "end": 1066.38,
                    "word": " data",
                    "probability": 0.810546875
                },
                {
                    "start": 1066.38,
                    "end": 1066.72,
                    "word": " coming",
                    "probability": 0.80419921875
                },
                {
                    "start": 1066.72,
                    "end": 1066.9,
                    "word": " over",
                    "probability": 0.88916015625
                }
            ]
        },
        {
            "id": 158,
            "seek": 93398,
            "start": 1066.9,
            "end": 1074.2,
            "text": " the wire, I can make an inference of where it should belong in this tree. So if it was already...",
            "tokens": [
                51344,
                264,
                6234,
                11,
                286,
                393,
                652,
                364,
                38253,
                295,
                689,
                309,
                820,
                5784,
                294,
                341,
                4230,
                13,
                407,
                498,
                309,
                390,
                1217,
                485,
                51708
            ],
            "temperature": 0.0,
            "avg_logprob": -0.17920919066789198,
            "compression_ratio": 1.5782608695652174,
            "no_speech_prob": 0.12939453125,
            "words": [
                {
                    "start": 1066.9,
                    "end": 1067.02,
                    "word": " the",
                    "probability": 0.818359375
                },
                {
                    "start": 1067.02,
                    "end": 1067.26,
                    "word": " wire",
                    "probability": 0.8037109375
                },
                {
                    "start": 1067.26,
                    "end": 1067.48,
                    "word": ",",
                    "probability": 0.8828125
                },
                {
                    "start": 1067.48,
                    "end": 1067.92,
                    "word": " I",
                    "probability": 0.9970703125
                },
                {
                    "start": 1067.92,
                    "end": 1068.3,
                    "word": " can",
                    "probability": 0.89208984375
                },
                {
                    "start": 1068.3,
                    "end": 1068.54,
                    "word": " make",
                    "probability": 0.880859375
                },
                {
                    "start": 1068.54,
                    "end": 1068.68,
                    "word": " an",
                    "probability": 0.90380859375
                },
                {
                    "start": 1068.68,
                    "end": 1069.04,
                    "word": " inference",
                    "probability": 0.80712890625
                },
                {
                    "start": 1069.04,
                    "end": 1069.4,
                    "word": " of",
                    "probability": 0.8974609375
                },
                {
                    "start": 1069.4,
                    "end": 1069.6,
                    "word": " where",
                    "probability": 0.90625
                },
                {
                    "start": 1069.6,
                    "end": 1069.78,
                    "word": " it",
                    "probability": 0.896484375
                },
                {
                    "start": 1069.78,
                    "end": 1070.06,
                    "word": " should",
                    "probability": 0.9189453125
                },
                {
                    "start": 1070.06,
                    "end": 1070.56,
                    "word": " belong",
                    "probability": 0.763671875
                },
                {
                    "start": 1070.56,
                    "end": 1071.1,
                    "word": " in",
                    "probability": 0.89599609375
                },
                {
                    "start": 1071.1,
                    "end": 1071.26,
                    "word": " this",
                    "probability": 0.849609375
                },
                {
                    "start": 1071.26,
                    "end": 1071.5,
                    "word": " tree",
                    "probability": 0.810546875
                },
                {
                    "start": 1071.5,
                    "end": 1071.98,
                    "word": ".",
                    "probability": 0.951171875
                },
                {
                    "start": 1072.02,
                    "end": 1072.26,
                    "word": " So",
                    "probability": 0.8759765625
                },
                {
                    "start": 1072.26,
                    "end": 1073.2,
                    "word": " if",
                    "probability": 0.8251953125
                },
                {
                    "start": 1073.2,
                    "end": 1073.34,
                    "word": " it",
                    "probability": 0.904296875
                },
                {
                    "start": 1073.34,
                    "end": 1073.64,
                    "word": " was",
                    "probability": 0.8818359375
                },
                {
                    "start": 1073.64,
                    "end": 1074.2,
                    "word": " already...",
                    "probability": 0.71630859375
                }
            ]
        },
        {
            "id": 159,
            "seek": 96176,
            "start": 1074.2,
            "end": 1078.4,
            "text": " if it was in my training set and didn't get involved in building the tree,",
            "tokens": [
                50364,
                498,
                309,
                390,
                294,
                452,
                3097,
                992,
                293,
                994,
                380,
                483,
                3288,
                294,
                2390,
                264,
                4230,
                11,
                50572
            ],
            "temperature": 0.0,
            "avg_logprob": -0.19192708482344945,
            "compression_ratio": 1.8181818181818181,
            "no_speech_prob": 0.00012934207916259766,
            "words": [
                {
                    "start": 1074.2,
                    "end": 1074.4,
                    "word": " if",
                    "probability": 0.12384033203125
                },
                {
                    "start": 1074.4,
                    "end": 1074.56,
                    "word": " it",
                    "probability": 0.8857421875
                },
                {
                    "start": 1074.56,
                    "end": 1074.74,
                    "word": " was",
                    "probability": 0.6220703125
                },
                {
                    "start": 1074.74,
                    "end": 1074.86,
                    "word": " in",
                    "probability": 0.8876953125
                },
                {
                    "start": 1074.86,
                    "end": 1075.0,
                    "word": " my",
                    "probability": 0.904296875
                },
                {
                    "start": 1075.0,
                    "end": 1075.26,
                    "word": " training",
                    "probability": 0.7861328125
                },
                {
                    "start": 1075.26,
                    "end": 1075.58,
                    "word": " set",
                    "probability": 0.8935546875
                },
                {
                    "start": 1075.58,
                    "end": 1075.78,
                    "word": " and",
                    "probability": 0.8505859375
                },
                {
                    "start": 1075.78,
                    "end": 1076.18,
                    "word": " didn't",
                    "probability": 0.92578125
                },
                {
                    "start": 1076.18,
                    "end": 1076.4,
                    "word": " get",
                    "probability": 0.89990234375
                },
                {
                    "start": 1076.4,
                    "end": 1077.12,
                    "word": " involved",
                    "probability": 0.7587890625
                },
                {
                    "start": 1077.12,
                    "end": 1077.32,
                    "word": " in",
                    "probability": 0.892578125
                },
                {
                    "start": 1077.32,
                    "end": 1077.58,
                    "word": " building",
                    "probability": 0.81640625
                },
                {
                    "start": 1077.58,
                    "end": 1077.78,
                    "word": " the",
                    "probability": 0.81005859375
                },
                {
                    "start": 1077.78,
                    "end": 1078.04,
                    "word": " tree",
                    "probability": 0.8251953125
                },
                {
                    "start": 1078.04,
                    "end": 1078.4,
                    "word": ",",
                    "probability": 0.483642578125
                }
            ]
        },
        {
            "id": 160,
            "seek": 96176,
            "start": 1078.4,
            "end": 1083.46,
            "text": " where would have it ended up? So I have a point here. So this is my new point that came over the",
            "tokens": [
                50572,
                689,
                576,
                362,
                309,
                4590,
                493,
                30,
                407,
                286,
                362,
                257,
                935,
                510,
                13,
                407,
                341,
                307,
                452,
                777,
                935,
                300,
                1361,
                670,
                264,
                50824
            ],
            "temperature": 0.0,
            "avg_logprob": -0.19192708482344945,
            "compression_ratio": 1.8181818181818181,
            "no_speech_prob": 0.00012934207916259766,
            "words": [
                {
                    "start": 1078.4,
                    "end": 1078.62,
                    "word": " where",
                    "probability": 0.8916015625
                },
                {
                    "start": 1078.62,
                    "end": 1078.8,
                    "word": " would",
                    "probability": 0.84521484375
                },
                {
                    "start": 1078.8,
                    "end": 1078.92,
                    "word": " have",
                    "probability": 0.4853515625
                },
                {
                    "start": 1078.92,
                    "end": 1079.04,
                    "word": " it",
                    "probability": 0.88232421875
                },
                {
                    "start": 1079.04,
                    "end": 1079.22,
                    "word": " ended",
                    "probability": 0.74658203125
                },
                {
                    "start": 1079.22,
                    "end": 1079.48,
                    "word": " up",
                    "probability": 0.9052734375
                },
                {
                    "start": 1079.48,
                    "end": 1079.92,
                    "word": "?",
                    "probability": 0.87451171875
                },
                {
                    "start": 1080.06,
                    "end": 1080.32,
                    "word": " So",
                    "probability": 0.81298828125
                },
                {
                    "start": 1080.32,
                    "end": 1080.9,
                    "word": " I",
                    "probability": 0.6953125
                },
                {
                    "start": 1080.9,
                    "end": 1081.06,
                    "word": " have",
                    "probability": 0.88037109375
                },
                {
                    "start": 1081.16,
                    "end": 1081.32,
                    "word": " a point",
                    "probability": 0.8486328125
                },
                {
                    "start": 1081.32,
                    "end": 1081.6,
                    "word": " here",
                    "probability": 0.81298828125
                },
                {
                    "start": 1081.6,
                    "end": 1081.76,
                    "word": ".",
                    "probability": 0.67333984375
                },
                {
                    "start": 1081.76,
                    "end": 1081.92,
                    "word": " So",
                    "probability": 0.78369140625
                },
                {
                    "start": 1081.92,
                    "end": 1082.08,
                    "word": " this",
                    "probability": 0.87060546875
                },
                {
                    "start": 1082.08,
                    "end": 1082.2,
                    "word": " is",
                    "probability": 0.9013671875
                },
                {
                    "start": 1082.2,
                    "end": 1082.34,
                    "word": " my",
                    "probability": 0.90869140625
                },
                {
                    "start": 1082.34,
                    "end": 1082.5,
                    "word": " new",
                    "probability": 0.814453125
                },
                {
                    "start": 1082.5,
                    "end": 1082.78,
                    "word": " point",
                    "probability": 0.8671875
                },
                {
                    "start": 1082.78,
                    "end": 1082.98,
                    "word": " that",
                    "probability": 0.70068359375
                },
                {
                    "start": 1082.98,
                    "end": 1083.16,
                    "word": " came",
                    "probability": 0.8095703125
                },
                {
                    "start": 1083.16,
                    "end": 1083.3,
                    "word": " over",
                    "probability": 0.8828125
                },
                {
                    "start": 1083.3,
                    "end": 1083.46,
                    "word": " the",
                    "probability": 0.81787109375
                }
            ]
        },
        {
            "id": 161,
            "seek": 96176,
            "start": 1083.46,
            "end": 1091.02,
            "text": " wire. And I queried and said, oh well, this point is covered by this parent, and it's covered by",
            "tokens": [
                50824,
                6234,
                13,
                400,
                286,
                7083,
                1091,
                293,
                848,
                11,
                1954,
                731,
                11,
                341,
                935,
                307,
                5343,
                538,
                341,
                2596,
                11,
                293,
                309,
                311,
                5343,
                538,
                51204
            ],
            "temperature": 0.0,
            "avg_logprob": -0.19192708482344945,
            "compression_ratio": 1.8181818181818181,
            "no_speech_prob": 0.00012934207916259766,
            "words": [
                {
                    "start": 1083.46,
                    "end": 1083.72,
                    "word": " wire",
                    "probability": 0.806640625
                },
                {
                    "start": 1083.72,
                    "end": 1084.14,
                    "word": ".",
                    "probability": 0.283447265625
                },
                {
                    "start": 1084.14,
                    "end": 1084.22,
                    "word": " And",
                    "probability": 0.359375
                },
                {
                    "start": 1084.22,
                    "end": 1084.32,
                    "word": " I",
                    "probability": 0.98193359375
                },
                {
                    "start": 1084.32,
                    "end": 1084.76,
                    "word": " queried",
                    "probability": 0.90966796875
                },
                {
                    "start": 1084.76,
                    "end": 1085.02,
                    "word": " and",
                    "probability": 0.8486328125
                },
                {
                    "start": 1085.02,
                    "end": 1085.2,
                    "word": " said",
                    "probability": 0.87841796875
                },
                {
                    "start": 1085.2,
                    "end": 1085.42,
                    "word": ",",
                    "probability": 0.65771484375
                },
                {
                    "start": 1085.42,
                    "end": 1085.5,
                    "word": " oh",
                    "probability": 0.50244140625
                },
                {
                    "start": 1085.5,
                    "end": 1085.72,
                    "word": " well",
                    "probability": 0.4130859375
                },
                {
                    "start": 1085.72,
                    "end": 1085.8,
                    "word": ",",
                    "probability": 0.5380859375
                },
                {
                    "start": 1085.8,
                    "end": 1086.16,
                    "word": " this",
                    "probability": 0.87548828125
                },
                {
                    "start": 1086.16,
                    "end": 1086.6,
                    "word": " point",
                    "probability": 0.85595703125
                },
                {
                    "start": 1086.6,
                    "end": 1087.02,
                    "word": " is",
                    "probability": 0.9013671875
                },
                {
                    "start": 1087.02,
                    "end": 1087.34,
                    "word": " covered",
                    "probability": 0.703125
                },
                {
                    "start": 1087.34,
                    "end": 1087.78,
                    "word": " by",
                    "probability": 0.904296875
                },
                {
                    "start": 1087.78,
                    "end": 1088.1,
                    "word": " this",
                    "probability": 0.88037109375
                },
                {
                    "start": 1088.1,
                    "end": 1088.52,
                    "word": " parent",
                    "probability": 0.7314453125
                },
                {
                    "start": 1088.52,
                    "end": 1089.12,
                    "word": ",",
                    "probability": 0.272705078125
                },
                {
                    "start": 1089.12,
                    "end": 1090.24,
                    "word": " and",
                    "probability": 0.796875
                },
                {
                    "start": 1090.24,
                    "end": 1090.7,
                    "word": " it's",
                    "probability": 0.938720703125
                },
                {
                    "start": 1090.7,
                    "end": 1090.86,
                    "word": " covered",
                    "probability": 0.72509765625
                },
                {
                    "start": 1090.86,
                    "end": 1091.02,
                    "word": " by",
                    "probability": 0.90380859375
                }
            ]
        },
        {
            "id": 162,
            "seek": 96176,
            "start": 1091.02,
            "end": 1097.46,
            "text": " that parent, and that parent, and all the way up. So the path for this point goes up this way.",
            "tokens": [
                51204,
                300,
                2596,
                11,
                293,
                300,
                2596,
                11,
                293,
                439,
                264,
                636,
                493,
                13,
                407,
                264,
                3100,
                337,
                341,
                935,
                1709,
                493,
                341,
                636,
                13,
                51524
            ],
            "temperature": 0.0,
            "avg_logprob": -0.19192708482344945,
            "compression_ratio": 1.8181818181818181,
            "no_speech_prob": 0.00012934207916259766,
            "words": [
                {
                    "start": 1091.02,
                    "end": 1091.2,
                    "word": " that",
                    "probability": 0.8837890625
                },
                {
                    "start": 1091.2,
                    "end": 1091.46,
                    "word": " parent",
                    "probability": 0.7548828125
                },
                {
                    "start": 1091.46,
                    "end": 1091.52,
                    "word": ",",
                    "probability": 0.309326171875
                },
                {
                    "start": 1091.52,
                    "end": 1091.6,
                    "word": " and",
                    "probability": 0.8828125
                },
                {
                    "start": 1091.6,
                    "end": 1091.74,
                    "word": " that",
                    "probability": 0.8671875
                },
                {
                    "start": 1091.74,
                    "end": 1092.0,
                    "word": " parent",
                    "probability": 0.775390625
                },
                {
                    "start": 1092.0,
                    "end": 1092.06,
                    "word": ",",
                    "probability": 0.734375
                },
                {
                    "start": 1092.06,
                    "end": 1092.12,
                    "word": " and",
                    "probability": 0.46435546875
                },
                {
                    "start": 1092.12,
                    "end": 1092.26,
                    "word": " all",
                    "probability": 0.8984375
                },
                {
                    "start": 1092.26,
                    "end": 1092.36,
                    "word": " the",
                    "probability": 0.81298828125
                },
                {
                    "start": 1092.36,
                    "end": 1092.46,
                    "word": " way",
                    "probability": 0.89990234375
                },
                {
                    "start": 1092.46,
                    "end": 1092.68,
                    "word": " up",
                    "probability": 0.8994140625
                },
                {
                    "start": 1092.68,
                    "end": 1093.12,
                    "word": ".",
                    "probability": 0.9736328125
                },
                {
                    "start": 1093.18,
                    "end": 1093.52,
                    "word": " So",
                    "probability": 0.896484375
                },
                {
                    "start": 1093.52,
                    "end": 1093.68,
                    "word": " the",
                    "probability": 0.8095703125
                },
                {
                    "start": 1093.68,
                    "end": 1094.12,
                    "word": " path",
                    "probability": 0.8544921875
                },
                {
                    "start": 1094.12,
                    "end": 1094.42,
                    "word": " for",
                    "probability": 0.8818359375
                },
                {
                    "start": 1094.42,
                    "end": 1094.62,
                    "word": " this",
                    "probability": 0.88671875
                },
                {
                    "start": 1094.62,
                    "end": 1094.98,
                    "word": " point",
                    "probability": 0.859375
                },
                {
                    "start": 1094.98,
                    "end": 1096.26,
                    "word": " goes",
                    "probability": 0.8154296875
                },
                {
                    "start": 1096.26,
                    "end": 1096.56,
                    "word": " up",
                    "probability": 0.90087890625
                },
                {
                    "start": 1096.56,
                    "end": 1097.02,
                    "word": " this",
                    "probability": 0.88232421875
                },
                {
                    "start": 1097.02,
                    "end": 1097.3,
                    "word": " way",
                    "probability": 0.89111328125
                },
                {
                    "start": 1097.3,
                    "end": 1097.46,
                    "word": ".",
                    "probability": 0.8916015625
                }
            ]
        },
        {
            "id": 163,
            "seek": 96176,
            "start": 1097.46,
            "end": 1101.98,
            "text": " And now for each of those elements in those paths, I can increment my probabilistic distribution",
            "tokens": [
                51524,
                400,
                586,
                337,
                1184,
                295,
                729,
                4959,
                294,
                729,
                14518,
                11,
                286,
                393,
                26200,
                452,
                31959,
                3142,
                7316,
                51768
            ],
            "temperature": 0.0,
            "avg_logprob": -0.19192708482344945,
            "compression_ratio": 1.8181818181818181,
            "no_speech_prob": 0.00012934207916259766,
            "words": [
                {
                    "start": 1097.46,
                    "end": 1097.62,
                    "word": " And",
                    "probability": 0.87451171875
                },
                {
                    "start": 1097.62,
                    "end": 1097.84,
                    "word": " now",
                    "probability": 0.892578125
                },
                {
                    "start": 1097.84,
                    "end": 1098.04,
                    "word": " for",
                    "probability": 0.77978515625
                },
                {
                    "start": 1098.04,
                    "end": 1098.26,
                    "word": " each",
                    "probability": 0.865234375
                },
                {
                    "start": 1098.26,
                    "end": 1098.38,
                    "word": " of",
                    "probability": 0.900390625
                },
                {
                    "start": 1098.38,
                    "end": 1098.62,
                    "word": " those",
                    "probability": 0.84521484375
                },
                {
                    "start": 1098.62,
                    "end": 1098.96,
                    "word": " elements",
                    "probability": 0.7900390625
                },
                {
                    "start": 1098.96,
                    "end": 1099.1,
                    "word": " in",
                    "probability": 0.82958984375
                },
                {
                    "start": 1099.1,
                    "end": 1099.26,
                    "word": " those",
                    "probability": 0.71630859375
                },
                {
                    "start": 1099.26,
                    "end": 1099.54,
                    "word": " paths",
                    "probability": 0.71728515625
                },
                {
                    "start": 1099.54,
                    "end": 1099.62,
                    "word": ",",
                    "probability": 0.81787109375
                },
                {
                    "start": 1099.62,
                    "end": 1099.84,
                    "word": " I",
                    "probability": 0.9931640625
                },
                {
                    "start": 1099.84,
                    "end": 1100.02,
                    "word": " can",
                    "probability": 0.890625
                },
                {
                    "start": 1100.02,
                    "end": 1100.46,
                    "word": " increment",
                    "probability": 0.8564453125
                },
                {
                    "start": 1100.46,
                    "end": 1100.82,
                    "word": " my",
                    "probability": 0.90185546875
                },
                {
                    "start": 1100.82,
                    "end": 1101.44,
                    "word": " probabilistic",
                    "probability": 0.5946044921875
                },
                {
                    "start": 1101.44,
                    "end": 1101.98,
                    "word": " distribution",
                    "probability": 0.72607421875
                }
            ]
        },
        {
            "id": 164,
            "seek": 98604,
            "start": 1101.98,
            "end": 1106.86,
            "text": " and get a posterior one that involves... that includes the new information that I got,",
            "tokens": [
                50380,
                293,
                483,
                257,
                33529,
                472,
                300,
                11626,
                485,
                300,
                5974,
                264,
                777,
                1589,
                300,
                286,
                658,
                11,
                50600
            ],
            "temperature": 0.0,
            "avg_logprob": -0.2314967046442785,
            "compression_ratio": 1.853061224489796,
            "no_speech_prob": 0.0014781951904296875,
            "words": [
                {
                    "start": 1101.98,
                    "end": 1102.48,
                    "word": " and",
                    "probability": 0.32373046875
                },
                {
                    "start": 1102.48,
                    "end": 1102.68,
                    "word": " get",
                    "probability": 0.73876953125
                },
                {
                    "start": 1102.88,
                    "end": 1103.3,
                    "word": " a posterior",
                    "probability": 0.67529296875
                },
                {
                    "start": 1103.3,
                    "end": 1103.68,
                    "word": " one",
                    "probability": 0.76806640625
                },
                {
                    "start": 1103.68,
                    "end": 1103.88,
                    "word": " that",
                    "probability": 0.83203125
                },
                {
                    "start": 1103.88,
                    "end": 1104.42,
                    "word": " involves...",
                    "probability": 0.05239105224609375
                },
                {
                    "start": 1104.42,
                    "end": 1104.54,
                    "word": " that",
                    "probability": 0.69970703125
                },
                {
                    "start": 1104.54,
                    "end": 1105.02,
                    "word": " includes",
                    "probability": 0.830078125
                },
                {
                    "start": 1105.02,
                    "end": 1105.26,
                    "word": " the",
                    "probability": 0.517578125
                },
                {
                    "start": 1105.26,
                    "end": 1105.52,
                    "word": " new",
                    "probability": 0.810546875
                },
                {
                    "start": 1105.52,
                    "end": 1105.94,
                    "word": " information",
                    "probability": 0.72216796875
                },
                {
                    "start": 1105.94,
                    "end": 1106.16,
                    "word": " that",
                    "probability": 0.8271484375
                },
                {
                    "start": 1106.16,
                    "end": 1106.24,
                    "word": " I",
                    "probability": 0.9912109375
                },
                {
                    "start": 1106.24,
                    "end": 1106.48,
                    "word": " got",
                    "probability": 0.82421875
                },
                {
                    "start": 1106.48,
                    "end": 1106.86,
                    "word": ",",
                    "probability": 0.2271728515625
                }
            ]
        },
        {
            "id": 165,
            "seek": 98604,
            "start": 1106.86,
            "end": 1111.32,
            "text": " that there's a new point over here. And now I can do that again with another point.",
            "tokens": [
                50600,
                300,
                456,
                311,
                257,
                777,
                935,
                670,
                510,
                13,
                400,
                586,
                286,
                393,
                360,
                300,
                797,
                365,
                1071,
                935,
                13,
                50824
            ],
            "temperature": 0.0,
            "avg_logprob": -0.2314967046442785,
            "compression_ratio": 1.853061224489796,
            "no_speech_prob": 0.0014781951904296875,
            "words": [
                {
                    "start": 1106.86,
                    "end": 1106.98,
                    "word": " that",
                    "probability": 0.7333984375
                },
                {
                    "start": 1106.98,
                    "end": 1107.18,
                    "word": " there's",
                    "probability": 0.6767578125
                },
                {
                    "start": 1107.26,
                    "end": 1107.4,
                    "word": " a new",
                    "probability": 0.810546875
                },
                {
                    "start": 1107.4,
                    "end": 1107.62,
                    "word": " point",
                    "probability": 0.85546875
                },
                {
                    "start": 1107.62,
                    "end": 1107.8,
                    "word": " over",
                    "probability": 0.87109375
                },
                {
                    "start": 1107.8,
                    "end": 1108.08,
                    "word": " here",
                    "probability": 0.81201171875
                },
                {
                    "start": 1108.08,
                    "end": 1108.48,
                    "word": ".",
                    "probability": 0.89599609375
                },
                {
                    "start": 1108.84,
                    "end": 1109.24,
                    "word": " And",
                    "probability": 0.8408203125
                },
                {
                    "start": 1109.24,
                    "end": 1109.44,
                    "word": " now",
                    "probability": 0.888671875
                },
                {
                    "start": 1109.44,
                    "end": 1109.58,
                    "word": " I",
                    "probability": 0.90966796875
                },
                {
                    "start": 1109.58,
                    "end": 1109.68,
                    "word": " can",
                    "probability": 0.8896484375
                },
                {
                    "start": 1109.68,
                    "end": 1109.78,
                    "word": " do",
                    "probability": 0.9072265625
                },
                {
                    "start": 1109.78,
                    "end": 1109.92,
                    "word": " that",
                    "probability": 0.88134765625
                },
                {
                    "start": 1109.92,
                    "end": 1110.2,
                    "word": " again",
                    "probability": 0.916015625
                },
                {
                    "start": 1110.2,
                    "end": 1110.38,
                    "word": " with",
                    "probability": 0.7421875
                },
                {
                    "start": 1110.38,
                    "end": 1110.58,
                    "word": " another",
                    "probability": 0.81884765625
                },
                {
                    "start": 1110.58,
                    "end": 1110.92,
                    "word": " point",
                    "probability": 0.86279296875
                },
                {
                    "start": 1110.92,
                    "end": 1111.32,
                    "word": ".",
                    "probability": 0.7568359375
                }
            ]
        },
        {
            "id": 166,
            "seek": 98604,
            "start": 1111.32,
            "end": 1116.1,
            "text": " Oh yep, cool. I have a posterior... this point's been added, and you can see that the distribution",
            "tokens": [
                50824,
                876,
                18633,
                11,
                1627,
                13,
                286,
                362,
                257,
                33529,
                485,
                341,
                935,
                311,
                668,
                3869,
                11,
                293,
                291,
                393,
                536,
                300,
                264,
                7316,
                51076
            ],
            "temperature": 0.0,
            "avg_logprob": -0.2314967046442785,
            "compression_ratio": 1.853061224489796,
            "no_speech_prob": 0.0014781951904296875,
            "words": [
                {
                    "start": 1111.32,
                    "end": 1111.46,
                    "word": " Oh",
                    "probability": 0.41552734375
                },
                {
                    "start": 1111.46,
                    "end": 1111.78,
                    "word": " yep",
                    "probability": 0.046295166015625
                },
                {
                    "start": 1111.78,
                    "end": 1112.06,
                    "word": ",",
                    "probability": 0.646484375
                },
                {
                    "start": 1112.06,
                    "end": 1112.3,
                    "word": " cool",
                    "probability": 0.84521484375
                },
                {
                    "start": 1112.3,
                    "end": 1112.42,
                    "word": ".",
                    "probability": 0.413330078125
                },
                {
                    "start": 1112.42,
                    "end": 1112.5,
                    "word": " I",
                    "probability": 0.9873046875
                },
                {
                    "start": 1112.5,
                    "end": 1112.66,
                    "word": " have",
                    "probability": 0.8828125
                },
                {
                    "start": 1112.78,
                    "end": 1113.48,
                    "word": " a posterior...",
                    "probability": 0.60498046875
                },
                {
                    "start": 1113.48,
                    "end": 1113.76,
                    "word": " this",
                    "probability": 0.5791015625
                },
                {
                    "start": 1113.76,
                    "end": 1114.24,
                    "word": " point's",
                    "probability": 0.75439453125
                },
                {
                    "start": 1114.24,
                    "end": 1114.36,
                    "word": " been",
                    "probability": 0.8154296875
                },
                {
                    "start": 1114.36,
                    "end": 1114.64,
                    "word": " added",
                    "probability": 0.7958984375
                },
                {
                    "start": 1114.64,
                    "end": 1114.8,
                    "word": ",",
                    "probability": 0.32421875
                },
                {
                    "start": 1114.8,
                    "end": 1114.86,
                    "word": " and",
                    "probability": 0.8876953125
                },
                {
                    "start": 1114.86,
                    "end": 1114.98,
                    "word": " you",
                    "probability": 0.89501953125
                },
                {
                    "start": 1114.98,
                    "end": 1115.1,
                    "word": " can",
                    "probability": 0.89599609375
                },
                {
                    "start": 1115.1,
                    "end": 1115.28,
                    "word": " see",
                    "probability": 0.8173828125
                },
                {
                    "start": 1115.28,
                    "end": 1115.5,
                    "word": " that",
                    "probability": 0.85986328125
                },
                {
                    "start": 1115.5,
                    "end": 1115.68,
                    "word": " the",
                    "probability": 0.8076171875
                },
                {
                    "start": 1115.68,
                    "end": 1116.1,
                    "word": " distribution",
                    "probability": 0.74609375
                }
            ]
        },
        {
            "id": 167,
            "seek": 98604,
            "start": 1116.1,
            "end": 1121.14,
            "text": " changes. And then I can add some more points, and I can add some more points. So these are",
            "tokens": [
                51076,
                2962,
                13,
                400,
                550,
                286,
                393,
                909,
                512,
                544,
                2793,
                11,
                293,
                286,
                393,
                909,
                512,
                544,
                2793,
                13,
                407,
                613,
                366,
                51320
            ],
            "temperature": 0.0,
            "avg_logprob": -0.2314967046442785,
            "compression_ratio": 1.853061224489796,
            "no_speech_prob": 0.0014781951904296875,
            "words": [
                {
                    "start": 1116.1,
                    "end": 1116.6,
                    "word": " changes",
                    "probability": 0.79150390625
                },
                {
                    "start": 1116.6,
                    "end": 1117.0,
                    "word": ".",
                    "probability": 0.6572265625
                },
                {
                    "start": 1117.18,
                    "end": 1117.38,
                    "word": " And",
                    "probability": 0.85986328125
                },
                {
                    "start": 1117.38,
                    "end": 1117.48,
                    "word": " then",
                    "probability": 0.62890625
                },
                {
                    "start": 1117.48,
                    "end": 1117.56,
                    "word": " I",
                    "probability": 0.99267578125
                },
                {
                    "start": 1117.56,
                    "end": 1117.68,
                    "word": " can",
                    "probability": 0.876953125
                },
                {
                    "start": 1117.68,
                    "end": 1117.84,
                    "word": " add",
                    "probability": 0.79541015625
                },
                {
                    "start": 1117.84,
                    "end": 1117.98,
                    "word": " some",
                    "probability": 0.81005859375
                },
                {
                    "start": 1117.98,
                    "end": 1118.1,
                    "word": " more",
                    "probability": 0.87744140625
                },
                {
                    "start": 1118.1,
                    "end": 1118.38,
                    "word": " points",
                    "probability": 0.779296875
                },
                {
                    "start": 1118.38,
                    "end": 1118.96,
                    "word": ",",
                    "probability": 0.55712890625
                },
                {
                    "start": 1118.96,
                    "end": 1119.2,
                    "word": " and",
                    "probability": 0.87255859375
                },
                {
                    "start": 1119.2,
                    "end": 1119.3,
                    "word": " I",
                    "probability": 0.97265625
                },
                {
                    "start": 1119.3,
                    "end": 1119.4,
                    "word": " can",
                    "probability": 0.884765625
                },
                {
                    "start": 1119.4,
                    "end": 1119.54,
                    "word": " add",
                    "probability": 0.80078125
                },
                {
                    "start": 1119.54,
                    "end": 1119.66,
                    "word": " some",
                    "probability": 0.818359375
                },
                {
                    "start": 1119.66,
                    "end": 1119.82,
                    "word": " more",
                    "probability": 0.87744140625
                },
                {
                    "start": 1119.82,
                    "end": 1120.08,
                    "word": " points",
                    "probability": 0.76806640625
                },
                {
                    "start": 1120.08,
                    "end": 1120.32,
                    "word": ".",
                    "probability": 0.96826171875
                },
                {
                    "start": 1120.32,
                    "end": 1120.6,
                    "word": " So",
                    "probability": 0.8955078125
                },
                {
                    "start": 1120.6,
                    "end": 1120.98,
                    "word": " these",
                    "probability": 0.63037109375
                },
                {
                    "start": 1120.98,
                    "end": 1121.14,
                    "word": " are",
                    "probability": 0.869140625
                }
            ]
        },
        {
            "id": 168,
            "seek": 98604,
            "start": 1121.14,
            "end": 1126.26,
            "text": " all my posterior. And I can update my original thing by taking the path of each of the points",
            "tokens": [
                51320,
                439,
                452,
                33529,
                13,
                400,
                286,
                393,
                5623,
                452,
                3380,
                551,
                538,
                1940,
                264,
                3100,
                295,
                1184,
                295,
                264,
                2793,
                51600
            ],
            "temperature": 0.0,
            "avg_logprob": -0.2314967046442785,
            "compression_ratio": 1.853061224489796,
            "no_speech_prob": 0.0014781951904296875,
            "words": [
                {
                    "start": 1121.14,
                    "end": 1121.4,
                    "word": " all",
                    "probability": 0.89599609375
                },
                {
                    "start": 1121.4,
                    "end": 1121.74,
                    "word": " my",
                    "probability": 0.8779296875
                },
                {
                    "start": 1121.74,
                    "end": 1122.34,
                    "word": " posterior",
                    "probability": 0.57275390625
                },
                {
                    "start": 1122.34,
                    "end": 1122.5,
                    "word": ".",
                    "probability": 0.276123046875
                },
                {
                    "start": 1122.5,
                    "end": 1122.6,
                    "word": " And",
                    "probability": 0.84130859375
                },
                {
                    "start": 1122.6,
                    "end": 1122.72,
                    "word": " I",
                    "probability": 0.99755859375
                },
                {
                    "start": 1122.72,
                    "end": 1122.92,
                    "word": " can",
                    "probability": 0.8564453125
                },
                {
                    "start": 1122.92,
                    "end": 1123.4,
                    "word": " update",
                    "probability": 0.76904296875
                },
                {
                    "start": 1123.4,
                    "end": 1123.64,
                    "word": " my",
                    "probability": 0.8876953125
                },
                {
                    "start": 1123.64,
                    "end": 1123.9,
                    "word": " original",
                    "probability": 0.81787109375
                },
                {
                    "start": 1123.9,
                    "end": 1124.14,
                    "word": " thing",
                    "probability": 0.80810546875
                },
                {
                    "start": 1124.14,
                    "end": 1124.32,
                    "word": " by",
                    "probability": 0.89453125
                },
                {
                    "start": 1124.32,
                    "end": 1124.6,
                    "word": " taking",
                    "probability": 0.80908203125
                },
                {
                    "start": 1124.6,
                    "end": 1124.82,
                    "word": " the",
                    "probability": 0.79833984375
                },
                {
                    "start": 1124.82,
                    "end": 1125.22,
                    "word": " path",
                    "probability": 0.79248046875
                },
                {
                    "start": 1125.22,
                    "end": 1125.62,
                    "word": " of",
                    "probability": 0.9052734375
                },
                {
                    "start": 1125.62,
                    "end": 1125.78,
                    "word": " each",
                    "probability": 0.8564453125
                },
                {
                    "start": 1125.78,
                    "end": 1125.88,
                    "word": " of",
                    "probability": 0.8984375
                },
                {
                    "start": 1125.88,
                    "end": 1125.98,
                    "word": " the",
                    "probability": 0.81298828125
                },
                {
                    "start": 1125.98,
                    "end": 1126.26,
                    "word": " points",
                    "probability": 0.7783203125
                }
            ]
        },
        {
            "id": 169,
            "seek": 101456,
            "start": 1126.26,
            "end": 1133.54,
            "text": " and updating all the little discrete distributions that I have. So this enables me to",
            "tokens": [
                50384,
                293,
                25113,
                439,
                264,
                707,
                27706,
                37870,
                300,
                286,
                362,
                13,
                407,
                341,
                17077,
                385,
                281,
                50704
            ],
            "temperature": 0.0,
            "avg_logprob": -0.18850160562075102,
            "compression_ratio": 1.7871287128712872,
            "no_speech_prob": 0.00572967529296875,
            "words": [
                {
                    "start": 1126.26,
                    "end": 1126.96,
                    "word": " and",
                    "probability": 0.2215576171875
                },
                {
                    "start": 1126.96,
                    "end": 1127.28,
                    "word": " updating",
                    "probability": 0.72314453125
                },
                {
                    "start": 1127.28,
                    "end": 1127.62,
                    "word": " all",
                    "probability": 0.876953125
                },
                {
                    "start": 1127.62,
                    "end": 1127.8,
                    "word": " the",
                    "probability": 0.7841796875
                },
                {
                    "start": 1127.8,
                    "end": 1128.08,
                    "word": " little",
                    "probability": 0.66552734375
                },
                {
                    "start": 1128.08,
                    "end": 1129.08,
                    "word": " discrete",
                    "probability": 0.80078125
                },
                {
                    "start": 1129.08,
                    "end": 1129.62,
                    "word": " distributions",
                    "probability": 0.7900390625
                },
                {
                    "start": 1129.62,
                    "end": 1129.88,
                    "word": " that",
                    "probability": 0.85107421875
                },
                {
                    "start": 1129.88,
                    "end": 1129.98,
                    "word": " I",
                    "probability": 0.96435546875
                },
                {
                    "start": 1129.98,
                    "end": 1130.32,
                    "word": " have",
                    "probability": 0.86962890625
                },
                {
                    "start": 1130.32,
                    "end": 1131.12,
                    "word": ".",
                    "probability": 0.84130859375
                },
                {
                    "start": 1131.32,
                    "end": 1131.94,
                    "word": " So",
                    "probability": 0.78076171875
                },
                {
                    "start": 1131.94,
                    "end": 1132.14,
                    "word": " this",
                    "probability": 0.8115234375
                },
                {
                    "start": 1132.14,
                    "end": 1132.54,
                    "word": " enables",
                    "probability": 0.75439453125
                },
                {
                    "start": 1132.54,
                    "end": 1132.84,
                    "word": " me",
                    "probability": 0.90380859375
                },
                {
                    "start": 1132.84,
                    "end": 1133.54,
                    "word": " to",
                    "probability": 0.90478515625
                }
            ]
        },
        {
            "id": 170,
            "seek": 101456,
            "start": 1134.1,
            "end": 1140.94,
            "text": " sort of quickly make a posterior distribution out of my prior distribution, and everything",
            "tokens": [
                50760,
                1333,
                295,
                2661,
                652,
                257,
                33529,
                7316,
                484,
                295,
                452,
                4059,
                7316,
                11,
                293,
                1203,
                51112
            ],
            "temperature": 0.0,
            "avg_logprob": -0.18850160562075102,
            "compression_ratio": 1.7871287128712872,
            "no_speech_prob": 0.00572967529296875,
            "words": [
                {
                    "start": 1134.1,
                    "end": 1134.5,
                    "word": " sort",
                    "probability": 0.364501953125
                },
                {
                    "start": 1134.5,
                    "end": 1135.24,
                    "word": " of",
                    "probability": 0.9013671875
                },
                {
                    "start": 1135.24,
                    "end": 1136.76,
                    "word": " quickly",
                    "probability": 0.81494140625
                },
                {
                    "start": 1136.76,
                    "end": 1137.76,
                    "word": " make",
                    "probability": 0.849609375
                },
                {
                    "start": 1137.9,
                    "end": 1138.22,
                    "word": " a posterior",
                    "probability": 0.763671875
                },
                {
                    "start": 1138.22,
                    "end": 1138.84,
                    "word": " distribution",
                    "probability": 0.6943359375
                },
                {
                    "start": 1138.84,
                    "end": 1139.46,
                    "word": " out",
                    "probability": 0.8046875
                },
                {
                    "start": 1139.46,
                    "end": 1139.54,
                    "word": " of",
                    "probability": 0.91162109375
                },
                {
                    "start": 1139.54,
                    "end": 1139.66,
                    "word": " my",
                    "probability": 0.9111328125
                },
                {
                    "start": 1139.66,
                    "end": 1139.88,
                    "word": " prior",
                    "probability": 0.82080078125
                },
                {
                    "start": 1139.88,
                    "end": 1140.32,
                    "word": " distribution",
                    "probability": 0.72412109375
                },
                {
                    "start": 1140.32,
                    "end": 1140.52,
                    "word": ",",
                    "probability": 0.2193603515625
                },
                {
                    "start": 1140.52,
                    "end": 1140.56,
                    "word": " and",
                    "probability": 0.51318359375
                },
                {
                    "start": 1140.56,
                    "end": 1140.94,
                    "word": " everything",
                    "probability": 0.8544921875
                }
            ]
        },
        {
            "id": 171,
            "seek": 101456,
            "start": 1140.94,
            "end": 1148.1,
            "text": " is discrete. And then I can take the KL divergence between my posterior discrete distributions and",
            "tokens": [
                51112,
                307,
                27706,
                13,
                400,
                550,
                286,
                393,
                747,
                264,
                47991,
                47387,
                1296,
                452,
                33529,
                27706,
                37870,
                293,
                51448
            ],
            "temperature": 0.0,
            "avg_logprob": -0.18850160562075102,
            "compression_ratio": 1.7871287128712872,
            "no_speech_prob": 0.00572967529296875,
            "words": [
                {
                    "start": 1140.94,
                    "end": 1141.42,
                    "word": " is",
                    "probability": 0.90185546875
                },
                {
                    "start": 1141.42,
                    "end": 1141.92,
                    "word": " discrete",
                    "probability": 0.78125
                },
                {
                    "start": 1141.92,
                    "end": 1142.6,
                    "word": ".",
                    "probability": 0.767578125
                },
                {
                    "start": 1142.6,
                    "end": 1143.0,
                    "word": " And",
                    "probability": 0.81787109375
                },
                {
                    "start": 1143.0,
                    "end": 1143.1,
                    "word": " then",
                    "probability": 0.7314453125
                },
                {
                    "start": 1143.1,
                    "end": 1143.2,
                    "word": " I",
                    "probability": 0.97998046875
                },
                {
                    "start": 1143.2,
                    "end": 1143.32,
                    "word": " can",
                    "probability": 0.88916015625
                },
                {
                    "start": 1143.32,
                    "end": 1143.54,
                    "word": " take",
                    "probability": 0.814453125
                },
                {
                    "start": 1143.54,
                    "end": 1143.7,
                    "word": " the",
                    "probability": 0.822265625
                },
                {
                    "start": 1143.7,
                    "end": 1143.92,
                    "word": " KL",
                    "probability": 0.94384765625
                },
                {
                    "start": 1143.92,
                    "end": 1144.34,
                    "word": " divergence",
                    "probability": 0.7216796875
                },
                {
                    "start": 1144.34,
                    "end": 1144.88,
                    "word": " between",
                    "probability": 0.779296875
                },
                {
                    "start": 1144.88,
                    "end": 1145.28,
                    "word": " my",
                    "probability": 0.9052734375
                },
                {
                    "start": 1145.28,
                    "end": 1145.76,
                    "word": " posterior",
                    "probability": 0.69970703125
                },
                {
                    "start": 1145.76,
                    "end": 1146.9,
                    "word": " discrete",
                    "probability": 0.865234375
                },
                {
                    "start": 1146.9,
                    "end": 1147.44,
                    "word": " distributions",
                    "probability": 0.7958984375
                },
                {
                    "start": 1147.44,
                    "end": 1148.1,
                    "word": " and",
                    "probability": 0.8134765625
                }
            ]
        },
        {
            "id": 172,
            "seek": 101456,
            "start": 1148.1,
            "end": 1154.78,
            "text": " my prior discrete distributions. So in this case, this one has a probability of 0.33,",
            "tokens": [
                51448,
                452,
                4059,
                27706,
                37870,
                13,
                407,
                294,
                341,
                1389,
                11,
                341,
                472,
                575,
                257,
                8482,
                295,
                1958,
                13,
                10191,
                11,
                51796
            ],
            "temperature": 0.0,
            "avg_logprob": -0.18850160562075102,
            "compression_ratio": 1.7871287128712872,
            "no_speech_prob": 0.00572967529296875,
            "words": [
                {
                    "start": 1148.1,
                    "end": 1148.24,
                    "word": " my",
                    "probability": 0.91015625
                },
                {
                    "start": 1148.24,
                    "end": 1148.58,
                    "word": " prior",
                    "probability": 0.8427734375
                },
                {
                    "start": 1148.58,
                    "end": 1149.18,
                    "word": " discrete",
                    "probability": 0.8720703125
                },
                {
                    "start": 1149.18,
                    "end": 1149.62,
                    "word": " distributions",
                    "probability": 0.505859375
                },
                {
                    "start": 1149.62,
                    "end": 1150.14,
                    "word": ".",
                    "probability": 0.96533203125
                },
                {
                    "start": 1150.14,
                    "end": 1150.3,
                    "word": " So",
                    "probability": 0.8798828125
                },
                {
                    "start": 1150.3,
                    "end": 1150.44,
                    "word": " in",
                    "probability": 0.86865234375
                },
                {
                    "start": 1150.44,
                    "end": 1150.6,
                    "word": " this",
                    "probability": 0.87939453125
                },
                {
                    "start": 1150.6,
                    "end": 1150.9,
                    "word": " case",
                    "probability": 0.8017578125
                },
                {
                    "start": 1150.9,
                    "end": 1151.48,
                    "word": ",",
                    "probability": 0.83251953125
                },
                {
                    "start": 1151.48,
                    "end": 1151.92,
                    "word": " this",
                    "probability": 0.87353515625
                },
                {
                    "start": 1151.92,
                    "end": 1152.24,
                    "word": " one",
                    "probability": 0.81201171875
                },
                {
                    "start": 1152.24,
                    "end": 1152.64,
                    "word": " has",
                    "probability": 0.86328125
                },
                {
                    "start": 1152.88,
                    "end": 1153.38,
                    "word": " a probability",
                    "probability": 0.8994140625
                },
                {
                    "start": 1153.38,
                    "end": 1153.88,
                    "word": " of",
                    "probability": 0.90283203125
                },
                {
                    "start": 1153.88,
                    "end": 1154.08,
                    "word": " 0",
                    "probability": 0.8671875
                },
                {
                    "start": 1154.08,
                    "end": 1154.38,
                    "word": ".33",
                    "probability": 0.99658203125
                },
                {
                    "start": 1154.38,
                    "end": 1154.78,
                    "word": ",",
                    "probability": 0.50390625
                }
            ]
        },
        {
            "id": 173,
            "seek": 104200,
            "start": 1154.78,
            "end": 1161.1,
            "text": " and here I have a probability of 0.66. And afterwards, after I've included all my training",
            "tokens": [
                50372,
                293,
                510,
                286,
                362,
                257,
                8482,
                295,
                1958,
                13,
                15237,
                13,
                400,
                10543,
                11,
                934,
                286,
                600,
                5556,
                439,
                452,
                3097,
                50680
            ],
            "temperature": 0.0,
            "avg_logprob": -0.16586538977347887,
            "compression_ratio": 2.088888888888889,
            "no_speech_prob": 0.0014781951904296875,
            "words": [
                {
                    "start": 1154.78,
                    "end": 1155.3,
                    "word": " and",
                    "probability": 0.2437744140625
                },
                {
                    "start": 1155.3,
                    "end": 1155.66,
                    "word": " here",
                    "probability": 0.75634765625
                },
                {
                    "start": 1155.66,
                    "end": 1155.84,
                    "word": " I",
                    "probability": 0.9287109375
                },
                {
                    "start": 1155.84,
                    "end": 1155.96,
                    "word": " have",
                    "probability": 0.85986328125
                },
                {
                    "start": 1156.06,
                    "end": 1156.4,
                    "word": " a probability",
                    "probability": 0.880859375
                },
                {
                    "start": 1156.4,
                    "end": 1156.8,
                    "word": " of",
                    "probability": 0.89453125
                },
                {
                    "start": 1156.8,
                    "end": 1156.96,
                    "word": " 0",
                    "probability": 0.6650390625
                },
                {
                    "start": 1156.96,
                    "end": 1157.36,
                    "word": ".66",
                    "probability": 0.989990234375
                },
                {
                    "start": 1157.36,
                    "end": 1157.8,
                    "word": ".",
                    "probability": 0.1783447265625
                },
                {
                    "start": 1157.92,
                    "end": 1158.08,
                    "word": " And",
                    "probability": 0.44091796875
                },
                {
                    "start": 1158.08,
                    "end": 1158.84,
                    "word": " afterwards",
                    "probability": 0.7421875
                },
                {
                    "start": 1158.84,
                    "end": 1159.58,
                    "word": ",",
                    "probability": 0.78759765625
                },
                {
                    "start": 1159.58,
                    "end": 1159.94,
                    "word": " after",
                    "probability": 0.73828125
                },
                {
                    "start": 1159.94,
                    "end": 1160.16,
                    "word": " I've",
                    "probability": 0.923583984375
                },
                {
                    "start": 1160.16,
                    "end": 1160.48,
                    "word": " included",
                    "probability": 0.7470703125
                },
                {
                    "start": 1160.48,
                    "end": 1160.68,
                    "word": " all",
                    "probability": 0.89013671875
                },
                {
                    "start": 1160.68,
                    "end": 1160.84,
                    "word": " my",
                    "probability": 0.8740234375
                },
                {
                    "start": 1160.84,
                    "end": 1161.1,
                    "word": " training",
                    "probability": 0.8203125
                }
            ]
        },
        {
            "id": 174,
            "seek": 104200,
            "start": 1161.1,
            "end": 1167.06,
            "text": " set, this is a probability of like 0.4. And this, you know, this is a probability of 0.6,",
            "tokens": [
                50680,
                992,
                11,
                341,
                307,
                257,
                8482,
                295,
                411,
                1958,
                13,
                19,
                13,
                400,
                341,
                11,
                291,
                458,
                11,
                341,
                307,
                257,
                8482,
                295,
                1958,
                13,
                21,
                11,
                50972
            ],
            "temperature": 0.0,
            "avg_logprob": -0.16586538977347887,
            "compression_ratio": 2.088888888888889,
            "no_speech_prob": 0.0014781951904296875,
            "words": [
                {
                    "start": 1161.1,
                    "end": 1161.36,
                    "word": " set",
                    "probability": 0.84521484375
                },
                {
                    "start": 1161.36,
                    "end": 1161.56,
                    "word": ",",
                    "probability": 0.76171875
                },
                {
                    "start": 1161.56,
                    "end": 1161.7,
                    "word": " this",
                    "probability": 0.85009765625
                },
                {
                    "start": 1161.7,
                    "end": 1161.84,
                    "word": " is",
                    "probability": 0.896484375
                },
                {
                    "start": 1161.98,
                    "end": 1162.5,
                    "word": " a probability",
                    "probability": 0.89990234375
                },
                {
                    "start": 1162.5,
                    "end": 1162.94,
                    "word": " of",
                    "probability": 0.90576171875
                },
                {
                    "start": 1162.94,
                    "end": 1163.14,
                    "word": " like",
                    "probability": 0.4892578125
                },
                {
                    "start": 1163.14,
                    "end": 1163.56,
                    "word": " 0",
                    "probability": 0.78955078125
                },
                {
                    "start": 1163.56,
                    "end": 1164.42,
                    "word": ".4",
                    "probability": 0.975341796875
                },
                {
                    "start": 1164.42,
                    "end": 1164.66,
                    "word": ".",
                    "probability": 0.1668701171875
                },
                {
                    "start": 1164.66,
                    "end": 1164.66,
                    "word": " And",
                    "probability": 0.60791015625
                },
                {
                    "start": 1164.66,
                    "end": 1164.84,
                    "word": " this",
                    "probability": 0.71728515625
                },
                {
                    "start": 1164.84,
                    "end": 1165.06,
                    "word": ",",
                    "probability": 0.38818359375
                },
                {
                    "start": 1165.06,
                    "end": 1165.3,
                    "word": " you",
                    "probability": 0.77490234375
                },
                {
                    "start": 1165.3,
                    "end": 1165.42,
                    "word": " know",
                    "probability": 0.81298828125
                },
                {
                    "start": 1165.42,
                    "end": 1165.48,
                    "word": ",",
                    "probability": 0.9326171875
                },
                {
                    "start": 1165.48,
                    "end": 1165.66,
                    "word": " this",
                    "probability": 0.87060546875
                },
                {
                    "start": 1165.66,
                    "end": 1165.78,
                    "word": " is",
                    "probability": 0.90283203125
                },
                {
                    "start": 1165.9,
                    "end": 1166.24,
                    "word": " a probability",
                    "probability": 0.9033203125
                },
                {
                    "start": 1166.24,
                    "end": 1166.5,
                    "word": " of",
                    "probability": 0.89599609375
                },
                {
                    "start": 1166.5,
                    "end": 1166.64,
                    "word": " 0",
                    "probability": 0.982421875
                },
                {
                    "start": 1166.64,
                    "end": 1166.98,
                    "word": ".6",
                    "probability": 0.995361328125
                },
                {
                    "start": 1166.98,
                    "end": 1167.06,
                    "word": ",",
                    "probability": 0.337890625
                }
            ]
        },
        {
            "id": 175,
            "seek": 104200,
            "start": 1167.06,
            "end": 1171.98,
            "text": " and this is a probability of 0.4. And those are very different. And I can plug those into the",
            "tokens": [
                50972,
                293,
                341,
                307,
                257,
                8482,
                295,
                1958,
                13,
                19,
                13,
                400,
                729,
                366,
                588,
                819,
                13,
                400,
                286,
                393,
                5452,
                729,
                666,
                264,
                51216
            ],
            "temperature": 0.0,
            "avg_logprob": -0.16586538977347887,
            "compression_ratio": 2.088888888888889,
            "no_speech_prob": 0.0014781951904296875,
            "words": [
                {
                    "start": 1167.06,
                    "end": 1167.16,
                    "word": " and",
                    "probability": 0.8818359375
                },
                {
                    "start": 1167.16,
                    "end": 1167.3,
                    "word": " this",
                    "probability": 0.8798828125
                },
                {
                    "start": 1167.3,
                    "end": 1167.42,
                    "word": " is",
                    "probability": 0.900390625
                },
                {
                    "start": 1167.56,
                    "end": 1167.96,
                    "word": " a probability",
                    "probability": 0.91748046875
                },
                {
                    "start": 1167.96,
                    "end": 1168.22,
                    "word": " of",
                    "probability": 0.90283203125
                },
                {
                    "start": 1168.22,
                    "end": 1168.36,
                    "word": " 0",
                    "probability": 0.990234375
                },
                {
                    "start": 1168.36,
                    "end": 1168.72,
                    "word": ".4",
                    "probability": 0.998291015625
                },
                {
                    "start": 1168.72,
                    "end": 1169.02,
                    "word": ".",
                    "probability": 0.90234375
                },
                {
                    "start": 1169.02,
                    "end": 1169.28,
                    "word": " And",
                    "probability": 0.84814453125
                },
                {
                    "start": 1169.28,
                    "end": 1169.42,
                    "word": " those",
                    "probability": 0.5732421875
                },
                {
                    "start": 1169.42,
                    "end": 1170.0,
                    "word": " are",
                    "probability": 0.88134765625
                },
                {
                    "start": 1170.0,
                    "end": 1170.2,
                    "word": " very",
                    "probability": 0.734375
                },
                {
                    "start": 1170.2,
                    "end": 1170.52,
                    "word": " different",
                    "probability": 0.68798828125
                },
                {
                    "start": 1170.52,
                    "end": 1170.9,
                    "word": ".",
                    "probability": 0.5869140625
                },
                {
                    "start": 1170.9,
                    "end": 1171.06,
                    "word": " And",
                    "probability": 0.87255859375
                },
                {
                    "start": 1171.06,
                    "end": 1171.16,
                    "word": " I",
                    "probability": 0.99658203125
                },
                {
                    "start": 1171.16,
                    "end": 1171.26,
                    "word": " can",
                    "probability": 0.8896484375
                },
                {
                    "start": 1171.26,
                    "end": 1171.48,
                    "word": " plug",
                    "probability": 0.8466796875
                },
                {
                    "start": 1171.48,
                    "end": 1171.66,
                    "word": " those",
                    "probability": 0.85498046875
                },
                {
                    "start": 1171.66,
                    "end": 1171.82,
                    "word": " into",
                    "probability": 0.7412109375
                },
                {
                    "start": 1171.82,
                    "end": 1171.98,
                    "word": " the",
                    "probability": 0.80419921875
                }
            ]
        },
        {
            "id": 176,
            "seek": 104200,
            "start": 1171.98,
            "end": 1176.9,
            "text": " KL divergence thing and get a positive number for the KL divergence at this node. And then I can do",
            "tokens": [
                51216,
                47991,
                47387,
                551,
                293,
                483,
                257,
                3353,
                1230,
                337,
                264,
                47991,
                47387,
                412,
                341,
                9984,
                13,
                400,
                550,
                286,
                393,
                360,
                51464
            ],
            "temperature": 0.0,
            "avg_logprob": -0.16586538977347887,
            "compression_ratio": 2.088888888888889,
            "no_speech_prob": 0.0014781951904296875,
            "words": [
                {
                    "start": 1171.98,
                    "end": 1172.14,
                    "word": " KL",
                    "probability": 0.84033203125
                },
                {
                    "start": 1172.14,
                    "end": 1172.52,
                    "word": " divergence",
                    "probability": 0.38525390625
                },
                {
                    "start": 1172.52,
                    "end": 1172.98,
                    "word": " thing",
                    "probability": 0.77001953125
                },
                {
                    "start": 1172.98,
                    "end": 1173.14,
                    "word": " and",
                    "probability": 0.7421875
                },
                {
                    "start": 1173.14,
                    "end": 1173.28,
                    "word": " get",
                    "probability": 0.89990234375
                },
                {
                    "start": 1173.42,
                    "end": 1173.68,
                    "word": " a positive",
                    "probability": 0.87109375
                },
                {
                    "start": 1173.68,
                    "end": 1174.04,
                    "word": " number",
                    "probability": 0.8779296875
                },
                {
                    "start": 1174.04,
                    "end": 1174.48,
                    "word": " for",
                    "probability": 0.82666015625
                },
                {
                    "start": 1174.48,
                    "end": 1174.6,
                    "word": " the",
                    "probability": 0.81982421875
                },
                {
                    "start": 1174.6,
                    "end": 1174.78,
                    "word": " KL",
                    "probability": 0.94091796875
                },
                {
                    "start": 1174.78,
                    "end": 1175.14,
                    "word": " divergence",
                    "probability": 0.76123046875
                },
                {
                    "start": 1175.14,
                    "end": 1175.52,
                    "word": " at",
                    "probability": 0.90283203125
                },
                {
                    "start": 1175.52,
                    "end": 1175.78,
                    "word": " this",
                    "probability": 0.88037109375
                },
                {
                    "start": 1175.78,
                    "end": 1176.04,
                    "word": " node",
                    "probability": 0.80078125
                },
                {
                    "start": 1176.04,
                    "end": 1176.28,
                    "word": ".",
                    "probability": 0.63623046875
                },
                {
                    "start": 1176.28,
                    "end": 1176.56,
                    "word": " And",
                    "probability": 0.8564453125
                },
                {
                    "start": 1176.56,
                    "end": 1176.64,
                    "word": " then",
                    "probability": 0.63818359375
                },
                {
                    "start": 1176.64,
                    "end": 1176.72,
                    "word": " I",
                    "probability": 0.9873046875
                },
                {
                    "start": 1176.72,
                    "end": 1176.8,
                    "word": " can",
                    "probability": 0.75146484375
                },
                {
                    "start": 1176.8,
                    "end": 1176.9,
                    "word": " do",
                    "probability": 0.9140625
                }
            ]
        },
        {
            "id": 177,
            "seek": 104200,
            "start": 1176.9,
            "end": 1182.22,
            "text": " it for this node, and this node, and this node, and get all the KL divergences and add them up.",
            "tokens": [
                51464,
                309,
                337,
                341,
                9984,
                11,
                293,
                341,
                9984,
                11,
                293,
                341,
                9984,
                11,
                293,
                483,
                439,
                264,
                47991,
                18558,
                1766,
                887,
                293,
                909,
                552,
                493,
                13,
                51732
            ],
            "temperature": 0.0,
            "avg_logprob": -0.16586538977347887,
            "compression_ratio": 2.088888888888889,
            "no_speech_prob": 0.0014781951904296875,
            "words": [
                {
                    "start": 1176.9,
                    "end": 1177.02,
                    "word": " it",
                    "probability": 0.900390625
                },
                {
                    "start": 1177.02,
                    "end": 1177.14,
                    "word": " for",
                    "probability": 0.8564453125
                },
                {
                    "start": 1177.14,
                    "end": 1177.36,
                    "word": " this",
                    "probability": 0.8837890625
                },
                {
                    "start": 1177.36,
                    "end": 1177.64,
                    "word": " node",
                    "probability": 0.8037109375
                },
                {
                    "start": 1177.64,
                    "end": 1177.94,
                    "word": ",",
                    "probability": 0.533203125
                },
                {
                    "start": 1177.94,
                    "end": 1178.22,
                    "word": " and",
                    "probability": 0.88720703125
                },
                {
                    "start": 1178.22,
                    "end": 1178.44,
                    "word": " this",
                    "probability": 0.873046875
                },
                {
                    "start": 1178.44,
                    "end": 1178.74,
                    "word": " node",
                    "probability": 0.80322265625
                },
                {
                    "start": 1178.74,
                    "end": 1178.82,
                    "word": ",",
                    "probability": 0.86376953125
                },
                {
                    "start": 1178.82,
                    "end": 1179.0,
                    "word": " and",
                    "probability": 0.89111328125
                },
                {
                    "start": 1179.0,
                    "end": 1179.18,
                    "word": " this",
                    "probability": 0.87890625
                },
                {
                    "start": 1179.18,
                    "end": 1179.4,
                    "word": " node",
                    "probability": 0.8046875
                },
                {
                    "start": 1179.4,
                    "end": 1179.42,
                    "word": ",",
                    "probability": 0.6748046875
                },
                {
                    "start": 1179.42,
                    "end": 1179.54,
                    "word": " and",
                    "probability": 0.890625
                },
                {
                    "start": 1179.54,
                    "end": 1179.7,
                    "word": " get",
                    "probability": 0.896484375
                },
                {
                    "start": 1179.7,
                    "end": 1180.08,
                    "word": " all",
                    "probability": 0.8994140625
                },
                {
                    "start": 1180.08,
                    "end": 1180.24,
                    "word": " the",
                    "probability": 0.69873046875
                },
                {
                    "start": 1180.24,
                    "end": 1180.48,
                    "word": " KL",
                    "probability": 0.96435546875
                },
                {
                    "start": 1180.48,
                    "end": 1180.98,
                    "word": " divergences",
                    "probability": 0.8673502604166666
                },
                {
                    "start": 1180.98,
                    "end": 1181.16,
                    "word": " and",
                    "probability": 0.802734375
                },
                {
                    "start": 1181.16,
                    "end": 1181.42,
                    "word": " add",
                    "probability": 0.78564453125
                },
                {
                    "start": 1181.42,
                    "end": 1181.56,
                    "word": " them",
                    "probability": 0.73583984375
                },
                {
                    "start": 1181.56,
                    "end": 1181.78,
                    "word": " up",
                    "probability": 0.91455078125
                },
                {
                    "start": 1181.78,
                    "end": 1182.22,
                    "word": ".",
                    "probability": 0.984375
                }
            ]
        },
        {
            "id": 178,
            "seek": 106474,
            "start": 1182.22,
            "end": 1186.42,
            "text": " And then I can get the total KL divergence with respect to my original tree",
            "tokens": [
                50364,
                400,
                550,
                286,
                393,
                483,
                264,
                3217,
                47991,
                47387,
                365,
                3104,
                281,
                452,
                3380,
                4230,
                50572
            ],
            "temperature": 0.0,
            "avg_logprob": -0.22623240108221349,
            "compression_ratio": 1.641711229946524,
            "no_speech_prob": 0.0010814666748046875,
            "words": [
                {
                    "start": 1182.22,
                    "end": 1182.46,
                    "word": " And",
                    "probability": 0.239990234375
                },
                {
                    "start": 1182.46,
                    "end": 1182.56,
                    "word": " then",
                    "probability": 0.69384765625
                },
                {
                    "start": 1182.56,
                    "end": 1182.64,
                    "word": " I",
                    "probability": 0.96923828125
                },
                {
                    "start": 1182.64,
                    "end": 1182.74,
                    "word": " can",
                    "probability": 0.8916015625
                },
                {
                    "start": 1182.74,
                    "end": 1182.9,
                    "word": " get",
                    "probability": 0.89794921875
                },
                {
                    "start": 1182.9,
                    "end": 1183.06,
                    "word": " the",
                    "probability": 0.7978515625
                },
                {
                    "start": 1183.06,
                    "end": 1183.32,
                    "word": " total",
                    "probability": 0.73095703125
                },
                {
                    "start": 1183.32,
                    "end": 1183.6,
                    "word": " KL",
                    "probability": 0.85546875
                },
                {
                    "start": 1183.6,
                    "end": 1184.74,
                    "word": " divergence",
                    "probability": 0.47607421875
                },
                {
                    "start": 1184.74,
                    "end": 1185.12,
                    "word": " with",
                    "probability": 0.73876953125
                },
                {
                    "start": 1185.12,
                    "end": 1185.36,
                    "word": " respect",
                    "probability": 0.76416015625
                },
                {
                    "start": 1185.36,
                    "end": 1185.54,
                    "word": " to",
                    "probability": 0.90234375
                },
                {
                    "start": 1185.54,
                    "end": 1185.72,
                    "word": " my",
                    "probability": 0.8955078125
                },
                {
                    "start": 1185.72,
                    "end": 1186.02,
                    "word": " original",
                    "probability": 0.8251953125
                },
                {
                    "start": 1186.02,
                    "end": 1186.42,
                    "word": " tree",
                    "probability": 0.75634765625
                }
            ]
        },
        {
            "id": 179,
            "seek": 106474,
            "start": 1186.94,
            "end": 1190.86,
            "text": " of the posterior distribution with respect to the prior distribution,",
            "tokens": [
                50600,
                295,
                264,
                33529,
                7316,
                365,
                3104,
                281,
                264,
                4059,
                7316,
                11,
                50784
            ],
            "temperature": 0.0,
            "avg_logprob": -0.22623240108221349,
            "compression_ratio": 1.641711229946524,
            "no_speech_prob": 0.0010814666748046875,
            "words": [
                {
                    "start": 1186.94,
                    "end": 1187.28,
                    "word": " of",
                    "probability": 0.83349609375
                },
                {
                    "start": 1187.28,
                    "end": 1187.8,
                    "word": " the",
                    "probability": 0.8115234375
                },
                {
                    "start": 1187.8,
                    "end": 1188.3,
                    "word": " posterior",
                    "probability": 0.75
                },
                {
                    "start": 1188.3,
                    "end": 1188.86,
                    "word": " distribution",
                    "probability": 0.73974609375
                },
                {
                    "start": 1188.86,
                    "end": 1189.26,
                    "word": " with",
                    "probability": 0.76806640625
                },
                {
                    "start": 1189.26,
                    "end": 1189.56,
                    "word": " respect",
                    "probability": 0.7626953125
                },
                {
                    "start": 1189.56,
                    "end": 1189.72,
                    "word": " to",
                    "probability": 0.88916015625
                },
                {
                    "start": 1189.72,
                    "end": 1189.84,
                    "word": " the",
                    "probability": 0.8125
                },
                {
                    "start": 1189.84,
                    "end": 1190.04,
                    "word": " prior",
                    "probability": 0.84521484375
                },
                {
                    "start": 1190.04,
                    "end": 1190.58,
                    "word": " distribution",
                    "probability": 0.73681640625
                },
                {
                    "start": 1190.58,
                    "end": 1190.86,
                    "word": ",",
                    "probability": 0.2490234375
                }
            ]
        },
        {
            "id": 180,
            "seek": 106474,
            "start": 1192.56,
            "end": 1200.52,
            "text": " which is really useful. And so this basically, yeah, does it. And the code for doing this",
            "tokens": [
                50896,
                597,
                307,
                534,
                4420,
                13,
                400,
                370,
                341,
                1936,
                11,
                1338,
                11,
                775,
                309,
                13,
                400,
                264,
                3089,
                337,
                884,
                341,
                51264
            ],
            "temperature": 0.0,
            "avg_logprob": -0.22623240108221349,
            "compression_ratio": 1.641711229946524,
            "no_speech_prob": 0.0010814666748046875,
            "words": [
                {
                    "start": 1192.56,
                    "end": 1193.12,
                    "word": " which",
                    "probability": 0.90869140625
                },
                {
                    "start": 1193.12,
                    "end": 1193.28,
                    "word": " is",
                    "probability": 0.90380859375
                },
                {
                    "start": 1193.28,
                    "end": 1193.54,
                    "word": " really",
                    "probability": 0.78271484375
                },
                {
                    "start": 1193.54,
                    "end": 1194.16,
                    "word": " useful",
                    "probability": 0.77734375
                },
                {
                    "start": 1194.16,
                    "end": 1194.68,
                    "word": ".",
                    "probability": 0.77490234375
                },
                {
                    "start": 1194.68,
                    "end": 1194.68,
                    "word": " And",
                    "probability": 0.6884765625
                },
                {
                    "start": 1194.68,
                    "end": 1194.78,
                    "word": " so",
                    "probability": 0.85009765625
                },
                {
                    "start": 1194.78,
                    "end": 1194.98,
                    "word": " this",
                    "probability": 0.82861328125
                },
                {
                    "start": 1194.98,
                    "end": 1195.4,
                    "word": " basically",
                    "probability": 0.8857421875
                },
                {
                    "start": 1195.4,
                    "end": 1195.68,
                    "word": ",",
                    "probability": 0.56103515625
                },
                {
                    "start": 1195.68,
                    "end": 1195.96,
                    "word": " yeah",
                    "probability": 0.86962890625
                },
                {
                    "start": 1195.96,
                    "end": 1196.42,
                    "word": ",",
                    "probability": 0.86962890625
                },
                {
                    "start": 1196.42,
                    "end": 1196.66,
                    "word": " does",
                    "probability": 0.459228515625
                },
                {
                    "start": 1196.66,
                    "end": 1196.84,
                    "word": " it",
                    "probability": 0.88525390625
                },
                {
                    "start": 1196.84,
                    "end": 1197.4,
                    "word": ".",
                    "probability": 0.859375
                },
                {
                    "start": 1197.7,
                    "end": 1197.9,
                    "word": " And",
                    "probability": 0.671875
                },
                {
                    "start": 1197.9,
                    "end": 1198.18,
                    "word": " the",
                    "probability": 0.353271484375
                },
                {
                    "start": 1198.18,
                    "end": 1199.74,
                    "word": " code",
                    "probability": 0.828125
                },
                {
                    "start": 1199.74,
                    "end": 1199.92,
                    "word": " for",
                    "probability": 0.90087890625
                },
                {
                    "start": 1199.92,
                    "end": 1200.1,
                    "word": " doing",
                    "probability": 0.91455078125
                },
                {
                    "start": 1200.1,
                    "end": 1200.52,
                    "word": " this",
                    "probability": 0.87939453125
                }
            ]
        },
        {
            "id": 181,
            "seek": 106474,
            "start": 1201.12,
            "end": 1204.96,
            "text": " enables you to do it very quickly, because it's basically just counting",
            "tokens": [
                51316,
                17077,
                291,
                281,
                360,
                309,
                588,
                2661,
                11,
                570,
                309,
                311,
                1936,
                445,
                13251,
                51500
            ],
            "temperature": 0.0,
            "avg_logprob": -0.22623240108221349,
            "compression_ratio": 1.641711229946524,
            "no_speech_prob": 0.0010814666748046875,
            "words": [
                {
                    "start": 1201.12,
                    "end": 1201.68,
                    "word": " enables",
                    "probability": 0.70751953125
                },
                {
                    "start": 1201.68,
                    "end": 1201.84,
                    "word": " you",
                    "probability": 0.896484375
                },
                {
                    "start": 1201.84,
                    "end": 1201.96,
                    "word": " to",
                    "probability": 0.9091796875
                },
                {
                    "start": 1201.96,
                    "end": 1202.08,
                    "word": " do",
                    "probability": 0.912109375
                },
                {
                    "start": 1202.08,
                    "end": 1202.4,
                    "word": " it",
                    "probability": 0.90380859375
                },
                {
                    "start": 1202.4,
                    "end": 1203.18,
                    "word": " very",
                    "probability": 0.72314453125
                },
                {
                    "start": 1203.18,
                    "end": 1203.48,
                    "word": " quickly",
                    "probability": 0.8828125
                },
                {
                    "start": 1203.48,
                    "end": 1203.66,
                    "word": ",",
                    "probability": 0.346435546875
                },
                {
                    "start": 1203.66,
                    "end": 1203.9,
                    "word": " because",
                    "probability": 0.83251953125
                },
                {
                    "start": 1203.9,
                    "end": 1204.06,
                    "word": " it's",
                    "probability": 0.949951171875
                },
                {
                    "start": 1204.06,
                    "end": 1204.34,
                    "word": " basically",
                    "probability": 0.89501953125
                },
                {
                    "start": 1204.34,
                    "end": 1204.6,
                    "word": " just",
                    "probability": 0.8876953125
                },
                {
                    "start": 1204.6,
                    "end": 1204.96,
                    "word": " counting",
                    "probability": 0.7587890625
                }
            ]
        },
        {
            "id": 182,
            "seek": 109066,
            "start": 1206.24,
            "end": 1215.68,
            "text": " a couple logs and some other statistical equations. And the slide just covers how you do it.",
            "tokens": [
                50392,
                257,
                1916,
                20820,
                293,
                512,
                661,
                22820,
                11787,
                13,
                400,
                264,
                4137,
                445,
                10538,
                577,
                291,
                360,
                309,
                13,
                50856
            ],
            "temperature": 0.0,
            "avg_logprob": -0.2511660474449841,
            "compression_ratio": 1.5135135135135136,
            "no_speech_prob": 0.00046563148498535156,
            "words": [
                {
                    "start": 1206.24,
                    "end": 1206.46,
                    "word": " a couple",
                    "probability": 0.7255859375
                },
                {
                    "start": 1206.46,
                    "end": 1206.86,
                    "word": " logs",
                    "probability": 0.352294921875
                },
                {
                    "start": 1206.86,
                    "end": 1207.2,
                    "word": " and",
                    "probability": 0.7568359375
                },
                {
                    "start": 1207.2,
                    "end": 1207.36,
                    "word": " some",
                    "probability": 0.82421875
                },
                {
                    "start": 1207.36,
                    "end": 1207.66,
                    "word": " other",
                    "probability": 0.78662109375
                },
                {
                    "start": 1207.66,
                    "end": 1208.82,
                    "word": " statistical",
                    "probability": 0.6396484375
                },
                {
                    "start": 1208.82,
                    "end": 1209.58,
                    "word": " equations",
                    "probability": 0.81201171875
                },
                {
                    "start": 1209.58,
                    "end": 1210.22,
                    "word": ".",
                    "probability": 0.60498046875
                },
                {
                    "start": 1211.06,
                    "end": 1211.7,
                    "word": " And",
                    "probability": 0.611328125
                },
                {
                    "start": 1211.7,
                    "end": 1213.3,
                    "word": " the",
                    "probability": 0.30517578125
                },
                {
                    "start": 1213.3,
                    "end": 1213.58,
                    "word": " slide",
                    "probability": 0.86767578125
                },
                {
                    "start": 1213.58,
                    "end": 1213.76,
                    "word": " just",
                    "probability": 0.88916015625
                },
                {
                    "start": 1213.76,
                    "end": 1214.02,
                    "word": " covers",
                    "probability": 0.7919921875
                },
                {
                    "start": 1214.02,
                    "end": 1214.42,
                    "word": " how",
                    "probability": 0.86572265625
                },
                {
                    "start": 1214.42,
                    "end": 1214.66,
                    "word": " you",
                    "probability": 0.5302734375
                },
                {
                    "start": 1214.66,
                    "end": 1214.78,
                    "word": " do",
                    "probability": 0.904296875
                },
                {
                    "start": 1214.78,
                    "end": 1215.04,
                    "word": " it",
                    "probability": 0.90087890625
                },
                {
                    "start": 1215.04,
                    "end": 1215.68,
                    "word": ".",
                    "probability": 0.958984375
                }
            ]
        },
        {
            "id": 183,
            "seek": 109066,
            "start": 1216.26,
            "end": 1224.5,
            "text": " So let's go back to that original test set. So I've got basically a blown up version of",
            "tokens": [
                50932,
                407,
                718,
                311,
                352,
                646,
                281,
                300,
                3380,
                1500,
                992,
                13,
                407,
                286,
                600,
                658,
                1936,
                257,
                16479,
                493,
                3037,
                295,
                51336
            ],
            "temperature": 0.0,
            "avg_logprob": -0.2511660474449841,
            "compression_ratio": 1.5135135135135136,
            "no_speech_prob": 0.00046563148498535156,
            "words": [
                {
                    "start": 1216.26,
                    "end": 1216.62,
                    "word": " So",
                    "probability": 0.865234375
                },
                {
                    "start": 1216.62,
                    "end": 1217.66,
                    "word": " let's",
                    "probability": 0.783447265625
                },
                {
                    "start": 1217.66,
                    "end": 1217.82,
                    "word": " go",
                    "probability": 0.900390625
                },
                {
                    "start": 1217.82,
                    "end": 1218.16,
                    "word": " back",
                    "probability": 0.81298828125
                },
                {
                    "start": 1218.16,
                    "end": 1218.76,
                    "word": " to",
                    "probability": 0.90380859375
                },
                {
                    "start": 1218.76,
                    "end": 1219.32,
                    "word": " that",
                    "probability": 0.88037109375
                },
                {
                    "start": 1219.32,
                    "end": 1219.8,
                    "word": " original",
                    "probability": 0.82177734375
                },
                {
                    "start": 1219.8,
                    "end": 1220.76,
                    "word": " test",
                    "probability": 0.810546875
                },
                {
                    "start": 1220.76,
                    "end": 1221.04,
                    "word": " set",
                    "probability": 0.8759765625
                },
                {
                    "start": 1221.04,
                    "end": 1221.48,
                    "word": ".",
                    "probability": 0.919921875
                },
                {
                    "start": 1221.48,
                    "end": 1221.9,
                    "word": " So",
                    "probability": 0.82421875
                },
                {
                    "start": 1221.9,
                    "end": 1222.32,
                    "word": " I've",
                    "probability": 0.92333984375
                },
                {
                    "start": 1222.32,
                    "end": 1222.54,
                    "word": " got",
                    "probability": 0.9072265625
                },
                {
                    "start": 1222.54,
                    "end": 1223.1,
                    "word": " basically",
                    "probability": 0.44921875
                },
                {
                    "start": 1223.38,
                    "end": 1223.68,
                    "word": " a blown",
                    "probability": 0.8271484375
                },
                {
                    "start": 1223.68,
                    "end": 1223.84,
                    "word": " up",
                    "probability": 0.5576171875
                },
                {
                    "start": 1223.84,
                    "end": 1224.22,
                    "word": " version",
                    "probability": 0.7451171875
                },
                {
                    "start": 1224.22,
                    "end": 1224.5,
                    "word": " of",
                    "probability": 0.90087890625
                }
            ]
        },
        {
            "id": 184,
            "seek": 109066,
            "start": 1225.02,
            "end": 1230.88,
            "text": " those slides of the Gaussian distributions where I had them, where a large number of points from my",
            "tokens": [
                51368,
                729,
                9788,
                295,
                264,
                39148,
                37870,
                689,
                286,
                632,
                552,
                11,
                689,
                257,
                2416,
                1230,
                295,
                2793,
                490,
                452,
                51648
            ],
            "temperature": 0.0,
            "avg_logprob": -0.2511660474449841,
            "compression_ratio": 1.5135135135135136,
            "no_speech_prob": 0.00046563148498535156,
            "words": [
                {
                    "start": 1225.02,
                    "end": 1225.34,
                    "word": " those",
                    "probability": 0.53369140625
                },
                {
                    "start": 1225.34,
                    "end": 1225.8,
                    "word": " slides",
                    "probability": 0.9248046875
                },
                {
                    "start": 1225.8,
                    "end": 1226.08,
                    "word": " of",
                    "probability": 0.85693359375
                },
                {
                    "start": 1226.08,
                    "end": 1226.4,
                    "word": " the",
                    "probability": 0.8076171875
                },
                {
                    "start": 1226.4,
                    "end": 1226.7,
                    "word": " Gaussian",
                    "probability": 0.78955078125
                },
                {
                    "start": 1226.7,
                    "end": 1227.22,
                    "word": " distributions",
                    "probability": 0.7705078125
                },
                {
                    "start": 1227.22,
                    "end": 1227.46,
                    "word": " where",
                    "probability": 0.7529296875
                },
                {
                    "start": 1227.46,
                    "end": 1227.56,
                    "word": " I",
                    "probability": 0.97900390625
                },
                {
                    "start": 1227.56,
                    "end": 1227.76,
                    "word": " had",
                    "probability": 0.82080078125
                },
                {
                    "start": 1227.76,
                    "end": 1227.96,
                    "word": " them",
                    "probability": 0.7236328125
                },
                {
                    "start": 1227.96,
                    "end": 1228.28,
                    "word": ",",
                    "probability": 0.316162109375
                },
                {
                    "start": 1228.28,
                    "end": 1228.52,
                    "word": " where",
                    "probability": 0.7958984375
                },
                {
                    "start": 1228.84,
                    "end": 1229.1,
                    "word": " a large",
                    "probability": 0.92138671875
                },
                {
                    "start": 1229.1,
                    "end": 1229.36,
                    "word": " number",
                    "probability": 0.8837890625
                },
                {
                    "start": 1229.36,
                    "end": 1229.6,
                    "word": " of",
                    "probability": 0.90625
                },
                {
                    "start": 1229.6,
                    "end": 1230.06,
                    "word": " points",
                    "probability": 0.79833984375
                },
                {
                    "start": 1230.06,
                    "end": 1230.46,
                    "word": " from",
                    "probability": 0.8271484375
                },
                {
                    "start": 1230.46,
                    "end": 1230.88,
                    "word": " my",
                    "probability": 0.91650390625
                }
            ]
        },
        {
            "id": 185,
            "seek": 111640,
            "start": 1230.88,
            "end": 1236.28,
            "text": " training distribution, which was a single Gaussian sampled at the center of the thing.",
            "tokens": [
                50388,
                3097,
                7316,
                11,
                597,
                390,
                257,
                2167,
                39148,
                3247,
                15551,
                412,
                264,
                3056,
                295,
                264,
                551,
                13,
                50632
            ],
            "temperature": 0.0,
            "avg_logprob": -0.2151162809757299,
            "compression_ratio": 1.78,
            "no_speech_prob": 0.0010318756103515625,
            "words": [
                {
                    "start": 1230.88,
                    "end": 1231.72,
                    "word": " training",
                    "probability": 0.5595703125
                },
                {
                    "start": 1231.72,
                    "end": 1232.22,
                    "word": " distribution",
                    "probability": 0.64453125
                },
                {
                    "start": 1232.22,
                    "end": 1232.68,
                    "word": ",",
                    "probability": 0.5126953125
                },
                {
                    "start": 1232.68,
                    "end": 1232.82,
                    "word": " which",
                    "probability": 0.9013671875
                },
                {
                    "start": 1232.82,
                    "end": 1232.94,
                    "word": " was",
                    "probability": 0.7880859375
                },
                {
                    "start": 1233.08,
                    "end": 1233.3,
                    "word": " a single",
                    "probability": 0.86669921875
                },
                {
                    "start": 1233.3,
                    "end": 1233.64,
                    "word": " Gaussian",
                    "probability": 0.91943359375
                },
                {
                    "start": 1233.64,
                    "end": 1234.32,
                    "word": " sampled",
                    "probability": 0.725341796875
                },
                {
                    "start": 1234.32,
                    "end": 1234.52,
                    "word": " at",
                    "probability": 0.77587890625
                },
                {
                    "start": 1234.52,
                    "end": 1234.9,
                    "word": " the",
                    "probability": 0.787109375
                },
                {
                    "start": 1234.9,
                    "end": 1235.2,
                    "word": " center",
                    "probability": 0.677734375
                },
                {
                    "start": 1235.2,
                    "end": 1235.54,
                    "word": " of",
                    "probability": 0.72021484375
                },
                {
                    "start": 1235.54,
                    "end": 1235.74,
                    "word": " the",
                    "probability": 0.74658203125
                },
                {
                    "start": 1235.74,
                    "end": 1235.92,
                    "word": " thing",
                    "probability": 0.80322265625
                },
                {
                    "start": 1235.92,
                    "end": 1236.28,
                    "word": ".",
                    "probability": 0.431640625
                }
            ]
        },
        {
            "id": 186,
            "seek": 111640,
            "start": 1236.28,
            "end": 1242.68,
            "text": " And then I have a small number of points from a test distribution. And in this case, I have",
            "tokens": [
                50632,
                400,
                550,
                286,
                362,
                257,
                1359,
                1230,
                295,
                2793,
                490,
                257,
                1500,
                7316,
                13,
                400,
                294,
                341,
                1389,
                11,
                286,
                362,
                50948
            ],
            "temperature": 0.0,
            "avg_logprob": -0.2151162809757299,
            "compression_ratio": 1.78,
            "no_speech_prob": 0.0010318756103515625,
            "words": [
                {
                    "start": 1236.28,
                    "end": 1236.52,
                    "word": " And",
                    "probability": 0.82177734375
                },
                {
                    "start": 1236.52,
                    "end": 1236.62,
                    "word": " then",
                    "probability": 0.71826171875
                },
                {
                    "start": 1236.62,
                    "end": 1236.7,
                    "word": " I",
                    "probability": 0.95703125
                },
                {
                    "start": 1236.7,
                    "end": 1236.76,
                    "word": " have",
                    "probability": 0.86865234375
                },
                {
                    "start": 1236.86,
                    "end": 1237.02,
                    "word": " a small",
                    "probability": 0.84765625
                },
                {
                    "start": 1237.02,
                    "end": 1237.22,
                    "word": " number",
                    "probability": 0.88671875
                },
                {
                    "start": 1237.22,
                    "end": 1237.4,
                    "word": " of",
                    "probability": 0.90283203125
                },
                {
                    "start": 1237.4,
                    "end": 1237.68,
                    "word": " points",
                    "probability": 0.7841796875
                },
                {
                    "start": 1237.68,
                    "end": 1238.72,
                    "word": " from",
                    "probability": 0.75732421875
                },
                {
                    "start": 1239.02,
                    "end": 1239.26,
                    "word": " a test",
                    "probability": 0.8271484375
                },
                {
                    "start": 1239.26,
                    "end": 1239.8,
                    "word": " distribution",
                    "probability": 0.75048828125
                },
                {
                    "start": 1239.8,
                    "end": 1240.44,
                    "word": ".",
                    "probability": 0.91943359375
                },
                {
                    "start": 1241.12,
                    "end": 1241.76,
                    "word": " And",
                    "probability": 0.78857421875
                },
                {
                    "start": 1241.76,
                    "end": 1241.96,
                    "word": " in",
                    "probability": 0.8759765625
                },
                {
                    "start": 1241.96,
                    "end": 1242.1,
                    "word": " this",
                    "probability": 0.88037109375
                },
                {
                    "start": 1242.1,
                    "end": 1242.3,
                    "word": " case",
                    "probability": 0.8095703125
                },
                {
                    "start": 1242.3,
                    "end": 1242.38,
                    "word": ",",
                    "probability": 0.7705078125
                },
                {
                    "start": 1242.38,
                    "end": 1242.52,
                    "word": " I",
                    "probability": 0.9990234375
                },
                {
                    "start": 1242.52,
                    "end": 1242.68,
                    "word": " have",
                    "probability": 0.88671875
                }
            ]
        },
        {
            "id": 187,
            "seek": 111640,
            "start": 1242.68,
            "end": 1248.3,
            "text": " 100,000 points from my training distribution and only 500 points from my test distribution.",
            "tokens": [
                50948,
                2319,
                11,
                1360,
                2793,
                490,
                452,
                3097,
                7316,
                293,
                787,
                5923,
                2793,
                490,
                452,
                1500,
                7316,
                13,
                51236
            ],
            "temperature": 0.0,
            "avg_logprob": -0.2151162809757299,
            "compression_ratio": 1.78,
            "no_speech_prob": 0.0010318756103515625,
            "words": [
                {
                    "start": 1242.68,
                    "end": 1243.08,
                    "word": " 100",
                    "probability": 0.4990234375
                },
                {
                    "start": 1243.08,
                    "end": 1243.42,
                    "word": ",000",
                    "probability": 0.950927734375
                },
                {
                    "start": 1243.42,
                    "end": 1243.78,
                    "word": " points",
                    "probability": 0.79931640625
                },
                {
                    "start": 1243.78,
                    "end": 1243.92,
                    "word": " from",
                    "probability": 0.43408203125
                },
                {
                    "start": 1243.92,
                    "end": 1244.08,
                    "word": " my",
                    "probability": 0.9013671875
                },
                {
                    "start": 1244.08,
                    "end": 1244.4,
                    "word": " training",
                    "probability": 0.80078125
                },
                {
                    "start": 1244.4,
                    "end": 1244.94,
                    "word": " distribution",
                    "probability": 0.73779296875
                },
                {
                    "start": 1244.94,
                    "end": 1245.56,
                    "word": " and",
                    "probability": 0.60498046875
                },
                {
                    "start": 1245.56,
                    "end": 1245.84,
                    "word": " only",
                    "probability": 0.85595703125
                },
                {
                    "start": 1245.84,
                    "end": 1246.26,
                    "word": " 500",
                    "probability": 0.89599609375
                },
                {
                    "start": 1246.26,
                    "end": 1246.66,
                    "word": " points",
                    "probability": 0.77197265625
                },
                {
                    "start": 1246.66,
                    "end": 1246.8,
                    "word": " from",
                    "probability": 0.68212890625
                },
                {
                    "start": 1246.8,
                    "end": 1246.92,
                    "word": " my",
                    "probability": 0.75537109375
                },
                {
                    "start": 1246.92,
                    "end": 1247.4,
                    "word": " test",
                    "probability": 0.7685546875
                },
                {
                    "start": 1247.4,
                    "end": 1247.84,
                    "word": " distribution",
                    "probability": 0.734375
                },
                {
                    "start": 1247.84,
                    "end": 1248.3,
                    "word": ".",
                    "probability": 0.8515625
                }
            ]
        },
        {
            "id": 188,
            "seek": 111640,
            "start": 1248.3,
            "end": 1256.62,
            "text": " And these both have a covariance matrix of all ones. So it's actually quite, and it's",
            "tokens": [
                51236,
                400,
                613,
                1293,
                362,
                257,
                49851,
                719,
                8141,
                295,
                439,
                2306,
                13,
                407,
                309,
                311,
                767,
                1596,
                11,
                293,
                309,
                311,
                51656
            ],
            "temperature": 0.0,
            "avg_logprob": -0.2151162809757299,
            "compression_ratio": 1.78,
            "no_speech_prob": 0.0010318756103515625,
            "words": [
                {
                    "start": 1248.3,
                    "end": 1248.62,
                    "word": " And",
                    "probability": 0.8740234375
                },
                {
                    "start": 1248.62,
                    "end": 1248.86,
                    "word": " these",
                    "probability": 0.71630859375
                },
                {
                    "start": 1248.86,
                    "end": 1249.3,
                    "word": " both",
                    "probability": 0.80810546875
                },
                {
                    "start": 1249.3,
                    "end": 1249.8,
                    "word": " have",
                    "probability": 0.88330078125
                },
                {
                    "start": 1250.5,
                    "end": 1251.68,
                    "word": " a covariance",
                    "probability": 0.8623046875
                },
                {
                    "start": 1251.68,
                    "end": 1252.66,
                    "word": " matrix",
                    "probability": 0.81787109375
                },
                {
                    "start": 1252.66,
                    "end": 1253.22,
                    "word": " of",
                    "probability": 0.904296875
                },
                {
                    "start": 1253.22,
                    "end": 1253.7,
                    "word": " all",
                    "probability": 0.90283203125
                },
                {
                    "start": 1253.7,
                    "end": 1254.14,
                    "word": " ones",
                    "probability": 0.5205078125
                },
                {
                    "start": 1254.14,
                    "end": 1254.34,
                    "word": ".",
                    "probability": 0.65771484375
                },
                {
                    "start": 1254.34,
                    "end": 1254.42,
                    "word": " So",
                    "probability": 0.82421875
                },
                {
                    "start": 1254.42,
                    "end": 1254.96,
                    "word": " it's",
                    "probability": 0.923583984375
                },
                {
                    "start": 1254.96,
                    "end": 1255.3,
                    "word": " actually",
                    "probability": 0.7763671875
                },
                {
                    "start": 1255.3,
                    "end": 1255.78,
                    "word": " quite",
                    "probability": 0.88623046875
                },
                {
                    "start": 1255.78,
                    "end": 1256.24,
                    "word": ",",
                    "probability": 0.36572265625
                },
                {
                    "start": 1256.24,
                    "end": 1256.46,
                    "word": " and",
                    "probability": 0.58984375
                },
                {
                    "start": 1256.46,
                    "end": 1256.62,
                    "word": " it's",
                    "probability": 0.942138671875
                }
            ]
        },
        {
            "id": 189,
            "seek": 114020,
            "start": 1256.62,
            "end": 1264.16,
            "text": " 20-dimensional. This is actually harder than most datasets. MNIST is about 10-dimensional,",
            "tokens": [
                50368,
                945,
                12,
                18759,
                13,
                639,
                307,
                767,
                6081,
                813,
                881,
                42856,
                13,
                376,
                45,
                19756,
                307,
                466,
                1266,
                12,
                18759,
                11,
                50736
            ],
            "temperature": 0.0,
            "avg_logprob": -0.22583575616049212,
            "compression_ratio": 1.5377358490566038,
            "no_speech_prob": 0.05340576171875,
            "words": [
                {
                    "start": 1256.62,
                    "end": 1257.02,
                    "word": " 20",
                    "probability": 0.328369140625
                },
                {
                    "start": 1257.02,
                    "end": 1257.42,
                    "word": "-dimensional",
                    "probability": 0.6312255859375
                },
                {
                    "start": 1257.42,
                    "end": 1257.86,
                    "word": ".",
                    "probability": 0.7314453125
                },
                {
                    "start": 1257.88,
                    "end": 1258.12,
                    "word": " This",
                    "probability": 0.7060546875
                },
                {
                    "start": 1258.12,
                    "end": 1258.34,
                    "word": " is",
                    "probability": 0.9052734375
                },
                {
                    "start": 1258.34,
                    "end": 1258.98,
                    "word": " actually",
                    "probability": 0.7802734375
                },
                {
                    "start": 1258.98,
                    "end": 1259.82,
                    "word": " harder",
                    "probability": 0.75244140625
                },
                {
                    "start": 1259.82,
                    "end": 1260.12,
                    "word": " than",
                    "probability": 0.88525390625
                },
                {
                    "start": 1260.12,
                    "end": 1260.38,
                    "word": " most",
                    "probability": 0.82080078125
                },
                {
                    "start": 1260.38,
                    "end": 1260.78,
                    "word": " datasets",
                    "probability": 0.5322265625
                },
                {
                    "start": 1260.78,
                    "end": 1261.22,
                    "word": ".",
                    "probability": 0.65673828125
                },
                {
                    "start": 1261.68,
                    "end": 1262.12,
                    "word": " MNIST",
                    "probability": 0.86279296875
                },
                {
                    "start": 1262.12,
                    "end": 1262.48,
                    "word": " is",
                    "probability": 0.90234375
                },
                {
                    "start": 1262.48,
                    "end": 1262.86,
                    "word": " about",
                    "probability": 0.80859375
                },
                {
                    "start": 1262.86,
                    "end": 1263.6,
                    "word": " 10",
                    "probability": 0.8720703125
                },
                {
                    "start": 1263.6,
                    "end": 1263.96,
                    "word": "-dimensional",
                    "probability": 0.912841796875
                },
                {
                    "start": 1263.96,
                    "end": 1264.16,
                    "word": ",",
                    "probability": 0.2958984375
                }
            ]
        },
        {
            "id": 190,
            "seek": 114020,
            "start": 1264.16,
            "end": 1267.14,
            "text": " if you really get down to it. If you want to know what that is, you can",
            "tokens": [
                50736,
                498,
                291,
                534,
                483,
                760,
                281,
                309,
                13,
                759,
                291,
                528,
                281,
                458,
                437,
                300,
                307,
                11,
                291,
                393,
                50884
            ],
            "temperature": 0.0,
            "avg_logprob": -0.22583575616049212,
            "compression_ratio": 1.5377358490566038,
            "no_speech_prob": 0.05340576171875,
            "words": [
                {
                    "start": 1264.16,
                    "end": 1264.26,
                    "word": " if",
                    "probability": 0.88720703125
                },
                {
                    "start": 1264.26,
                    "end": 1264.36,
                    "word": " you",
                    "probability": 0.89697265625
                },
                {
                    "start": 1264.36,
                    "end": 1264.62,
                    "word": " really",
                    "probability": 0.75146484375
                },
                {
                    "start": 1264.62,
                    "end": 1265.22,
                    "word": " get",
                    "probability": 0.87939453125
                },
                {
                    "start": 1265.22,
                    "end": 1265.42,
                    "word": " down",
                    "probability": 0.81982421875
                },
                {
                    "start": 1265.42,
                    "end": 1265.54,
                    "word": " to",
                    "probability": 0.90380859375
                },
                {
                    "start": 1265.54,
                    "end": 1265.76,
                    "word": " it",
                    "probability": 0.89208984375
                },
                {
                    "start": 1265.76,
                    "end": 1265.88,
                    "word": ".",
                    "probability": 0.80029296875
                },
                {
                    "start": 1265.88,
                    "end": 1265.98,
                    "word": " If",
                    "probability": 0.50634765625
                },
                {
                    "start": 1265.98,
                    "end": 1266.02,
                    "word": " you",
                    "probability": 0.89208984375
                },
                {
                    "start": 1266.02,
                    "end": 1266.12,
                    "word": " want",
                    "probability": 0.81103515625
                },
                {
                    "start": 1266.12,
                    "end": 1266.2,
                    "word": " to",
                    "probability": 0.90234375
                },
                {
                    "start": 1266.2,
                    "end": 1266.28,
                    "word": " know",
                    "probability": 0.82470703125
                },
                {
                    "start": 1266.28,
                    "end": 1266.4,
                    "word": " what",
                    "probability": 0.880859375
                },
                {
                    "start": 1266.4,
                    "end": 1266.52,
                    "word": " that",
                    "probability": 0.87255859375
                },
                {
                    "start": 1266.52,
                    "end": 1266.7,
                    "word": " is",
                    "probability": 0.90380859375
                },
                {
                    "start": 1266.7,
                    "end": 1266.8,
                    "word": ",",
                    "probability": 0.75341796875
                },
                {
                    "start": 1266.8,
                    "end": 1266.92,
                    "word": " you",
                    "probability": 0.8857421875
                },
                {
                    "start": 1266.92,
                    "end": 1267.14,
                    "word": " can",
                    "probability": 0.89697265625
                }
            ]
        },
        {
            "id": 191,
            "seek": 114020,
            "start": 1267.6,
            "end": 1272.26,
            "text": " contact me on Twitter and ask me why MNIST might be lower dimensional.",
            "tokens": [
                50912,
                3385,
                385,
                322,
                5794,
                293,
                1029,
                385,
                983,
                376,
                45,
                19756,
                1062,
                312,
                3126,
                18795,
                13,
                51128
            ],
            "temperature": 0.0,
            "avg_logprob": -0.22583575616049212,
            "compression_ratio": 1.5377358490566038,
            "no_speech_prob": 0.05340576171875,
            "words": [
                {
                    "start": 1267.6,
                    "end": 1268.04,
                    "word": " contact",
                    "probability": 0.84130859375
                },
                {
                    "start": 1268.04,
                    "end": 1268.22,
                    "word": " me",
                    "probability": 0.9091796875
                },
                {
                    "start": 1268.22,
                    "end": 1268.36,
                    "word": " on",
                    "probability": 0.88818359375
                },
                {
                    "start": 1268.36,
                    "end": 1268.58,
                    "word": " Twitter",
                    "probability": 0.6845703125
                },
                {
                    "start": 1268.58,
                    "end": 1268.88,
                    "word": " and",
                    "probability": 0.82568359375
                },
                {
                    "start": 1268.88,
                    "end": 1269.08,
                    "word": " ask",
                    "probability": 0.783203125
                },
                {
                    "start": 1269.08,
                    "end": 1269.22,
                    "word": " me",
                    "probability": 0.89208984375
                },
                {
                    "start": 1269.22,
                    "end": 1269.5,
                    "word": " why",
                    "probability": 0.81494140625
                },
                {
                    "start": 1269.5,
                    "end": 1269.82,
                    "word": " MNIST",
                    "probability": 0.88916015625
                },
                {
                    "start": 1269.82,
                    "end": 1270.06,
                    "word": " might",
                    "probability": 0.80810546875
                },
                {
                    "start": 1270.06,
                    "end": 1270.24,
                    "word": " be",
                    "probability": 0.90283203125
                },
                {
                    "start": 1270.24,
                    "end": 1271.36,
                    "word": " lower",
                    "probability": 0.7412109375
                },
                {
                    "start": 1271.36,
                    "end": 1271.82,
                    "word": " dimensional",
                    "probability": 0.56982421875
                },
                {
                    "start": 1271.82,
                    "end": 1272.26,
                    "word": ".",
                    "probability": 0.9736328125
                }
            ]
        },
        {
            "id": 192,
            "seek": 114020,
            "start": 1275.14,
            "end": 1280.42,
            "text": " And between the training distribution, I build the tree, fix it, and then I make a posterior",
            "tokens": [
                51256,
                400,
                1296,
                264,
                3097,
                7316,
                11,
                286,
                1322,
                264,
                4230,
                11,
                3191,
                309,
                11,
                293,
                550,
                286,
                652,
                257,
                33529,
                51552
            ],
            "temperature": 0.0,
            "avg_logprob": -0.22583575616049212,
            "compression_ratio": 1.5377358490566038,
            "no_speech_prob": 0.05340576171875,
            "words": [
                {
                    "start": 1275.14,
                    "end": 1275.58,
                    "word": " And",
                    "probability": 0.607421875
                },
                {
                    "start": 1275.58,
                    "end": 1276.02,
                    "word": " between",
                    "probability": 0.75244140625
                },
                {
                    "start": 1276.02,
                    "end": 1276.64,
                    "word": " the",
                    "probability": 0.80908203125
                },
                {
                    "start": 1276.64,
                    "end": 1277.1,
                    "word": " training",
                    "probability": 0.79736328125
                },
                {
                    "start": 1277.1,
                    "end": 1277.58,
                    "word": " distribution",
                    "probability": 0.73046875
                },
                {
                    "start": 1277.58,
                    "end": 1277.92,
                    "word": ",",
                    "probability": 0.4384765625
                },
                {
                    "start": 1277.92,
                    "end": 1277.92,
                    "word": " I",
                    "probability": 0.6396484375
                },
                {
                    "start": 1277.92,
                    "end": 1278.24,
                    "word": " build",
                    "probability": 0.796875
                },
                {
                    "start": 1278.24,
                    "end": 1278.38,
                    "word": " the",
                    "probability": 0.60009765625
                },
                {
                    "start": 1278.38,
                    "end": 1278.52,
                    "word": " tree",
                    "probability": 0.81201171875
                },
                {
                    "start": 1278.52,
                    "end": 1278.64,
                    "word": ",",
                    "probability": 0.8818359375
                },
                {
                    "start": 1278.64,
                    "end": 1278.78,
                    "word": " fix",
                    "probability": 0.9013671875
                },
                {
                    "start": 1278.78,
                    "end": 1278.96,
                    "word": " it",
                    "probability": 0.90380859375
                },
                {
                    "start": 1278.96,
                    "end": 1279.02,
                    "word": ",",
                    "probability": 0.85888671875
                },
                {
                    "start": 1279.02,
                    "end": 1279.2,
                    "word": " and",
                    "probability": 0.88330078125
                },
                {
                    "start": 1279.2,
                    "end": 1279.34,
                    "word": " then",
                    "probability": 0.72265625
                },
                {
                    "start": 1279.34,
                    "end": 1279.52,
                    "word": " I",
                    "probability": 0.98779296875
                },
                {
                    "start": 1279.52,
                    "end": 1280.02,
                    "word": " make",
                    "probability": 0.8671875
                },
                {
                    "start": 1280.14,
                    "end": 1280.42,
                    "word": " a posterior",
                    "probability": 0.7197265625
                }
            ]
        },
        {
            "id": 193,
            "seek": 116852,
            "start": 1280.42,
            "end": 1287.14,
            "text": " distribution for a bunch of test distributions. And I take the KL divergence. And you can see",
            "tokens": [
                50364,
                7316,
                337,
                257,
                3840,
                295,
                1500,
                37870,
                13,
                400,
                286,
                747,
                264,
                47991,
                47387,
                13,
                400,
                291,
                393,
                536,
                50696
            ],
            "temperature": 0.0,
            "avg_logprob": -0.20622417453891975,
            "compression_ratio": 1.6770833333333333,
            "no_speech_prob": 0.310546875,
            "words": [
                {
                    "start": 1280.42,
                    "end": 1280.82,
                    "word": " distribution",
                    "probability": 0.46923828125
                },
                {
                    "start": 1280.82,
                    "end": 1281.14,
                    "word": " for",
                    "probability": 0.85546875
                },
                {
                    "start": 1281.52,
                    "end": 1281.78,
                    "word": " a bunch",
                    "probability": 0.79931640625
                },
                {
                    "start": 1281.78,
                    "end": 1282.4,
                    "word": " of",
                    "probability": 0.90185546875
                },
                {
                    "start": 1282.4,
                    "end": 1282.78,
                    "word": " test",
                    "probability": 0.82568359375
                },
                {
                    "start": 1282.78,
                    "end": 1283.34,
                    "word": " distributions",
                    "probability": 0.8310546875
                },
                {
                    "start": 1283.34,
                    "end": 1283.78,
                    "word": ".",
                    "probability": 0.281494140625
                },
                {
                    "start": 1283.8,
                    "end": 1284.16,
                    "word": " And",
                    "probability": 0.2359619140625
                },
                {
                    "start": 1284.16,
                    "end": 1284.24,
                    "word": " I",
                    "probability": 0.9404296875
                },
                {
                    "start": 1284.24,
                    "end": 1284.58,
                    "word": " take",
                    "probability": 0.80810546875
                },
                {
                    "start": 1284.58,
                    "end": 1285.1,
                    "word": " the",
                    "probability": 0.8125
                },
                {
                    "start": 1285.1,
                    "end": 1285.36,
                    "word": " KL",
                    "probability": 0.7783203125
                },
                {
                    "start": 1285.36,
                    "end": 1285.82,
                    "word": " divergence",
                    "probability": 0.716796875
                },
                {
                    "start": 1285.82,
                    "end": 1286.26,
                    "word": ".",
                    "probability": 0.55517578125
                },
                {
                    "start": 1286.38,
                    "end": 1286.72,
                    "word": " And",
                    "probability": 0.8271484375
                },
                {
                    "start": 1286.72,
                    "end": 1286.82,
                    "word": " you",
                    "probability": 0.88818359375
                },
                {
                    "start": 1286.82,
                    "end": 1286.96,
                    "word": " can",
                    "probability": 0.89697265625
                },
                {
                    "start": 1286.96,
                    "end": 1287.14,
                    "word": " see",
                    "probability": 0.81787109375
                }
            ]
        },
        {
            "id": 194,
            "seek": 116852,
            "start": 1287.14,
            "end": 1293.74,
            "text": " it's a bit noisy. The reason being is I generate a lot of the exact position of your points. It's",
            "tokens": [
                50696,
                309,
                311,
                257,
                857,
                24518,
                13,
                440,
                1778,
                885,
                307,
                286,
                8460,
                257,
                688,
                295,
                264,
                1900,
                2535,
                295,
                428,
                2793,
                13,
                467,
                311,
                51028
            ],
            "temperature": 0.0,
            "avg_logprob": -0.20622417453891975,
            "compression_ratio": 1.6770833333333333,
            "no_speech_prob": 0.310546875,
            "words": [
                {
                    "start": 1287.14,
                    "end": 1287.32,
                    "word": " it's",
                    "probability": 0.91259765625
                },
                {
                    "start": 1287.42,
                    "end": 1287.54,
                    "word": " a bit",
                    "probability": 0.90234375
                },
                {
                    "start": 1287.54,
                    "end": 1287.88,
                    "word": " noisy",
                    "probability": 0.7724609375
                },
                {
                    "start": 1287.88,
                    "end": 1288.3,
                    "word": ".",
                    "probability": 0.61669921875
                },
                {
                    "start": 1288.48,
                    "end": 1288.92,
                    "word": " The",
                    "probability": 0.76513671875
                },
                {
                    "start": 1288.92,
                    "end": 1289.18,
                    "word": " reason",
                    "probability": 0.91943359375
                },
                {
                    "start": 1289.18,
                    "end": 1289.46,
                    "word": " being",
                    "probability": 0.87646484375
                },
                {
                    "start": 1289.46,
                    "end": 1289.66,
                    "word": " is",
                    "probability": 0.81005859375
                },
                {
                    "start": 1289.66,
                    "end": 1289.8,
                    "word": " I",
                    "probability": 0.86474609375
                },
                {
                    "start": 1289.8,
                    "end": 1290.28,
                    "word": " generate",
                    "probability": 0.8046875
                },
                {
                    "start": 1290.48,
                    "end": 1290.7,
                    "word": " a lot",
                    "probability": 0.8974609375
                },
                {
                    "start": 1290.7,
                    "end": 1290.92,
                    "word": " of",
                    "probability": 0.89599609375
                },
                {
                    "start": 1290.92,
                    "end": 1291.7,
                    "word": " the",
                    "probability": 0.6513671875
                },
                {
                    "start": 1291.7,
                    "end": 1292.3,
                    "word": " exact",
                    "probability": 0.84716796875
                },
                {
                    "start": 1292.3,
                    "end": 1292.74,
                    "word": " position",
                    "probability": 0.82763671875
                },
                {
                    "start": 1292.74,
                    "end": 1292.88,
                    "word": " of",
                    "probability": 0.8984375
                },
                {
                    "start": 1292.88,
                    "end": 1293.0,
                    "word": " your",
                    "probability": 0.80908203125
                },
                {
                    "start": 1293.0,
                    "end": 1293.34,
                    "word": " points",
                    "probability": 0.73046875
                },
                {
                    "start": 1293.34,
                    "end": 1293.54,
                    "word": ".",
                    "probability": 0.546875
                },
                {
                    "start": 1293.54,
                    "end": 1293.74,
                    "word": " It's",
                    "probability": 0.9453125
                }
            ]
        },
        {
            "id": 195,
            "seek": 116852,
            "start": 1293.74,
            "end": 1297.94,
            "text": " fairly sensitive to these. So this is one of the problems. And one of the things I'll talk about",
            "tokens": [
                51028,
                6457,
                9477,
                281,
                613,
                13,
                407,
                341,
                307,
                472,
                295,
                264,
                2740,
                13,
                400,
                472,
                295,
                264,
                721,
                286,
                603,
                751,
                466,
                51236
            ],
            "temperature": 0.0,
            "avg_logprob": -0.20622417453891975,
            "compression_ratio": 1.6770833333333333,
            "no_speech_prob": 0.310546875,
            "words": [
                {
                    "start": 1293.74,
                    "end": 1294.1,
                    "word": " fairly",
                    "probability": 0.86279296875
                },
                {
                    "start": 1294.1,
                    "end": 1294.46,
                    "word": " sensitive",
                    "probability": 0.80517578125
                },
                {
                    "start": 1294.46,
                    "end": 1294.64,
                    "word": " to",
                    "probability": 0.81982421875
                },
                {
                    "start": 1294.64,
                    "end": 1294.76,
                    "word": " these",
                    "probability": 0.294677734375
                },
                {
                    "start": 1294.76,
                    "end": 1295.2,
                    "word": ".",
                    "probability": 0.87744140625
                },
                {
                    "start": 1295.34,
                    "end": 1295.66,
                    "word": " So",
                    "probability": 0.87158203125
                },
                {
                    "start": 1295.66,
                    "end": 1295.86,
                    "word": " this",
                    "probability": 0.8203125
                },
                {
                    "start": 1295.86,
                    "end": 1295.98,
                    "word": " is",
                    "probability": 0.90869140625
                },
                {
                    "start": 1295.98,
                    "end": 1296.2,
                    "word": " one",
                    "probability": 0.822265625
                },
                {
                    "start": 1296.2,
                    "end": 1296.26,
                    "word": " of",
                    "probability": 0.904296875
                },
                {
                    "start": 1296.26,
                    "end": 1296.36,
                    "word": " the",
                    "probability": 0.81884765625
                },
                {
                    "start": 1296.36,
                    "end": 1296.66,
                    "word": " problems",
                    "probability": 0.71435546875
                },
                {
                    "start": 1296.66,
                    "end": 1296.76,
                    "word": ".",
                    "probability": 0.2161865234375
                },
                {
                    "start": 1296.76,
                    "end": 1296.84,
                    "word": " And",
                    "probability": 0.89013671875
                },
                {
                    "start": 1296.84,
                    "end": 1297.02,
                    "word": " one",
                    "probability": 0.81982421875
                },
                {
                    "start": 1297.02,
                    "end": 1297.1,
                    "word": " of",
                    "probability": 0.90966796875
                },
                {
                    "start": 1297.1,
                    "end": 1297.22,
                    "word": " the",
                    "probability": 0.81640625
                },
                {
                    "start": 1297.22,
                    "end": 1297.4,
                    "word": " things",
                    "probability": 0.734375
                },
                {
                    "start": 1297.4,
                    "end": 1297.56,
                    "word": " I'll",
                    "probability": 0.960693359375
                },
                {
                    "start": 1297.56,
                    "end": 1297.74,
                    "word": " talk",
                    "probability": 0.818359375
                },
                {
                    "start": 1297.74,
                    "end": 1297.94,
                    "word": " about",
                    "probability": 0.818359375
                }
            ]
        },
        {
            "id": 196,
            "seek": 116852,
            "start": 1297.94,
            "end": 1303.88,
            "text": " at the end is like, there are ways to normalize this and resolve this, because I've designed it",
            "tokens": [
                51236,
                412,
                264,
                917,
                307,
                411,
                11,
                456,
                366,
                2098,
                281,
                2710,
                1125,
                341,
                293,
                14151,
                341,
                11,
                570,
                286,
                600,
                4761,
                309,
                51532
            ],
            "temperature": 0.0,
            "avg_logprob": -0.20622417453891975,
            "compression_ratio": 1.6770833333333333,
            "no_speech_prob": 0.310546875,
            "words": [
                {
                    "start": 1297.94,
                    "end": 1298.04,
                    "word": " at",
                    "probability": 0.875
                },
                {
                    "start": 1298.04,
                    "end": 1298.14,
                    "word": " the",
                    "probability": 0.81396484375
                },
                {
                    "start": 1298.14,
                    "end": 1298.26,
                    "word": " end",
                    "probability": 0.81884765625
                },
                {
                    "start": 1298.26,
                    "end": 1298.38,
                    "word": " is",
                    "probability": 0.845703125
                },
                {
                    "start": 1298.38,
                    "end": 1298.56,
                    "word": " like",
                    "probability": 0.66455078125
                },
                {
                    "start": 1298.56,
                    "end": 1298.82,
                    "word": ",",
                    "probability": 0.6064453125
                },
                {
                    "start": 1298.82,
                    "end": 1299.04,
                    "word": " there",
                    "probability": 0.83984375
                },
                {
                    "start": 1299.04,
                    "end": 1299.22,
                    "word": " are",
                    "probability": 0.8955078125
                },
                {
                    "start": 1299.22,
                    "end": 1299.56,
                    "word": " ways",
                    "probability": 0.79296875
                },
                {
                    "start": 1299.56,
                    "end": 1299.78,
                    "word": " to",
                    "probability": 0.9013671875
                },
                {
                    "start": 1299.78,
                    "end": 1300.62,
                    "word": " normalize",
                    "probability": 0.8486328125
                },
                {
                    "start": 1300.62,
                    "end": 1300.76,
                    "word": " this",
                    "probability": 0.8828125
                },
                {
                    "start": 1300.76,
                    "end": 1300.9,
                    "word": " and",
                    "probability": 0.88720703125
                },
                {
                    "start": 1300.9,
                    "end": 1301.26,
                    "word": " resolve",
                    "probability": 0.68994140625
                },
                {
                    "start": 1301.26,
                    "end": 1301.5,
                    "word": " this",
                    "probability": 0.8740234375
                },
                {
                    "start": 1301.5,
                    "end": 1301.8,
                    "word": ",",
                    "probability": 0.45263671875
                },
                {
                    "start": 1301.8,
                    "end": 1302.12,
                    "word": " because",
                    "probability": 0.83251953125
                },
                {
                    "start": 1302.12,
                    "end": 1303.54,
                    "word": " I've",
                    "probability": 0.890625
                },
                {
                    "start": 1303.54,
                    "end": 1303.74,
                    "word": " designed",
                    "probability": 0.78564453125
                },
                {
                    "start": 1303.74,
                    "end": 1303.88,
                    "word": " it",
                    "probability": 0.89794921875
                }
            ]
        },
        {
            "id": 197,
            "seek": 116852,
            "start": 1303.88,
            "end": 1308.74,
            "text": " for speed, not really for accuracy. And I'll get into why I did that afterwards, because honestly,",
            "tokens": [
                51532,
                337,
                3073,
                11,
                406,
                534,
                337,
                14170,
                13,
                400,
                286,
                603,
                483,
                666,
                983,
                286,
                630,
                300,
                10543,
                11,
                570,
                6095,
                11,
                51776
            ],
            "temperature": 0.0,
            "avg_logprob": -0.20622417453891975,
            "compression_ratio": 1.6770833333333333,
            "no_speech_prob": 0.310546875,
            "words": [
                {
                    "start": 1303.88,
                    "end": 1304.02,
                    "word": " for",
                    "probability": 0.89892578125
                },
                {
                    "start": 1304.02,
                    "end": 1304.42,
                    "word": " speed",
                    "probability": 0.86328125
                },
                {
                    "start": 1304.42,
                    "end": 1304.48,
                    "word": ",",
                    "probability": 0.93115234375
                },
                {
                    "start": 1304.48,
                    "end": 1304.76,
                    "word": " not",
                    "probability": 0.892578125
                },
                {
                    "start": 1304.76,
                    "end": 1305.04,
                    "word": " really",
                    "probability": 0.7998046875
                },
                {
                    "start": 1305.04,
                    "end": 1305.2,
                    "word": " for",
                    "probability": 0.89306640625
                },
                {
                    "start": 1305.2,
                    "end": 1305.54,
                    "word": " accuracy",
                    "probability": 0.7705078125
                },
                {
                    "start": 1305.54,
                    "end": 1305.98,
                    "word": ".",
                    "probability": 0.96240234375
                },
                {
                    "start": 1306.14,
                    "end": 1306.54,
                    "word": " And",
                    "probability": 0.8818359375
                },
                {
                    "start": 1306.54,
                    "end": 1306.66,
                    "word": " I'll",
                    "probability": 0.972900390625
                },
                {
                    "start": 1306.66,
                    "end": 1306.78,
                    "word": " get",
                    "probability": 0.89208984375
                },
                {
                    "start": 1306.78,
                    "end": 1306.96,
                    "word": " into",
                    "probability": 0.7451171875
                },
                {
                    "start": 1306.96,
                    "end": 1307.34,
                    "word": " why",
                    "probability": 0.77099609375
                },
                {
                    "start": 1307.34,
                    "end": 1307.44,
                    "word": " I",
                    "probability": 0.953125
                },
                {
                    "start": 1307.44,
                    "end": 1307.54,
                    "word": " did",
                    "probability": 0.90478515625
                },
                {
                    "start": 1307.54,
                    "end": 1307.68,
                    "word": " that",
                    "probability": 0.88134765625
                },
                {
                    "start": 1307.68,
                    "end": 1308.1,
                    "word": " afterwards",
                    "probability": 0.74072265625
                },
                {
                    "start": 1308.1,
                    "end": 1308.16,
                    "word": ",",
                    "probability": 0.458251953125
                },
                {
                    "start": 1308.16,
                    "end": 1308.38,
                    "word": " because",
                    "probability": 0.8330078125
                },
                {
                    "start": 1308.38,
                    "end": 1308.68,
                    "word": " honestly",
                    "probability": 0.76220703125
                },
                {
                    "start": 1308.68,
                    "end": 1308.74,
                    "word": ",",
                    "probability": 0.93896484375
                }
            ]
        },
        {
            "id": 198,
            "seek": 119852,
            "start": 1308.74,
            "end": 1314.4,
            "text": " it came to the drift stuff backwards. But you can see, the further you get, the further you",
            "tokens": [
                50388,
                309,
                1361,
                281,
                264,
                19699,
                1507,
                12204,
                13,
                583,
                291,
                393,
                536,
                11,
                264,
                3052,
                291,
                483,
                11,
                264,
                3052,
                291,
                50644
            ],
            "temperature": 0.0,
            "avg_logprob": -0.27234374105930326,
            "compression_ratio": 1.728448275862069,
            "no_speech_prob": 0.0074615478515625,
            "words": [
                {
                    "start": 1308.74,
                    "end": 1308.84,
                    "word": " it",
                    "probability": 0.2607421875
                },
                {
                    "start": 1308.84,
                    "end": 1309.04,
                    "word": " came",
                    "probability": 0.78515625
                },
                {
                    "start": 1309.04,
                    "end": 1309.82,
                    "word": " to",
                    "probability": 0.80712890625
                },
                {
                    "start": 1309.82,
                    "end": 1310.02,
                    "word": " the",
                    "probability": 0.79541015625
                },
                {
                    "start": 1310.02,
                    "end": 1310.22,
                    "word": " drift",
                    "probability": 0.80517578125
                },
                {
                    "start": 1310.22,
                    "end": 1310.44,
                    "word": " stuff",
                    "probability": 0.85791015625
                },
                {
                    "start": 1310.44,
                    "end": 1310.98,
                    "word": " backwards",
                    "probability": 0.77880859375
                },
                {
                    "start": 1310.98,
                    "end": 1311.46,
                    "word": ".",
                    "probability": 0.61767578125
                },
                {
                    "start": 1311.62,
                    "end": 1311.92,
                    "word": " But",
                    "probability": 0.8603515625
                },
                {
                    "start": 1311.92,
                    "end": 1312.04,
                    "word": " you",
                    "probability": 0.88525390625
                },
                {
                    "start": 1312.04,
                    "end": 1312.16,
                    "word": " can",
                    "probability": 0.8974609375
                },
                {
                    "start": 1312.16,
                    "end": 1312.38,
                    "word": " see",
                    "probability": 0.818359375
                },
                {
                    "start": 1312.38,
                    "end": 1312.94,
                    "word": ",",
                    "probability": 0.40234375
                },
                {
                    "start": 1312.94,
                    "end": 1312.96,
                    "word": " the",
                    "probability": 0.81103515625
                },
                {
                    "start": 1312.96,
                    "end": 1313.4,
                    "word": " further",
                    "probability": 0.8564453125
                },
                {
                    "start": 1313.4,
                    "end": 1313.6,
                    "word": " you",
                    "probability": 0.880859375
                },
                {
                    "start": 1313.6,
                    "end": 1313.76,
                    "word": " get",
                    "probability": 0.14990234375
                },
                {
                    "start": 1313.76,
                    "end": 1313.9,
                    "word": ",",
                    "probability": 0.68603515625
                },
                {
                    "start": 1313.9,
                    "end": 1313.98,
                    "word": " the",
                    "probability": 0.81201171875
                },
                {
                    "start": 1313.98,
                    "end": 1314.22,
                    "word": " further",
                    "probability": 0.85791015625
                },
                {
                    "start": 1314.22,
                    "end": 1314.4,
                    "word": " you",
                    "probability": 0.89501953125
                }
            ]
        },
        {
            "id": 199,
            "seek": 119852,
            "start": 1314.4,
            "end": 1320.68,
            "text": " separate the two distributions, the higher the number gets. So the sanity test, the unit test",
            "tokens": [
                50644,
                4994,
                264,
                732,
                37870,
                11,
                264,
                2946,
                264,
                1230,
                2170,
                13,
                407,
                264,
                47892,
                1500,
                11,
                264,
                4985,
                1500,
                50972
            ],
            "temperature": 0.0,
            "avg_logprob": -0.27234374105930326,
            "compression_ratio": 1.728448275862069,
            "no_speech_prob": 0.0074615478515625,
            "words": [
                {
                    "start": 1314.4,
                    "end": 1314.74,
                    "word": " separate",
                    "probability": 0.837890625
                },
                {
                    "start": 1314.74,
                    "end": 1314.92,
                    "word": " the",
                    "probability": 0.810546875
                },
                {
                    "start": 1314.92,
                    "end": 1315.24,
                    "word": " two",
                    "probability": 0.81884765625
                },
                {
                    "start": 1315.24,
                    "end": 1316.18,
                    "word": " distributions",
                    "probability": 0.83642578125
                },
                {
                    "start": 1316.18,
                    "end": 1316.78,
                    "word": ",",
                    "probability": 0.90966796875
                },
                {
                    "start": 1316.78,
                    "end": 1316.98,
                    "word": " the",
                    "probability": 0.81396484375
                },
                {
                    "start": 1316.98,
                    "end": 1318.1,
                    "word": " higher",
                    "probability": 0.84228515625
                },
                {
                    "start": 1318.1,
                    "end": 1318.26,
                    "word": " the",
                    "probability": 0.76904296875
                },
                {
                    "start": 1318.26,
                    "end": 1318.44,
                    "word": " number",
                    "probability": 0.8818359375
                },
                {
                    "start": 1318.44,
                    "end": 1318.74,
                    "word": " gets",
                    "probability": 0.78759765625
                },
                {
                    "start": 1318.74,
                    "end": 1319.04,
                    "word": ".",
                    "probability": 0.96533203125
                },
                {
                    "start": 1319.04,
                    "end": 1319.22,
                    "word": " So",
                    "probability": 0.89599609375
                },
                {
                    "start": 1319.22,
                    "end": 1319.48,
                    "word": " the",
                    "probability": 0.67529296875
                },
                {
                    "start": 1319.48,
                    "end": 1319.74,
                    "word": " sanity",
                    "probability": 0.8974609375
                },
                {
                    "start": 1319.74,
                    "end": 1320.04,
                    "word": " test",
                    "probability": 0.79248046875
                },
                {
                    "start": 1320.04,
                    "end": 1320.14,
                    "word": ",",
                    "probability": 0.970703125
                },
                {
                    "start": 1320.14,
                    "end": 1320.18,
                    "word": " the",
                    "probability": 0.80810546875
                },
                {
                    "start": 1320.18,
                    "end": 1320.36,
                    "word": " unit",
                    "probability": 0.84033203125
                },
                {
                    "start": 1320.36,
                    "end": 1320.68,
                    "word": " test",
                    "probability": 0.79736328125
                }
            ]
        },
        {
            "id": 200,
            "seek": 119852,
            "start": 1320.68,
            "end": 1327.16,
            "text": " of does this work is correct. And I'll have to compare this to other methods.",
            "tokens": [
                50972,
                295,
                775,
                341,
                589,
                307,
                3006,
                13,
                400,
                286,
                603,
                362,
                281,
                6794,
                341,
                281,
                661,
                7150,
                13,
                51260
            ],
            "temperature": 0.0,
            "avg_logprob": -0.27234374105930326,
            "compression_ratio": 1.728448275862069,
            "no_speech_prob": 0.0074615478515625,
            "words": [
                {
                    "start": 1320.68,
                    "end": 1321.16,
                    "word": " of",
                    "probability": 0.705078125
                },
                {
                    "start": 1321.16,
                    "end": 1321.34,
                    "word": " does",
                    "probability": 0.69189453125
                },
                {
                    "start": 1321.34,
                    "end": 1321.5,
                    "word": " this",
                    "probability": 0.88525390625
                },
                {
                    "start": 1321.5,
                    "end": 1321.82,
                    "word": " work",
                    "probability": 0.80419921875
                },
                {
                    "start": 1321.82,
                    "end": 1322.48,
                    "word": " is",
                    "probability": 0.70068359375
                },
                {
                    "start": 1322.48,
                    "end": 1322.72,
                    "word": " correct",
                    "probability": 0.83154296875
                },
                {
                    "start": 1322.72,
                    "end": 1323.2,
                    "word": ".",
                    "probability": 0.93505859375
                },
                {
                    "start": 1324.86,
                    "end": 1325.34,
                    "word": " And",
                    "probability": 0.74951171875
                },
                {
                    "start": 1325.34,
                    "end": 1325.5,
                    "word": " I'll",
                    "probability": 0.424560546875
                },
                {
                    "start": 1325.5,
                    "end": 1325.54,
                    "word": " have",
                    "probability": 0.69384765625
                },
                {
                    "start": 1325.54,
                    "end": 1325.58,
                    "word": " to",
                    "probability": 0.8974609375
                },
                {
                    "start": 1325.58,
                    "end": 1325.9,
                    "word": " compare",
                    "probability": 0.88232421875
                },
                {
                    "start": 1325.9,
                    "end": 1326.1,
                    "word": " this",
                    "probability": 0.8603515625
                },
                {
                    "start": 1326.1,
                    "end": 1326.2,
                    "word": " to",
                    "probability": 0.8779296875
                },
                {
                    "start": 1326.2,
                    "end": 1326.34,
                    "word": " other",
                    "probability": 0.7822265625
                },
                {
                    "start": 1326.34,
                    "end": 1326.68,
                    "word": " methods",
                    "probability": 0.82666015625
                },
                {
                    "start": 1326.68,
                    "end": 1327.16,
                    "word": ".",
                    "probability": 0.875
                }
            ]
        },
        {
            "id": 201,
            "seek": 119852,
            "start": 1327.86,
            "end": 1332.94,
            "text": " This takes big O of, sorry, Wasserstein takes big O of N to the fourth.",
            "tokens": [
                51316,
                639,
                2516,
                955,
                422,
                295,
                11,
                2597,
                11,
                17351,
                9089,
                2516,
                955,
                422,
                295,
                426,
                281,
                264,
                6409,
                13,
                51552
            ],
            "temperature": 0.0,
            "avg_logprob": -0.27234374105930326,
            "compression_ratio": 1.728448275862069,
            "no_speech_prob": 0.0074615478515625,
            "words": [
                {
                    "start": 1327.86,
                    "end": 1328.06,
                    "word": " This",
                    "probability": 0.73388671875
                },
                {
                    "start": 1328.06,
                    "end": 1328.44,
                    "word": " takes",
                    "probability": 0.72216796875
                },
                {
                    "start": 1328.44,
                    "end": 1328.92,
                    "word": " big",
                    "probability": 0.6396484375
                },
                {
                    "start": 1328.92,
                    "end": 1329.22,
                    "word": " O",
                    "probability": 0.85595703125
                },
                {
                    "start": 1329.22,
                    "end": 1329.52,
                    "word": " of",
                    "probability": 0.6982421875
                },
                {
                    "start": 1329.52,
                    "end": 1329.98,
                    "word": ",",
                    "probability": 0.490966796875
                },
                {
                    "start": 1329.98,
                    "end": 1330.34,
                    "word": " sorry",
                    "probability": 0.1781005859375
                },
                {
                    "start": 1330.34,
                    "end": 1330.52,
                    "word": ",",
                    "probability": 0.91943359375
                },
                {
                    "start": 1330.52,
                    "end": 1330.88,
                    "word": " Wasserstein",
                    "probability": 0.853271484375
                },
                {
                    "start": 1330.88,
                    "end": 1331.32,
                    "word": " takes",
                    "probability": 0.76611328125
                },
                {
                    "start": 1331.32,
                    "end": 1331.74,
                    "word": " big",
                    "probability": 0.83447265625
                },
                {
                    "start": 1331.74,
                    "end": 1331.88,
                    "word": " O",
                    "probability": 0.97216796875
                },
                {
                    "start": 1331.88,
                    "end": 1332.02,
                    "word": " of",
                    "probability": 0.654296875
                },
                {
                    "start": 1332.02,
                    "end": 1332.1,
                    "word": " N",
                    "probability": 0.58935546875
                },
                {
                    "start": 1332.1,
                    "end": 1332.18,
                    "word": " to",
                    "probability": 0.67724609375
                },
                {
                    "start": 1332.18,
                    "end": 1332.28,
                    "word": " the",
                    "probability": 0.80029296875
                },
                {
                    "start": 1332.28,
                    "end": 1332.46,
                    "word": " fourth",
                    "probability": 0.85498046875
                },
                {
                    "start": 1332.46,
                    "end": 1332.94,
                    "word": ".",
                    "probability": 0.95556640625
                }
            ]
        },
        {
            "id": 202,
            "seek": 119852,
            "start": 1333.44,
            "end": 1338.18,
            "text": " Wasserstein is the only other drift calculator that I really know",
            "tokens": [
                51600,
                17351,
                9089,
                307,
                264,
                787,
                661,
                19699,
                24993,
                300,
                286,
                534,
                458,
                51828
            ],
            "temperature": 0.0,
            "avg_logprob": -0.27234374105930326,
            "compression_ratio": 1.728448275862069,
            "no_speech_prob": 0.0074615478515625,
            "words": [
                {
                    "start": 1333.44,
                    "end": 1333.92,
                    "word": " Wasserstein",
                    "probability": 0.7137451171875
                },
                {
                    "start": 1333.92,
                    "end": 1334.1,
                    "word": " is",
                    "probability": 0.884765625
                },
                {
                    "start": 1334.1,
                    "end": 1334.26,
                    "word": " the",
                    "probability": 0.81689453125
                },
                {
                    "start": 1334.26,
                    "end": 1334.44,
                    "word": " only",
                    "probability": 0.85302734375
                },
                {
                    "start": 1334.44,
                    "end": 1334.9,
                    "word": " other",
                    "probability": 0.75927734375
                },
                {
                    "start": 1334.9,
                    "end": 1336.06,
                    "word": " drift",
                    "probability": 0.88232421875
                },
                {
                    "start": 1336.06,
                    "end": 1336.52,
                    "word": " calculator",
                    "probability": 0.86572265625
                },
                {
                    "start": 1336.52,
                    "end": 1336.8,
                    "word": " that",
                    "probability": 0.87939453125
                },
                {
                    "start": 1336.8,
                    "end": 1336.92,
                    "word": " I",
                    "probability": 0.998046875
                },
                {
                    "start": 1336.92,
                    "end": 1337.32,
                    "word": " really",
                    "probability": 0.3505859375
                },
                {
                    "start": 1337.32,
                    "end": 1338.18,
                    "word": " know",
                    "probability": 0.81396484375
                }
            ]
        },
        {
            "id": 203,
            "seek": 122594,
            "start": 1338.74,
            "end": 1345.28,
            "text": " works. And is model independent and doesn't have some funny business with",
            "tokens": [
                50372,
                1985,
                13,
                400,
                307,
                2316,
                6695,
                293,
                1177,
                380,
                362,
                512,
                4074,
                1606,
                365,
                50708
            ],
            "temperature": 0.0,
            "avg_logprob": -0.2041396111636967,
            "compression_ratio": 1.5339366515837105,
            "no_speech_prob": 0.0011692047119140625,
            "words": [
                {
                    "start": 1338.74,
                    "end": 1339.32,
                    "word": " works",
                    "probability": 0.1697998046875
                },
                {
                    "start": 1339.32,
                    "end": 1339.96,
                    "word": ".",
                    "probability": 0.15185546875
                },
                {
                    "start": 1340.74,
                    "end": 1341.02,
                    "word": " And",
                    "probability": 0.61572265625
                },
                {
                    "start": 1341.02,
                    "end": 1341.2,
                    "word": " is",
                    "probability": 0.81884765625
                },
                {
                    "start": 1341.2,
                    "end": 1341.72,
                    "word": " model",
                    "probability": 0.78515625
                },
                {
                    "start": 1341.72,
                    "end": 1342.28,
                    "word": " independent",
                    "probability": 0.54736328125
                },
                {
                    "start": 1342.28,
                    "end": 1342.8,
                    "word": " and",
                    "probability": 0.7646484375
                },
                {
                    "start": 1342.8,
                    "end": 1343.44,
                    "word": " doesn't",
                    "probability": 0.927001953125
                },
                {
                    "start": 1343.44,
                    "end": 1343.8,
                    "word": " have",
                    "probability": 0.89794921875
                },
                {
                    "start": 1343.8,
                    "end": 1344.42,
                    "word": " some",
                    "probability": 0.71923828125
                },
                {
                    "start": 1344.42,
                    "end": 1344.64,
                    "word": " funny",
                    "probability": 0.91455078125
                },
                {
                    "start": 1344.64,
                    "end": 1344.96,
                    "word": " business",
                    "probability": 0.88671875
                },
                {
                    "start": 1344.96,
                    "end": 1345.28,
                    "word": " with",
                    "probability": 0.8115234375
                }
            ]
        },
        {
            "id": 204,
            "seek": 122594,
            "start": 1345.28,
            "end": 1354.8,
            "text": " building an estimated distribution initially. This takes K log N. Another big difference between",
            "tokens": [
                50708,
                2390,
                364,
                14109,
                7316,
                9105,
                13,
                639,
                2516,
                591,
                3565,
                426,
                13,
                3996,
                955,
                2649,
                1296,
                51160
            ],
            "temperature": 0.0,
            "avg_logprob": -0.2041396111636967,
            "compression_ratio": 1.5339366515837105,
            "no_speech_prob": 0.0011692047119140625,
            "words": [
                {
                    "start": 1345.28,
                    "end": 1346.0,
                    "word": " building",
                    "probability": 0.78857421875
                },
                {
                    "start": 1346.0,
                    "end": 1346.18,
                    "word": " an",
                    "probability": 0.88720703125
                },
                {
                    "start": 1346.18,
                    "end": 1346.48,
                    "word": " estimated",
                    "probability": 0.77197265625
                },
                {
                    "start": 1346.48,
                    "end": 1346.98,
                    "word": " distribution",
                    "probability": 0.7451171875
                },
                {
                    "start": 1346.98,
                    "end": 1347.6,
                    "word": " initially",
                    "probability": 0.7197265625
                },
                {
                    "start": 1347.6,
                    "end": 1348.24,
                    "word": ".",
                    "probability": 0.94970703125
                },
                {
                    "start": 1349.78,
                    "end": 1350.42,
                    "word": " This",
                    "probability": 0.79052734375
                },
                {
                    "start": 1350.42,
                    "end": 1350.96,
                    "word": " takes",
                    "probability": 0.77392578125
                },
                {
                    "start": 1350.96,
                    "end": 1352.76,
                    "word": " K",
                    "probability": 0.438232421875
                },
                {
                    "start": 1352.76,
                    "end": 1352.96,
                    "word": " log",
                    "probability": 0.591796875
                },
                {
                    "start": 1352.96,
                    "end": 1353.14,
                    "word": " N",
                    "probability": 0.421630859375
                },
                {
                    "start": 1353.14,
                    "end": 1353.5,
                    "word": ".",
                    "probability": 0.92724609375
                },
                {
                    "start": 1353.5,
                    "end": 1353.94,
                    "word": " Another",
                    "probability": 0.81201171875
                },
                {
                    "start": 1353.94,
                    "end": 1354.2,
                    "word": " big",
                    "probability": 0.88671875
                },
                {
                    "start": 1354.2,
                    "end": 1354.5,
                    "word": " difference",
                    "probability": 0.7373046875
                },
                {
                    "start": 1354.5,
                    "end": 1354.8,
                    "word": " between",
                    "probability": 0.8037109375
                }
            ]
        },
        {
            "id": 205,
            "seek": 122594,
            "start": 1354.8,
            "end": 1359.32,
            "text": " Wasserstein is the traditional way of doing Wasserstein requires a relatively equal size",
            "tokens": [
                51160,
                17351,
                9089,
                307,
                264,
                5164,
                636,
                295,
                884,
                17351,
                9089,
                7029,
                257,
                7226,
                2681,
                2744,
                51392
            ],
            "temperature": 0.0,
            "avg_logprob": -0.2041396111636967,
            "compression_ratio": 1.5339366515837105,
            "no_speech_prob": 0.0011692047119140625,
            "words": [
                {
                    "start": 1354.8,
                    "end": 1355.14,
                    "word": " Wasserstein",
                    "probability": 0.818359375
                },
                {
                    "start": 1355.14,
                    "end": 1355.38,
                    "word": " is",
                    "probability": 0.791015625
                },
                {
                    "start": 1355.38,
                    "end": 1355.8,
                    "word": " the",
                    "probability": 0.72119140625
                },
                {
                    "start": 1355.8,
                    "end": 1356.2,
                    "word": " traditional",
                    "probability": 0.7373046875
                },
                {
                    "start": 1356.2,
                    "end": 1356.48,
                    "word": " way",
                    "probability": 0.91650390625
                },
                {
                    "start": 1356.48,
                    "end": 1356.58,
                    "word": " of",
                    "probability": 0.90478515625
                },
                {
                    "start": 1356.58,
                    "end": 1356.76,
                    "word": " doing",
                    "probability": 0.9130859375
                },
                {
                    "start": 1356.76,
                    "end": 1357.26,
                    "word": " Wasserstein",
                    "probability": 0.92041015625
                },
                {
                    "start": 1357.26,
                    "end": 1358.08,
                    "word": " requires",
                    "probability": 0.74462890625
                },
                {
                    "start": 1358.32,
                    "end": 1358.72,
                    "word": " a relatively",
                    "probability": 0.76123046875
                },
                {
                    "start": 1358.72,
                    "end": 1359.06,
                    "word": " equal",
                    "probability": 0.810546875
                },
                {
                    "start": 1359.06,
                    "end": 1359.32,
                    "word": " size",
                    "probability": 0.32568359375
                }
            ]
        },
        {
            "id": 206,
            "seek": 122594,
            "start": 1359.32,
            "end": 1366.16,
            "text": " test and training set. You can't have the test set be 0.5% of the training set.",
            "tokens": [
                51392,
                1500,
                293,
                3097,
                992,
                13,
                509,
                393,
                380,
                362,
                264,
                1500,
                992,
                312,
                1958,
                13,
                20,
                4,
                295,
                264,
                3097,
                992,
                13,
                51700
            ],
            "temperature": 0.0,
            "avg_logprob": -0.2041396111636967,
            "compression_ratio": 1.5339366515837105,
            "no_speech_prob": 0.0011692047119140625,
            "words": [
                {
                    "start": 1359.32,
                    "end": 1359.64,
                    "word": " test",
                    "probability": 0.7783203125
                },
                {
                    "start": 1359.64,
                    "end": 1359.8,
                    "word": " and",
                    "probability": 0.888671875
                },
                {
                    "start": 1359.8,
                    "end": 1360.12,
                    "word": " training",
                    "probability": 0.8134765625
                },
                {
                    "start": 1360.12,
                    "end": 1360.38,
                    "word": " set",
                    "probability": 0.87255859375
                },
                {
                    "start": 1360.38,
                    "end": 1361.02,
                    "word": ".",
                    "probability": 0.9287109375
                },
                {
                    "start": 1361.84,
                    "end": 1362.48,
                    "word": " You",
                    "probability": 0.88818359375
                },
                {
                    "start": 1362.48,
                    "end": 1362.78,
                    "word": " can't",
                    "probability": 0.944091796875
                },
                {
                    "start": 1362.78,
                    "end": 1362.92,
                    "word": " have",
                    "probability": 0.8896484375
                },
                {
                    "start": 1362.92,
                    "end": 1363.04,
                    "word": " the",
                    "probability": 0.81005859375
                },
                {
                    "start": 1363.04,
                    "end": 1363.26,
                    "word": " test",
                    "probability": 0.75048828125
                },
                {
                    "start": 1363.26,
                    "end": 1363.42,
                    "word": " set",
                    "probability": 0.7646484375
                },
                {
                    "start": 1363.42,
                    "end": 1363.66,
                    "word": " be",
                    "probability": 0.87744140625
                },
                {
                    "start": 1363.66,
                    "end": 1364.04,
                    "word": " 0",
                    "probability": 0.8544921875
                },
                {
                    "start": 1364.04,
                    "end": 1364.46,
                    "word": ".5",
                    "probability": 0.995361328125
                },
                {
                    "start": 1364.46,
                    "end": 1364.66,
                    "word": "%",
                    "probability": 0.7158203125
                },
                {
                    "start": 1364.66,
                    "end": 1364.96,
                    "word": " of",
                    "probability": 0.89794921875
                },
                {
                    "start": 1364.96,
                    "end": 1365.08,
                    "word": " the",
                    "probability": 0.8154296875
                },
                {
                    "start": 1365.08,
                    "end": 1365.28,
                    "word": " training",
                    "probability": 0.81689453125
                },
                {
                    "start": 1365.28,
                    "end": 1365.52,
                    "word": " set",
                    "probability": 0.89208984375
                },
                {
                    "start": 1365.52,
                    "end": 1366.16,
                    "word": ".",
                    "probability": 0.65869140625
                }
            ]
        },
        {
            "id": 207,
            "seek": 125240,
            "start": 1366.72,
            "end": 1373.34,
            "text": " That would just not work so great on the traditional method I know of. And this one,",
            "tokens": [
                50392,
                663,
                576,
                445,
                406,
                589,
                370,
                869,
                322,
                264,
                5164,
                3170,
                286,
                458,
                295,
                13,
                400,
                341,
                472,
                11,
                50684
            ],
            "temperature": 0.0,
            "avg_logprob": -0.2511160799435207,
            "compression_ratio": 1.6205357142857142,
            "no_speech_prob": 0.00844573974609375,
            "words": [
                {
                    "start": 1366.72,
                    "end": 1367.06,
                    "word": " That",
                    "probability": 0.178955078125
                },
                {
                    "start": 1367.06,
                    "end": 1367.3,
                    "word": " would",
                    "probability": 0.61767578125
                },
                {
                    "start": 1367.3,
                    "end": 1367.44,
                    "word": " just",
                    "probability": 0.75390625
                },
                {
                    "start": 1367.44,
                    "end": 1367.96,
                    "word": " not",
                    "probability": 0.33154296875
                },
                {
                    "start": 1367.96,
                    "end": 1368.46,
                    "word": " work",
                    "probability": 0.7880859375
                },
                {
                    "start": 1368.46,
                    "end": 1368.6,
                    "word": " so",
                    "probability": 0.8798828125
                },
                {
                    "start": 1368.6,
                    "end": 1368.84,
                    "word": " great",
                    "probability": 0.87841796875
                },
                {
                    "start": 1368.84,
                    "end": 1369.56,
                    "word": " on",
                    "probability": 0.80224609375
                },
                {
                    "start": 1369.56,
                    "end": 1369.68,
                    "word": " the",
                    "probability": 0.814453125
                },
                {
                    "start": 1369.68,
                    "end": 1370.1,
                    "word": " traditional",
                    "probability": 0.77685546875
                },
                {
                    "start": 1370.1,
                    "end": 1370.52,
                    "word": " method",
                    "probability": 0.75244140625
                },
                {
                    "start": 1370.52,
                    "end": 1370.72,
                    "word": " I",
                    "probability": 0.94775390625
                },
                {
                    "start": 1370.72,
                    "end": 1370.88,
                    "word": " know",
                    "probability": 0.8076171875
                },
                {
                    "start": 1370.88,
                    "end": 1371.1,
                    "word": " of",
                    "probability": 0.82861328125
                },
                {
                    "start": 1371.1,
                    "end": 1371.62,
                    "word": ".",
                    "probability": 0.89111328125
                },
                {
                    "start": 1371.74,
                    "end": 1372.26,
                    "word": " And",
                    "probability": 0.78662109375
                },
                {
                    "start": 1372.26,
                    "end": 1372.44,
                    "word": " this",
                    "probability": 0.8759765625
                },
                {
                    "start": 1372.44,
                    "end": 1372.74,
                    "word": " one",
                    "probability": 0.8154296875
                },
                {
                    "start": 1372.74,
                    "end": 1373.34,
                    "word": ",",
                    "probability": 0.9013671875
                }
            ]
        },
        {
            "id": 208,
            "seek": 125240,
            "start": 1373.34,
            "end": 1378.04,
            "text": " honestly, it's online. You can do, the test set can be a fraction of the size of the training",
            "tokens": [
                50716,
                6095,
                11,
                309,
                311,
                2950,
                13,
                509,
                393,
                360,
                11,
                264,
                1500,
                992,
                393,
                312,
                257,
                14135,
                295,
                264,
                2744,
                295,
                264,
                3097,
                50956
            ],
            "temperature": 0.0,
            "avg_logprob": -0.2511160799435207,
            "compression_ratio": 1.6205357142857142,
            "no_speech_prob": 0.00844573974609375,
            "words": [
                {
                    "start": 1373.34,
                    "end": 1373.68,
                    "word": " honestly",
                    "probability": 0.81787109375
                },
                {
                    "start": 1373.68,
                    "end": 1373.92,
                    "word": ",",
                    "probability": 0.8076171875
                },
                {
                    "start": 1373.92,
                    "end": 1374.14,
                    "word": " it's",
                    "probability": 0.94384765625
                },
                {
                    "start": 1374.14,
                    "end": 1374.52,
                    "word": " online",
                    "probability": 0.7998046875
                },
                {
                    "start": 1374.52,
                    "end": 1374.8,
                    "word": ".",
                    "probability": 0.6396484375
                },
                {
                    "start": 1374.8,
                    "end": 1374.88,
                    "word": " You",
                    "probability": 0.498046875
                },
                {
                    "start": 1374.88,
                    "end": 1375.04,
                    "word": " can",
                    "probability": 0.900390625
                },
                {
                    "start": 1375.04,
                    "end": 1375.3,
                    "word": " do",
                    "probability": 0.90283203125
                },
                {
                    "start": 1375.3,
                    "end": 1375.72,
                    "word": ",",
                    "probability": 0.31689453125
                },
                {
                    "start": 1375.72,
                    "end": 1375.82,
                    "word": " the",
                    "probability": 0.8037109375
                },
                {
                    "start": 1375.82,
                    "end": 1376.08,
                    "word": " test",
                    "probability": 0.8017578125
                },
                {
                    "start": 1376.08,
                    "end": 1376.2,
                    "word": " set",
                    "probability": 0.701171875
                },
                {
                    "start": 1376.2,
                    "end": 1376.34,
                    "word": " can",
                    "probability": 0.8779296875
                },
                {
                    "start": 1376.34,
                    "end": 1376.44,
                    "word": " be",
                    "probability": 0.89892578125
                },
                {
                    "start": 1376.6,
                    "end": 1376.88,
                    "word": " a fraction",
                    "probability": 0.7255859375
                },
                {
                    "start": 1376.88,
                    "end": 1377.04,
                    "word": " of",
                    "probability": 0.90576171875
                },
                {
                    "start": 1377.04,
                    "end": 1377.14,
                    "word": " the",
                    "probability": 0.81396484375
                },
                {
                    "start": 1377.14,
                    "end": 1377.44,
                    "word": " size",
                    "probability": 0.7958984375
                },
                {
                    "start": 1377.44,
                    "end": 1377.7,
                    "word": " of",
                    "probability": 0.9013671875
                },
                {
                    "start": 1377.7,
                    "end": 1377.82,
                    "word": " the",
                    "probability": 0.802734375
                },
                {
                    "start": 1377.82,
                    "end": 1378.04,
                    "word": " training",
                    "probability": 0.79296875
                }
            ]
        },
        {
            "id": 209,
            "seek": 125240,
            "start": 1378.04,
            "end": 1386.14,
            "text": " set. And because it's online, it can do it in real time. And this is stupidly fast. You can track",
            "tokens": [
                50956,
                992,
                13,
                400,
                570,
                309,
                311,
                2950,
                11,
                309,
                393,
                360,
                309,
                294,
                957,
                565,
                13,
                400,
                341,
                307,
                6631,
                356,
                2370,
                13,
                509,
                393,
                2837,
                51368
            ],
            "temperature": 0.0,
            "avg_logprob": -0.2511160799435207,
            "compression_ratio": 1.6205357142857142,
            "no_speech_prob": 0.00844573974609375,
            "words": [
                {
                    "start": 1378.04,
                    "end": 1378.28,
                    "word": " set",
                    "probability": 0.86572265625
                },
                {
                    "start": 1378.28,
                    "end": 1378.8,
                    "word": ".",
                    "probability": 0.8408203125
                },
                {
                    "start": 1378.8,
                    "end": 1379.06,
                    "word": " And",
                    "probability": 0.87158203125
                },
                {
                    "start": 1379.06,
                    "end": 1379.5,
                    "word": " because",
                    "probability": 0.78173828125
                },
                {
                    "start": 1379.5,
                    "end": 1379.68,
                    "word": " it's",
                    "probability": 0.948974609375
                },
                {
                    "start": 1379.68,
                    "end": 1379.96,
                    "word": " online",
                    "probability": 0.79052734375
                },
                {
                    "start": 1379.96,
                    "end": 1380.08,
                    "word": ",",
                    "probability": 0.92431640625
                },
                {
                    "start": 1380.08,
                    "end": 1380.2,
                    "word": " it",
                    "probability": 0.5966796875
                },
                {
                    "start": 1380.2,
                    "end": 1380.3,
                    "word": " can",
                    "probability": 0.90087890625
                },
                {
                    "start": 1380.3,
                    "end": 1380.58,
                    "word": " do",
                    "probability": 0.91357421875
                },
                {
                    "start": 1380.58,
                    "end": 1380.96,
                    "word": " it",
                    "probability": 0.91064453125
                },
                {
                    "start": 1380.96,
                    "end": 1381.22,
                    "word": " in",
                    "probability": 0.9033203125
                },
                {
                    "start": 1381.22,
                    "end": 1381.42,
                    "word": " real",
                    "probability": 0.8837890625
                },
                {
                    "start": 1381.42,
                    "end": 1381.72,
                    "word": " time",
                    "probability": 0.65234375
                },
                {
                    "start": 1381.72,
                    "end": 1382.24,
                    "word": ".",
                    "probability": 0.97314453125
                },
                {
                    "start": 1382.62,
                    "end": 1383.1,
                    "word": " And",
                    "probability": 0.8828125
                },
                {
                    "start": 1383.1,
                    "end": 1383.3,
                    "word": " this",
                    "probability": 0.8828125
                },
                {
                    "start": 1383.3,
                    "end": 1383.6,
                    "word": " is",
                    "probability": 0.90087890625
                },
                {
                    "start": 1383.6,
                    "end": 1384.84,
                    "word": " stupidly",
                    "probability": 0.906005859375
                },
                {
                    "start": 1384.84,
                    "end": 1385.24,
                    "word": " fast",
                    "probability": 0.82373046875
                },
                {
                    "start": 1385.24,
                    "end": 1385.36,
                    "word": ".",
                    "probability": 0.923828125
                },
                {
                    "start": 1385.36,
                    "end": 1385.44,
                    "word": " You",
                    "probability": 0.89453125
                },
                {
                    "start": 1385.44,
                    "end": 1385.62,
                    "word": " can",
                    "probability": 0.89697265625
                },
                {
                    "start": 1385.62,
                    "end": 1386.14,
                    "word": " track",
                    "probability": 0.869140625
                }
            ]
        },
        {
            "id": 210,
            "seek": 125240,
            "start": 1386.14,
            "end": 1392.62,
            "text": " for the Ember open source, Ember malware data set, you build a reasonable couple tree,",
            "tokens": [
                51368,
                337,
                264,
                3968,
                607,
                1269,
                4009,
                11,
                3968,
                607,
                40747,
                1412,
                992,
                11,
                291,
                1322,
                257,
                10585,
                1916,
                4230,
                11,
                51680
            ],
            "temperature": 0.0,
            "avg_logprob": -0.2511160799435207,
            "compression_ratio": 1.6205357142857142,
            "no_speech_prob": 0.00844573974609375,
            "words": [
                {
                    "start": 1386.14,
                    "end": 1386.68,
                    "word": " for",
                    "probability": 0.7919921875
                },
                {
                    "start": 1386.68,
                    "end": 1387.38,
                    "word": " the",
                    "probability": 0.78955078125
                },
                {
                    "start": 1387.38,
                    "end": 1387.64,
                    "word": " Ember",
                    "probability": 0.619873046875
                },
                {
                    "start": 1387.64,
                    "end": 1388.02,
                    "word": " open",
                    "probability": 0.486572265625
                },
                {
                    "start": 1388.02,
                    "end": 1388.46,
                    "word": " source",
                    "probability": 0.54443359375
                },
                {
                    "start": 1388.46,
                    "end": 1388.58,
                    "word": ",",
                    "probability": 0.442138671875
                },
                {
                    "start": 1388.58,
                    "end": 1388.84,
                    "word": " Ember",
                    "probability": 0.864501953125
                },
                {
                    "start": 1388.84,
                    "end": 1389.16,
                    "word": " malware",
                    "probability": 0.71923828125
                },
                {
                    "start": 1389.16,
                    "end": 1389.42,
                    "word": " data",
                    "probability": 0.4794921875
                },
                {
                    "start": 1389.42,
                    "end": 1389.8,
                    "word": " set",
                    "probability": 0.8935546875
                },
                {
                    "start": 1389.8,
                    "end": 1390.24,
                    "word": ",",
                    "probability": 0.27880859375
                },
                {
                    "start": 1390.24,
                    "end": 1390.98,
                    "word": " you",
                    "probability": 0.87744140625
                },
                {
                    "start": 1390.98,
                    "end": 1391.24,
                    "word": " build",
                    "probability": 0.8759765625
                },
                {
                    "start": 1391.38,
                    "end": 1391.74,
                    "word": " a reasonable",
                    "probability": 0.77978515625
                },
                {
                    "start": 1391.74,
                    "end": 1392.0,
                    "word": " couple",
                    "probability": 0.2022705078125
                },
                {
                    "start": 1392.0,
                    "end": 1392.32,
                    "word": " tree",
                    "probability": 0.438232421875
                },
                {
                    "start": 1392.32,
                    "end": 1392.62,
                    "word": ",",
                    "probability": 0.74755859375
                }
            ]
        },
        {
            "id": 211,
            "seek": 128124,
            "start": 1392.62,
            "end": 1401.58,
            "text": " you can track 16,000 new samples per second on my laptop. So that's fast enough that inference",
            "tokens": [
                50364,
                291,
                393,
                2837,
                3165,
                11,
                1360,
                777,
                10938,
                680,
                1150,
                322,
                452,
                10732,
                13,
                407,
                300,
                311,
                2370,
                1547,
                300,
                38253,
                50812
            ],
            "temperature": 0.0,
            "avg_logprob": -0.20570231467178188,
            "compression_ratio": 1.7219730941704037,
            "no_speech_prob": 0.00498199462890625,
            "words": [
                {
                    "start": 1392.62,
                    "end": 1392.76,
                    "word": " you",
                    "probability": 0.341796875
                },
                {
                    "start": 1392.76,
                    "end": 1392.92,
                    "word": " can",
                    "probability": 0.87890625
                },
                {
                    "start": 1392.92,
                    "end": 1393.18,
                    "word": " track",
                    "probability": 0.81005859375
                },
                {
                    "start": 1393.18,
                    "end": 1393.66,
                    "word": " 16",
                    "probability": 0.83544921875
                },
                {
                    "start": 1393.66,
                    "end": 1394.36,
                    "word": ",000",
                    "probability": 0.934814453125
                },
                {
                    "start": 1394.36,
                    "end": 1394.8,
                    "word": " new",
                    "probability": 0.7939453125
                },
                {
                    "start": 1394.8,
                    "end": 1395.18,
                    "word": " samples",
                    "probability": 0.76513671875
                },
                {
                    "start": 1395.18,
                    "end": 1395.8,
                    "word": " per",
                    "probability": 0.900390625
                },
                {
                    "start": 1395.8,
                    "end": 1396.28,
                    "word": " second",
                    "probability": 0.78955078125
                },
                {
                    "start": 1396.28,
                    "end": 1396.94,
                    "word": " on",
                    "probability": 0.8876953125
                },
                {
                    "start": 1396.94,
                    "end": 1397.18,
                    "word": " my",
                    "probability": 0.90380859375
                },
                {
                    "start": 1397.18,
                    "end": 1397.58,
                    "word": " laptop",
                    "probability": 0.77490234375
                },
                {
                    "start": 1397.58,
                    "end": 1398.1,
                    "word": ".",
                    "probability": 0.89013671875
                },
                {
                    "start": 1398.5,
                    "end": 1398.7,
                    "word": " So",
                    "probability": 0.72265625
                },
                {
                    "start": 1398.7,
                    "end": 1398.96,
                    "word": " that's",
                    "probability": 0.904541015625
                },
                {
                    "start": 1398.96,
                    "end": 1399.28,
                    "word": " fast",
                    "probability": 0.80712890625
                },
                {
                    "start": 1399.28,
                    "end": 1399.66,
                    "word": " enough",
                    "probability": 0.7880859375
                },
                {
                    "start": 1399.66,
                    "end": 1400.38,
                    "word": " that",
                    "probability": 0.84228515625
                },
                {
                    "start": 1400.38,
                    "end": 1401.58,
                    "word": " inference",
                    "probability": 0.80615234375
                }
            ]
        },
        {
            "id": 212,
            "seek": 128124,
            "start": 1401.58,
            "end": 1406.14,
            "text": " wouldn't be, this would not be the bottleneck. Inference would be the bottleneck if you would",
            "tokens": [
                50812,
                2759,
                380,
                312,
                11,
                341,
                576,
                406,
                312,
                264,
                44641,
                547,
                13,
                682,
                5158,
                576,
                312,
                264,
                44641,
                547,
                498,
                291,
                576,
                51032
            ],
            "temperature": 0.0,
            "avg_logprob": -0.20570231467178188,
            "compression_ratio": 1.7219730941704037,
            "no_speech_prob": 0.00498199462890625,
            "words": [
                {
                    "start": 1401.58,
                    "end": 1402.0,
                    "word": " wouldn't",
                    "probability": 0.814208984375
                },
                {
                    "start": 1402.0,
                    "end": 1402.16,
                    "word": " be",
                    "probability": 0.8828125
                },
                {
                    "start": 1402.16,
                    "end": 1402.4,
                    "word": ",",
                    "probability": 0.2264404296875
                },
                {
                    "start": 1402.4,
                    "end": 1402.84,
                    "word": " this",
                    "probability": 0.85498046875
                },
                {
                    "start": 1402.84,
                    "end": 1403.2,
                    "word": " would",
                    "probability": 0.76611328125
                },
                {
                    "start": 1403.2,
                    "end": 1403.36,
                    "word": " not",
                    "probability": 0.90087890625
                },
                {
                    "start": 1403.36,
                    "end": 1403.5,
                    "word": " be",
                    "probability": 0.90283203125
                },
                {
                    "start": 1403.5,
                    "end": 1403.62,
                    "word": " the",
                    "probability": 0.81494140625
                },
                {
                    "start": 1403.62,
                    "end": 1404.02,
                    "word": " bottleneck",
                    "probability": 0.92724609375
                },
                {
                    "start": 1404.02,
                    "end": 1404.16,
                    "word": ".",
                    "probability": 0.74365234375
                },
                {
                    "start": 1404.16,
                    "end": 1404.5,
                    "word": " Inference",
                    "probability": 0.856689453125
                },
                {
                    "start": 1404.5,
                    "end": 1404.88,
                    "word": " would",
                    "probability": 0.85693359375
                },
                {
                    "start": 1404.88,
                    "end": 1405.02,
                    "word": " be",
                    "probability": 0.90625
                },
                {
                    "start": 1405.02,
                    "end": 1405.22,
                    "word": " the",
                    "probability": 0.81494140625
                },
                {
                    "start": 1405.22,
                    "end": 1405.64,
                    "word": " bottleneck",
                    "probability": 0.897216796875
                },
                {
                    "start": 1405.64,
                    "end": 1405.78,
                    "word": " if",
                    "probability": 0.7685546875
                },
                {
                    "start": 1405.78,
                    "end": 1405.88,
                    "word": " you",
                    "probability": 0.888671875
                },
                {
                    "start": 1405.88,
                    "end": 1406.14,
                    "word": " would",
                    "probability": 0.128173828125
                }
            ]
        },
        {
            "id": 213,
            "seek": 128124,
            "start": 1406.9,
            "end": 1412.64,
            "text": " build a cloud distribution, a cloud model, and put this as a guard in front of that cloud model.",
            "tokens": [
                51092,
                1322,
                257,
                4588,
                7316,
                11,
                257,
                4588,
                2316,
                11,
                293,
                829,
                341,
                382,
                257,
                6290,
                294,
                1868,
                295,
                300,
                4588,
                2316,
                13,
                51340
            ],
            "temperature": 0.0,
            "avg_logprob": -0.20570231467178188,
            "compression_ratio": 1.7219730941704037,
            "no_speech_prob": 0.00498199462890625,
            "words": [
                {
                    "start": 1406.9,
                    "end": 1407.42,
                    "word": " build",
                    "probability": 0.54345703125
                },
                {
                    "start": 1407.56,
                    "end": 1407.74,
                    "word": " a cloud",
                    "probability": 0.8447265625
                },
                {
                    "start": 1407.74,
                    "end": 1408.24,
                    "word": " distribution",
                    "probability": 0.63525390625
                },
                {
                    "start": 1408.24,
                    "end": 1408.64,
                    "word": ",",
                    "probability": 0.93310546875
                },
                {
                    "start": 1408.8,
                    "end": 1409.06,
                    "word": " a cloud",
                    "probability": 0.896484375
                },
                {
                    "start": 1409.06,
                    "end": 1409.4,
                    "word": " model",
                    "probability": 0.8212890625
                },
                {
                    "start": 1409.4,
                    "end": 1409.64,
                    "word": ",",
                    "probability": 0.62255859375
                },
                {
                    "start": 1409.64,
                    "end": 1409.9,
                    "word": " and",
                    "probability": 0.89306640625
                },
                {
                    "start": 1409.9,
                    "end": 1410.08,
                    "word": " put",
                    "probability": 0.89501953125
                },
                {
                    "start": 1410.08,
                    "end": 1410.3,
                    "word": " this",
                    "probability": 0.8798828125
                },
                {
                    "start": 1410.3,
                    "end": 1410.44,
                    "word": " as",
                    "probability": 0.89892578125
                },
                {
                    "start": 1410.56,
                    "end": 1410.84,
                    "word": " a guard",
                    "probability": 0.82568359375
                },
                {
                    "start": 1410.84,
                    "end": 1411.04,
                    "word": " in",
                    "probability": 0.89892578125
                },
                {
                    "start": 1411.04,
                    "end": 1411.26,
                    "word": " front",
                    "probability": 0.9208984375
                },
                {
                    "start": 1411.26,
                    "end": 1411.38,
                    "word": " of",
                    "probability": 0.9052734375
                },
                {
                    "start": 1411.38,
                    "end": 1411.54,
                    "word": " that",
                    "probability": 0.873046875
                },
                {
                    "start": 1411.54,
                    "end": 1411.82,
                    "word": " cloud",
                    "probability": 0.9013671875
                },
                {
                    "start": 1411.82,
                    "end": 1412.12,
                    "word": " model",
                    "probability": 0.8232421875
                },
                {
                    "start": 1412.12,
                    "end": 1412.64,
                    "word": ".",
                    "probability": 0.99755859375
                }
            ]
        },
        {
            "id": 214,
            "seek": 128124,
            "start": 1413.82,
            "end": 1421.46,
            "text": " So we built it so that it can go fast enough to defend. And now here is where the thing originally",
            "tokens": [
                51436,
                407,
                321,
                3094,
                309,
                370,
                300,
                309,
                393,
                352,
                2370,
                1547,
                281,
                8602,
                13,
                400,
                586,
                510,
                307,
                689,
                264,
                551,
                7993,
                51808
            ],
            "temperature": 0.0,
            "avg_logprob": -0.20570231467178188,
            "compression_ratio": 1.7219730941704037,
            "no_speech_prob": 0.00498199462890625,
            "words": [
                {
                    "start": 1413.82,
                    "end": 1414.34,
                    "word": " So",
                    "probability": 0.9013671875
                },
                {
                    "start": 1414.34,
                    "end": 1414.6,
                    "word": " we",
                    "probability": 0.85693359375
                },
                {
                    "start": 1414.6,
                    "end": 1414.86,
                    "word": " built",
                    "probability": 0.78125
                },
                {
                    "start": 1414.86,
                    "end": 1415.0,
                    "word": " it",
                    "probability": 0.90576171875
                },
                {
                    "start": 1415.0,
                    "end": 1415.16,
                    "word": " so",
                    "probability": 0.88671875
                },
                {
                    "start": 1415.16,
                    "end": 1415.3,
                    "word": " that",
                    "probability": 0.8798828125
                },
                {
                    "start": 1415.3,
                    "end": 1415.66,
                    "word": " it",
                    "probability": 0.8994140625
                },
                {
                    "start": 1415.66,
                    "end": 1415.92,
                    "word": " can",
                    "probability": 0.88232421875
                },
                {
                    "start": 1415.92,
                    "end": 1416.16,
                    "word": " go",
                    "probability": 0.9091796875
                },
                {
                    "start": 1416.16,
                    "end": 1416.5,
                    "word": " fast",
                    "probability": 0.82080078125
                },
                {
                    "start": 1416.5,
                    "end": 1416.7,
                    "word": " enough",
                    "probability": 0.8056640625
                },
                {
                    "start": 1416.7,
                    "end": 1416.86,
                    "word": " to",
                    "probability": 0.90673828125
                },
                {
                    "start": 1416.86,
                    "end": 1417.28,
                    "word": " defend",
                    "probability": 0.81103515625
                },
                {
                    "start": 1417.28,
                    "end": 1417.8,
                    "word": ".",
                    "probability": 0.955078125
                },
                {
                    "start": 1418.66,
                    "end": 1418.82,
                    "word": " And",
                    "probability": 0.74560546875
                },
                {
                    "start": 1418.82,
                    "end": 1418.96,
                    "word": " now",
                    "probability": 0.888671875
                },
                {
                    "start": 1418.96,
                    "end": 1419.28,
                    "word": " here",
                    "probability": 0.7470703125
                },
                {
                    "start": 1419.28,
                    "end": 1419.46,
                    "word": " is",
                    "probability": 0.86767578125
                },
                {
                    "start": 1419.46,
                    "end": 1419.86,
                    "word": " where",
                    "probability": 0.90673828125
                },
                {
                    "start": 1419.86,
                    "end": 1420.5,
                    "word": " the",
                    "probability": 0.81201171875
                },
                {
                    "start": 1420.5,
                    "end": 1421.14,
                    "word": " thing",
                    "probability": 0.6083984375
                },
                {
                    "start": 1421.14,
                    "end": 1421.46,
                    "word": " originally",
                    "probability": 0.6943359375
                }
            ]
        },
        {
            "id": 215,
            "seek": 131124,
            "start": 1421.46,
            "end": 1428.72,
            "text": " came from, and then we backtracked onto Drift. So here's the test set attack. So this was",
            "tokens": [
                50364,
                1361,
                490,
                11,
                293,
                550,
                321,
                646,
                19466,
                292,
                3911,
                2491,
                2008,
                13,
                407,
                510,
                311,
                264,
                1500,
                992,
                2690,
                13,
                407,
                341,
                390,
                50720
            ],
            "temperature": 0.0,
            "avg_logprob": -0.25598403430999594,
            "compression_ratio": 1.5617021276595744,
            "no_speech_prob": 0.00970458984375,
            "words": [
                {
                    "start": 1421.46,
                    "end": 1421.72,
                    "word": " came",
                    "probability": 0.55859375
                },
                {
                    "start": 1421.72,
                    "end": 1421.94,
                    "word": " from",
                    "probability": 0.82275390625
                },
                {
                    "start": 1421.94,
                    "end": 1422.08,
                    "word": ",",
                    "probability": 0.203369140625
                },
                {
                    "start": 1422.08,
                    "end": 1422.08,
                    "word": " and",
                    "probability": 0.87451171875
                },
                {
                    "start": 1422.08,
                    "end": 1422.3,
                    "word": " then",
                    "probability": 0.5908203125
                },
                {
                    "start": 1422.3,
                    "end": 1422.72,
                    "word": " we",
                    "probability": 0.89599609375
                },
                {
                    "start": 1422.72,
                    "end": 1423.32,
                    "word": " backtracked",
                    "probability": 0.8108723958333334
                },
                {
                    "start": 1423.32,
                    "end": 1423.46,
                    "word": " onto",
                    "probability": 0.65771484375
                },
                {
                    "start": 1423.46,
                    "end": 1423.82,
                    "word": " Drift",
                    "probability": 0.654052734375
                },
                {
                    "start": 1423.82,
                    "end": 1424.42,
                    "word": ".",
                    "probability": 0.98388671875
                },
                {
                    "start": 1424.48,
                    "end": 1424.92,
                    "word": " So",
                    "probability": 0.78955078125
                },
                {
                    "start": 1424.92,
                    "end": 1425.26,
                    "word": " here's",
                    "probability": 0.85400390625
                },
                {
                    "start": 1425.26,
                    "end": 1425.4,
                    "word": " the",
                    "probability": 0.8154296875
                },
                {
                    "start": 1425.4,
                    "end": 1425.68,
                    "word": " test",
                    "probability": 0.71826171875
                },
                {
                    "start": 1425.68,
                    "end": 1425.8,
                    "word": " set",
                    "probability": 0.77685546875
                },
                {
                    "start": 1425.8,
                    "end": 1426.18,
                    "word": " attack",
                    "probability": 0.8291015625
                },
                {
                    "start": 1426.18,
                    "end": 1426.78,
                    "word": ".",
                    "probability": 0.89013671875
                },
                {
                    "start": 1427.4,
                    "end": 1428.0,
                    "word": " So",
                    "probability": 0.62109375
                },
                {
                    "start": 1428.0,
                    "end": 1428.44,
                    "word": " this",
                    "probability": 0.802734375
                },
                {
                    "start": 1428.44,
                    "end": 1428.72,
                    "word": " was",
                    "probability": 0.8896484375
                }
            ]
        },
        {
            "id": 216,
            "seek": 131124,
            "start": 1428.72,
            "end": 1437.12,
            "text": " originally categorized by Ian Goodfellow, and some defenses have been proposed by Nicholas Carlini",
            "tokens": [
                50748,
                7993,
                19250,
                1602,
                538,
                19595,
                2205,
                69,
                21348,
                11,
                293,
                512,
                35989,
                362,
                668,
                10348,
                538,
                22924,
                14256,
                3812,
                51148
            ],
            "temperature": 0.0,
            "avg_logprob": -0.25598403430999594,
            "compression_ratio": 1.5617021276595744,
            "no_speech_prob": 0.00970458984375,
            "words": [
                {
                    "start": 1428.72,
                    "end": 1429.72,
                    "word": " originally",
                    "probability": 0.5908203125
                },
                {
                    "start": 1429.72,
                    "end": 1431.36,
                    "word": " categorized",
                    "probability": 0.866455078125
                },
                {
                    "start": 1431.36,
                    "end": 1431.9,
                    "word": " by",
                    "probability": 0.9091796875
                },
                {
                    "start": 1431.9,
                    "end": 1432.28,
                    "word": " Ian",
                    "probability": 0.8818359375
                },
                {
                    "start": 1432.28,
                    "end": 1432.68,
                    "word": " Goodfellow",
                    "probability": 0.92919921875
                },
                {
                    "start": 1432.68,
                    "end": 1433.3,
                    "word": ",",
                    "probability": 0.56787109375
                },
                {
                    "start": 1433.3,
                    "end": 1433.94,
                    "word": " and",
                    "probability": 0.89453125
                },
                {
                    "start": 1433.94,
                    "end": 1434.2,
                    "word": " some",
                    "probability": 0.671875
                },
                {
                    "start": 1434.2,
                    "end": 1434.58,
                    "word": " defenses",
                    "probability": 0.81396484375
                },
                {
                    "start": 1434.58,
                    "end": 1434.82,
                    "word": " have",
                    "probability": 0.85791015625
                },
                {
                    "start": 1434.82,
                    "end": 1434.96,
                    "word": " been",
                    "probability": 0.8154296875
                },
                {
                    "start": 1434.96,
                    "end": 1435.46,
                    "word": " proposed",
                    "probability": 0.8095703125
                },
                {
                    "start": 1435.46,
                    "end": 1436.12,
                    "word": " by",
                    "probability": 0.90625
                },
                {
                    "start": 1436.12,
                    "end": 1436.68,
                    "word": " Nicholas",
                    "probability": 0.353515625
                },
                {
                    "start": 1436.68,
                    "end": 1437.12,
                    "word": " Carlini",
                    "probability": 0.716552734375
                }
            ]
        },
        {
            "id": 217,
            "seek": 131124,
            "start": 1438.14,
            "end": 1447.12,
            "text": " and others. So, but here is the gist of the attack. Normal users just query. They don't",
            "tokens": [
                51208,
                293,
                2357,
                13,
                407,
                11,
                457,
                510,
                307,
                264,
                290,
                468,
                295,
                264,
                2690,
                13,
                21277,
                5022,
                445,
                14581,
                13,
                814,
                500,
                380,
                51644
            ],
            "temperature": 0.0,
            "avg_logprob": -0.25598403430999594,
            "compression_ratio": 1.5617021276595744,
            "no_speech_prob": 0.00970458984375,
            "words": [
                {
                    "start": 1438.14,
                    "end": 1438.74,
                    "word": " and",
                    "probability": 0.72802734375
                },
                {
                    "start": 1438.74,
                    "end": 1439.14,
                    "word": " others",
                    "probability": 0.7705078125
                },
                {
                    "start": 1439.14,
                    "end": 1439.74,
                    "word": ".",
                    "probability": 0.86865234375
                },
                {
                    "start": 1441.54,
                    "end": 1442.14,
                    "word": " So",
                    "probability": 0.459228515625
                },
                {
                    "start": 1442.14,
                    "end": 1442.58,
                    "word": ",",
                    "probability": 0.49462890625
                },
                {
                    "start": 1442.58,
                    "end": 1442.82,
                    "word": " but",
                    "probability": 0.82275390625
                },
                {
                    "start": 1442.82,
                    "end": 1443.28,
                    "word": " here",
                    "probability": 0.77294921875
                },
                {
                    "start": 1443.28,
                    "end": 1443.48,
                    "word": " is",
                    "probability": 0.50927734375
                },
                {
                    "start": 1443.48,
                    "end": 1443.78,
                    "word": " the",
                    "probability": 0.80029296875
                },
                {
                    "start": 1443.78,
                    "end": 1444.12,
                    "word": " gist",
                    "probability": 0.954833984375
                },
                {
                    "start": 1444.12,
                    "end": 1444.24,
                    "word": " of",
                    "probability": 0.90625
                },
                {
                    "start": 1444.24,
                    "end": 1444.34,
                    "word": " the",
                    "probability": 0.8203125
                },
                {
                    "start": 1444.34,
                    "end": 1444.64,
                    "word": " attack",
                    "probability": 0.84375
                },
                {
                    "start": 1444.64,
                    "end": 1445.16,
                    "word": ".",
                    "probability": 0.94970703125
                },
                {
                    "start": 1445.16,
                    "end": 1445.48,
                    "word": " Normal",
                    "probability": 0.80078125
                },
                {
                    "start": 1445.48,
                    "end": 1445.9,
                    "word": " users",
                    "probability": 0.84521484375
                },
                {
                    "start": 1445.9,
                    "end": 1446.16,
                    "word": " just",
                    "probability": 0.8681640625
                },
                {
                    "start": 1446.16,
                    "end": 1446.42,
                    "word": " query",
                    "probability": 0.86572265625
                },
                {
                    "start": 1446.42,
                    "end": 1446.72,
                    "word": ".",
                    "probability": 0.77099609375
                },
                {
                    "start": 1446.72,
                    "end": 1446.92,
                    "word": " They",
                    "probability": 0.734375
                },
                {
                    "start": 1446.92,
                    "end": 1447.12,
                    "word": " don't",
                    "probability": 0.947021484375
                }
            ]
        },
        {
            "id": 218,
            "seek": 131124,
            "start": 1447.12,
            "end": 1450.84,
            "text": " care about whether they got a false positive or a false negative, they just keep querying.",
            "tokens": [
                51644,
                1127,
                466,
                1968,
                436,
                658,
                257,
                7908,
                3353,
                420,
                257,
                7908,
                3671,
                11,
                436,
                445,
                1066,
                7083,
                1840,
                13,
                51820
            ],
            "temperature": 0.0,
            "avg_logprob": -0.25598403430999594,
            "compression_ratio": 1.5617021276595744,
            "no_speech_prob": 0.00970458984375,
            "words": [
                {
                    "start": 1447.12,
                    "end": 1447.38,
                    "word": " care",
                    "probability": 0.80224609375
                },
                {
                    "start": 1447.38,
                    "end": 1447.6,
                    "word": " about",
                    "probability": 0.802734375
                },
                {
                    "start": 1447.6,
                    "end": 1447.8,
                    "word": " whether",
                    "probability": 0.84375
                },
                {
                    "start": 1447.8,
                    "end": 1447.98,
                    "word": " they",
                    "probability": 0.7392578125
                },
                {
                    "start": 1447.98,
                    "end": 1448.1,
                    "word": " got",
                    "probability": 0.791015625
                },
                {
                    "start": 1448.24,
                    "end": 1448.38,
                    "word": " a false",
                    "probability": 0.84814453125
                },
                {
                    "start": 1448.38,
                    "end": 1448.68,
                    "word": " positive",
                    "probability": 0.84423828125
                },
                {
                    "start": 1448.68,
                    "end": 1448.86,
                    "word": " or",
                    "probability": 0.88916015625
                },
                {
                    "start": 1448.96,
                    "end": 1449.08,
                    "word": " a false",
                    "probability": 0.85400390625
                },
                {
                    "start": 1449.08,
                    "end": 1449.42,
                    "word": " negative",
                    "probability": 0.87744140625
                },
                {
                    "start": 1449.42,
                    "end": 1449.58,
                    "word": ",",
                    "probability": 0.489501953125
                },
                {
                    "start": 1449.58,
                    "end": 1449.8,
                    "word": " they",
                    "probability": 0.732421875
                },
                {
                    "start": 1449.8,
                    "end": 1449.96,
                    "word": " just",
                    "probability": 0.88232421875
                },
                {
                    "start": 1449.96,
                    "end": 1450.16,
                    "word": " keep",
                    "probability": 0.8115234375
                },
                {
                    "start": 1450.16,
                    "end": 1450.54,
                    "word": " querying",
                    "probability": 0.93603515625
                },
                {
                    "start": 1450.54,
                    "end": 1450.84,
                    "word": ".",
                    "probability": 0.97705078125
                }
            ]
        },
        {
            "id": 219,
            "seek": 134062,
            "start": 1451.82,
            "end": 1458.28,
            "text": " And for us, we think of a normal user as being like an enterprise user, trying to figure out",
            "tokens": [
                50400,
                400,
                337,
                505,
                11,
                321,
                519,
                295,
                257,
                2710,
                4195,
                382,
                885,
                411,
                364,
                14132,
                4195,
                11,
                1382,
                281,
                2573,
                484,
                50700
            ],
            "temperature": 0.0,
            "avg_logprob": -0.2185763920346896,
            "compression_ratio": 1.7207207207207207,
            "no_speech_prob": 0.004467010498046875,
            "words": [
                {
                    "start": 1451.82,
                    "end": 1452.42,
                    "word": " And",
                    "probability": 0.47802734375
                },
                {
                    "start": 1452.42,
                    "end": 1452.82,
                    "word": " for",
                    "probability": 0.8828125
                },
                {
                    "start": 1452.82,
                    "end": 1453.14,
                    "word": " us",
                    "probability": 0.90380859375
                },
                {
                    "start": 1453.14,
                    "end": 1453.5,
                    "word": ",",
                    "probability": 0.67431640625
                },
                {
                    "start": 1453.5,
                    "end": 1454.24,
                    "word": " we",
                    "probability": 0.88720703125
                },
                {
                    "start": 1454.24,
                    "end": 1454.42,
                    "word": " think",
                    "probability": 0.65576171875
                },
                {
                    "start": 1454.42,
                    "end": 1454.54,
                    "word": " of",
                    "probability": 0.88525390625
                },
                {
                    "start": 1454.64,
                    "end": 1454.8,
                    "word": " a normal",
                    "probability": 0.8056640625
                },
                {
                    "start": 1454.8,
                    "end": 1455.2,
                    "word": " user",
                    "probability": 0.81005859375
                },
                {
                    "start": 1455.2,
                    "end": 1455.38,
                    "word": " as",
                    "probability": 0.86962890625
                },
                {
                    "start": 1455.38,
                    "end": 1455.56,
                    "word": " being",
                    "probability": 0.85693359375
                },
                {
                    "start": 1455.56,
                    "end": 1455.78,
                    "word": " like",
                    "probability": 0.67041015625
                },
                {
                    "start": 1455.78,
                    "end": 1456.04,
                    "word": " an",
                    "probability": 0.85595703125
                },
                {
                    "start": 1456.04,
                    "end": 1456.52,
                    "word": " enterprise",
                    "probability": 0.73974609375
                },
                {
                    "start": 1456.52,
                    "end": 1457.12,
                    "word": " user",
                    "probability": 0.82275390625
                },
                {
                    "start": 1457.12,
                    "end": 1457.24,
                    "word": ",",
                    "probability": 0.2724609375
                },
                {
                    "start": 1457.24,
                    "end": 1457.68,
                    "word": " trying",
                    "probability": 0.79150390625
                },
                {
                    "start": 1457.68,
                    "end": 1457.86,
                    "word": " to",
                    "probability": 0.9013671875
                },
                {
                    "start": 1457.86,
                    "end": 1458.04,
                    "word": " figure",
                    "probability": 0.9208984375
                },
                {
                    "start": 1458.04,
                    "end": 1458.28,
                    "word": " out",
                    "probability": 0.82421875
                }
            ]
        },
        {
            "id": 220,
            "seek": 134062,
            "start": 1458.28,
            "end": 1464.8,
            "text": " whether all the reports coming over the wire of new binaries are malicious or benign, or they'd",
            "tokens": [
                50700,
                1968,
                439,
                264,
                7122,
                1348,
                670,
                264,
                6234,
                295,
                777,
                5171,
                4889,
                366,
                33496,
                420,
                3271,
                788,
                11,
                420,
                436,
                1116,
                51024
            ],
            "temperature": 0.0,
            "avg_logprob": -0.2185763920346896,
            "compression_ratio": 1.7207207207207207,
            "no_speech_prob": 0.004467010498046875,
            "words": [
                {
                    "start": 1458.28,
                    "end": 1458.54,
                    "word": " whether",
                    "probability": 0.853515625
                },
                {
                    "start": 1458.54,
                    "end": 1459.28,
                    "word": " all",
                    "probability": 0.87646484375
                },
                {
                    "start": 1459.28,
                    "end": 1459.46,
                    "word": " the",
                    "probability": 0.78271484375
                },
                {
                    "start": 1459.46,
                    "end": 1459.94,
                    "word": " reports",
                    "probability": 0.779296875
                },
                {
                    "start": 1459.94,
                    "end": 1460.22,
                    "word": " coming",
                    "probability": 0.74755859375
                },
                {
                    "start": 1460.22,
                    "end": 1460.38,
                    "word": " over",
                    "probability": 0.8798828125
                },
                {
                    "start": 1460.38,
                    "end": 1460.54,
                    "word": " the",
                    "probability": 0.81689453125
                },
                {
                    "start": 1460.54,
                    "end": 1460.86,
                    "word": " wire",
                    "probability": 0.7998046875
                },
                {
                    "start": 1460.86,
                    "end": 1461.26,
                    "word": " of",
                    "probability": 0.72900390625
                },
                {
                    "start": 1461.26,
                    "end": 1461.64,
                    "word": " new",
                    "probability": 0.259033203125
                },
                {
                    "start": 1461.64,
                    "end": 1462.7,
                    "word": " binaries",
                    "probability": 0.805419921875
                },
                {
                    "start": 1462.7,
                    "end": 1463.12,
                    "word": " are",
                    "probability": 0.8876953125
                },
                {
                    "start": 1463.12,
                    "end": 1463.54,
                    "word": " malicious",
                    "probability": 0.85302734375
                },
                {
                    "start": 1463.54,
                    "end": 1463.8,
                    "word": " or",
                    "probability": 0.884765625
                },
                {
                    "start": 1463.8,
                    "end": 1464.1,
                    "word": " benign",
                    "probability": 0.947265625
                },
                {
                    "start": 1464.1,
                    "end": 1464.42,
                    "word": ",",
                    "probability": 0.68701171875
                },
                {
                    "start": 1464.42,
                    "end": 1464.6,
                    "word": " or",
                    "probability": 0.9033203125
                },
                {
                    "start": 1464.6,
                    "end": 1464.8,
                    "word": " they'd",
                    "probability": 0.755859375
                }
            ]
        },
        {
            "id": 221,
            "seek": 134062,
            "start": 1464.8,
            "end": 1471.56,
            "text": " be submitting every single binary that they see on their network to a service that will tell them",
            "tokens": [
                51024,
                312,
                31836,
                633,
                2167,
                17434,
                300,
                436,
                536,
                322,
                641,
                3209,
                281,
                257,
                2643,
                300,
                486,
                980,
                552,
                51360
            ],
            "temperature": 0.0,
            "avg_logprob": -0.2185763920346896,
            "compression_ratio": 1.7207207207207207,
            "no_speech_prob": 0.004467010498046875,
            "words": [
                {
                    "start": 1464.8,
                    "end": 1464.92,
                    "word": " be",
                    "probability": 0.9091796875
                },
                {
                    "start": 1464.92,
                    "end": 1465.22,
                    "word": " submitting",
                    "probability": 0.8564453125
                },
                {
                    "start": 1465.22,
                    "end": 1465.56,
                    "word": " every",
                    "probability": 0.75146484375
                },
                {
                    "start": 1465.56,
                    "end": 1465.78,
                    "word": " single",
                    "probability": 0.89599609375
                },
                {
                    "start": 1465.78,
                    "end": 1466.08,
                    "word": " binary",
                    "probability": 0.7509765625
                },
                {
                    "start": 1466.08,
                    "end": 1466.3,
                    "word": " that",
                    "probability": 0.8740234375
                },
                {
                    "start": 1466.3,
                    "end": 1466.44,
                    "word": " they",
                    "probability": 0.7333984375
                },
                {
                    "start": 1466.44,
                    "end": 1466.62,
                    "word": " see",
                    "probability": 0.82177734375
                },
                {
                    "start": 1466.62,
                    "end": 1466.76,
                    "word": " on",
                    "probability": 0.904296875
                },
                {
                    "start": 1466.76,
                    "end": 1466.88,
                    "word": " their",
                    "probability": 0.87890625
                },
                {
                    "start": 1466.88,
                    "end": 1467.3,
                    "word": " network",
                    "probability": 0.845703125
                },
                {
                    "start": 1467.3,
                    "end": 1467.86,
                    "word": " to",
                    "probability": 0.8505859375
                },
                {
                    "start": 1468.08,
                    "end": 1468.52,
                    "word": " a service",
                    "probability": 0.8017578125
                },
                {
                    "start": 1468.52,
                    "end": 1470.44,
                    "word": " that",
                    "probability": 0.411865234375
                },
                {
                    "start": 1470.44,
                    "end": 1471.18,
                    "word": " will",
                    "probability": 0.81884765625
                },
                {
                    "start": 1471.18,
                    "end": 1471.38,
                    "word": " tell",
                    "probability": 0.83056640625
                },
                {
                    "start": 1471.38,
                    "end": 1471.56,
                    "word": " them",
                    "probability": 0.72216796875
                }
            ]
        },
        {
            "id": 222,
            "seek": 134062,
            "start": 1471.56,
            "end": 1480.84,
            "text": " whether it was malicious or benign for an ADV product. But a malicious user would be submitting",
            "tokens": [
                51360,
                1968,
                309,
                390,
                33496,
                420,
                3271,
                788,
                337,
                364,
                9135,
                53,
                1674,
                13,
                583,
                257,
                33496,
                4195,
                576,
                312,
                31836,
                51844
            ],
            "temperature": 0.0,
            "avg_logprob": -0.2185763920346896,
            "compression_ratio": 1.7207207207207207,
            "no_speech_prob": 0.004467010498046875,
            "words": [
                {
                    "start": 1471.56,
                    "end": 1471.7,
                    "word": " whether",
                    "probability": 0.8330078125
                },
                {
                    "start": 1471.7,
                    "end": 1471.86,
                    "word": " it",
                    "probability": 0.88916015625
                },
                {
                    "start": 1471.86,
                    "end": 1471.98,
                    "word": " was",
                    "probability": 0.8916015625
                },
                {
                    "start": 1471.98,
                    "end": 1472.3,
                    "word": " malicious",
                    "probability": 0.8427734375
                },
                {
                    "start": 1472.3,
                    "end": 1472.56,
                    "word": " or",
                    "probability": 0.888671875
                },
                {
                    "start": 1472.56,
                    "end": 1472.86,
                    "word": " benign",
                    "probability": 0.9658203125
                },
                {
                    "start": 1472.86,
                    "end": 1473.44,
                    "word": " for",
                    "probability": 0.72900390625
                },
                {
                    "start": 1473.44,
                    "end": 1474.78,
                    "word": " an",
                    "probability": 0.8583984375
                },
                {
                    "start": 1474.78,
                    "end": 1475.24,
                    "word": " ADV",
                    "probability": 0.673828125
                },
                {
                    "start": 1475.24,
                    "end": 1475.56,
                    "word": " product",
                    "probability": 0.77880859375
                },
                {
                    "start": 1475.56,
                    "end": 1476.16,
                    "word": ".",
                    "probability": 0.99072265625
                },
                {
                    "start": 1477.6,
                    "end": 1478.2,
                    "word": " But",
                    "probability": 0.8583984375
                },
                {
                    "start": 1478.5,
                    "end": 1478.72,
                    "word": " a malicious",
                    "probability": 0.85302734375
                },
                {
                    "start": 1478.72,
                    "end": 1479.18,
                    "word": " user",
                    "probability": 0.8037109375
                },
                {
                    "start": 1479.18,
                    "end": 1479.48,
                    "word": " would",
                    "probability": 0.845703125
                },
                {
                    "start": 1479.48,
                    "end": 1479.98,
                    "word": " be",
                    "probability": 0.90966796875
                },
                {
                    "start": 1479.98,
                    "end": 1480.84,
                    "word": " submitting",
                    "probability": 0.86474609375
                }
            ]
        },
        {
            "id": 223,
            "seek": 136338,
            "start": 1480.84,
            "end": 1487.52,
            "text": " things with a specific goal in mind. They don't care about what they care deeply what the label",
            "tokens": [
                50376,
                721,
                365,
                257,
                2685,
                3387,
                294,
                1575,
                13,
                814,
                500,
                380,
                1127,
                466,
                437,
                436,
                1127,
                8760,
                437,
                264,
                7645,
                50700
            ],
            "temperature": 0.0,
            "avg_logprob": -0.22569444957706664,
            "compression_ratio": 1.7383177570093458,
            "no_speech_prob": 0.0182647705078125,
            "words": [
                {
                    "start": 1480.84,
                    "end": 1481.42,
                    "word": " things",
                    "probability": 0.2435302734375
                },
                {
                    "start": 1481.42,
                    "end": 1482.32,
                    "word": " with",
                    "probability": 0.7841796875
                },
                {
                    "start": 1482.48,
                    "end": 1482.76,
                    "word": " a specific",
                    "probability": 0.84521484375
                },
                {
                    "start": 1482.76,
                    "end": 1483.14,
                    "word": " goal",
                    "probability": 0.91845703125
                },
                {
                    "start": 1483.14,
                    "end": 1483.3,
                    "word": " in",
                    "probability": 0.90576171875
                },
                {
                    "start": 1483.3,
                    "end": 1483.48,
                    "word": " mind",
                    "probability": 0.80859375
                },
                {
                    "start": 1483.48,
                    "end": 1483.64,
                    "word": ".",
                    "probability": 0.73583984375
                },
                {
                    "start": 1483.64,
                    "end": 1483.78,
                    "word": " They",
                    "probability": 0.724609375
                },
                {
                    "start": 1483.78,
                    "end": 1484.14,
                    "word": " don't",
                    "probability": 0.8330078125
                },
                {
                    "start": 1484.14,
                    "end": 1484.66,
                    "word": " care",
                    "probability": 0.84814453125
                },
                {
                    "start": 1484.66,
                    "end": 1485.18,
                    "word": " about",
                    "probability": 0.69873046875
                },
                {
                    "start": 1485.18,
                    "end": 1485.54,
                    "word": " what",
                    "probability": 0.78125
                },
                {
                    "start": 1485.54,
                    "end": 1485.9,
                    "word": " they",
                    "probability": 0.4423828125
                },
                {
                    "start": 1485.9,
                    "end": 1486.38,
                    "word": " care",
                    "probability": 0.8759765625
                },
                {
                    "start": 1486.38,
                    "end": 1486.88,
                    "word": " deeply",
                    "probability": 0.72705078125
                },
                {
                    "start": 1486.88,
                    "end": 1487.18,
                    "word": " what",
                    "probability": 0.67236328125
                },
                {
                    "start": 1487.18,
                    "end": 1487.34,
                    "word": " the",
                    "probability": 0.473876953125
                },
                {
                    "start": 1487.34,
                    "end": 1487.52,
                    "word": " label",
                    "probability": 0.86181640625
                }
            ]
        },
        {
            "id": 224,
            "seek": 136338,
            "start": 1487.52,
            "end": 1493.14,
            "text": " is, and they care deeply about defining a false negative. They want to find a malicious sample",
            "tokens": [
                50700,
                307,
                11,
                293,
                436,
                1127,
                8760,
                466,
                17827,
                257,
                7908,
                3671,
                13,
                814,
                528,
                281,
                915,
                257,
                33496,
                6889,
                50976
            ],
            "temperature": 0.0,
            "avg_logprob": -0.22569444957706664,
            "compression_ratio": 1.7383177570093458,
            "no_speech_prob": 0.0182647705078125,
            "words": [
                {
                    "start": 1487.52,
                    "end": 1487.82,
                    "word": " is",
                    "probability": 0.9052734375
                },
                {
                    "start": 1487.82,
                    "end": 1487.84,
                    "word": ",",
                    "probability": 0.405517578125
                },
                {
                    "start": 1487.84,
                    "end": 1488.16,
                    "word": " and",
                    "probability": 0.87451171875
                },
                {
                    "start": 1488.16,
                    "end": 1488.28,
                    "word": " they",
                    "probability": 0.7236328125
                },
                {
                    "start": 1488.28,
                    "end": 1488.84,
                    "word": " care",
                    "probability": 0.732421875
                },
                {
                    "start": 1488.84,
                    "end": 1489.34,
                    "word": " deeply",
                    "probability": 0.7900390625
                },
                {
                    "start": 1489.34,
                    "end": 1489.98,
                    "word": " about",
                    "probability": 0.8154296875
                },
                {
                    "start": 1489.98,
                    "end": 1490.5,
                    "word": " defining",
                    "probability": 0.60009765625
                },
                {
                    "start": 1490.86,
                    "end": 1491.22,
                    "word": " a false",
                    "probability": 0.8359375
                },
                {
                    "start": 1491.22,
                    "end": 1491.56,
                    "word": " negative",
                    "probability": 0.837890625
                },
                {
                    "start": 1491.56,
                    "end": 1491.62,
                    "word": ".",
                    "probability": 0.5732421875
                },
                {
                    "start": 1491.62,
                    "end": 1491.76,
                    "word": " They",
                    "probability": 0.693359375
                },
                {
                    "start": 1491.76,
                    "end": 1491.92,
                    "word": " want",
                    "probability": 0.798828125
                },
                {
                    "start": 1491.92,
                    "end": 1492.0,
                    "word": " to",
                    "probability": 0.90625
                },
                {
                    "start": 1492.0,
                    "end": 1492.2,
                    "word": " find",
                    "probability": 0.80615234375
                },
                {
                    "start": 1492.34,
                    "end": 1492.74,
                    "word": " a malicious",
                    "probability": 0.81298828125
                },
                {
                    "start": 1492.74,
                    "end": 1493.14,
                    "word": " sample",
                    "probability": 0.52392578125
                }
            ]
        },
        {
            "id": 225,
            "seek": 136338,
            "start": 1493.14,
            "end": 1497.38,
            "text": " we classify as benign. So what they'll do is they'll try a bunch of stuff until they find",
            "tokens": [
                50976,
                321,
                33872,
                382,
                3271,
                788,
                13,
                407,
                437,
                436,
                603,
                360,
                307,
                436,
                603,
                853,
                257,
                3840,
                295,
                1507,
                1826,
                436,
                915,
                51188
            ],
            "temperature": 0.0,
            "avg_logprob": -0.22569444957706664,
            "compression_ratio": 1.7383177570093458,
            "no_speech_prob": 0.0182647705078125,
            "words": [
                {
                    "start": 1493.14,
                    "end": 1493.36,
                    "word": " we",
                    "probability": 0.8046875
                },
                {
                    "start": 1493.36,
                    "end": 1493.76,
                    "word": " classify",
                    "probability": 0.92041015625
                },
                {
                    "start": 1493.76,
                    "end": 1494.0,
                    "word": " as",
                    "probability": 0.89892578125
                },
                {
                    "start": 1494.0,
                    "end": 1494.4,
                    "word": " benign",
                    "probability": 0.927734375
                },
                {
                    "start": 1494.4,
                    "end": 1494.78,
                    "word": ".",
                    "probability": 0.9658203125
                },
                {
                    "start": 1494.78,
                    "end": 1495.14,
                    "word": " So",
                    "probability": 0.888671875
                },
                {
                    "start": 1495.14,
                    "end": 1495.26,
                    "word": " what",
                    "probability": 0.83984375
                },
                {
                    "start": 1495.26,
                    "end": 1495.42,
                    "word": " they'll",
                    "probability": 0.81689453125
                },
                {
                    "start": 1495.42,
                    "end": 1495.56,
                    "word": " do",
                    "probability": 0.9013671875
                },
                {
                    "start": 1495.56,
                    "end": 1495.68,
                    "word": " is",
                    "probability": 0.89208984375
                },
                {
                    "start": 1495.68,
                    "end": 1495.84,
                    "word": " they'll",
                    "probability": 0.75634765625
                },
                {
                    "start": 1495.84,
                    "end": 1496.04,
                    "word": " try",
                    "probability": 0.51318359375
                },
                {
                    "start": 1496.2,
                    "end": 1496.4,
                    "word": " a bunch",
                    "probability": 0.7998046875
                },
                {
                    "start": 1496.4,
                    "end": 1496.5,
                    "word": " of",
                    "probability": 0.90966796875
                },
                {
                    "start": 1496.5,
                    "end": 1496.72,
                    "word": " stuff",
                    "probability": 0.80078125
                },
                {
                    "start": 1496.72,
                    "end": 1496.96,
                    "word": " until",
                    "probability": 0.8056640625
                },
                {
                    "start": 1496.96,
                    "end": 1497.1,
                    "word": " they",
                    "probability": 0.720703125
                },
                {
                    "start": 1497.1,
                    "end": 1497.38,
                    "word": " find",
                    "probability": 0.8134765625
                }
            ]
        },
        {
            "id": 226,
            "seek": 136338,
            "start": 1497.38,
            "end": 1503.6,
            "text": " something that is misclassified as benign and is actually malicious and bypasses our model,",
            "tokens": [
                51188,
                746,
                300,
                307,
                3346,
                11665,
                2587,
                382,
                3271,
                788,
                293,
                307,
                767,
                33496,
                293,
                24996,
                279,
                527,
                2316,
                11,
                51496
            ],
            "temperature": 0.0,
            "avg_logprob": -0.22569444957706664,
            "compression_ratio": 1.7383177570093458,
            "no_speech_prob": 0.0182647705078125,
            "words": [
                {
                    "start": 1497.38,
                    "end": 1497.82,
                    "word": " something",
                    "probability": 0.73779296875
                },
                {
                    "start": 1497.82,
                    "end": 1498.62,
                    "word": " that",
                    "probability": 0.8232421875
                },
                {
                    "start": 1498.62,
                    "end": 1498.84,
                    "word": " is",
                    "probability": 0.9052734375
                },
                {
                    "start": 1498.84,
                    "end": 1499.82,
                    "word": " misclassified",
                    "probability": 0.9475911458333334
                },
                {
                    "start": 1499.82,
                    "end": 1499.98,
                    "word": " as",
                    "probability": 0.86181640625
                },
                {
                    "start": 1499.98,
                    "end": 1500.48,
                    "word": " benign",
                    "probability": 0.967041015625
                },
                {
                    "start": 1500.48,
                    "end": 1500.96,
                    "word": " and",
                    "probability": 0.263427734375
                },
                {
                    "start": 1500.96,
                    "end": 1501.08,
                    "word": " is",
                    "probability": 0.5458984375
                },
                {
                    "start": 1501.08,
                    "end": 1501.46,
                    "word": " actually",
                    "probability": 0.7861328125
                },
                {
                    "start": 1501.46,
                    "end": 1501.84,
                    "word": " malicious",
                    "probability": 0.83251953125
                },
                {
                    "start": 1501.84,
                    "end": 1502.12,
                    "word": " and",
                    "probability": 0.85595703125
                },
                {
                    "start": 1502.12,
                    "end": 1502.78,
                    "word": " bypasses",
                    "probability": 0.8681640625
                },
                {
                    "start": 1502.78,
                    "end": 1502.94,
                    "word": " our",
                    "probability": 0.81787109375
                },
                {
                    "start": 1502.94,
                    "end": 1503.28,
                    "word": " model",
                    "probability": 0.8134765625
                },
                {
                    "start": 1503.28,
                    "end": 1503.6,
                    "word": ",",
                    "probability": 0.310791015625
                }
            ]
        },
        {
            "id": 227,
            "seek": 139264,
            "start": 1503.6,
            "end": 1513.84,
            "text": " and then they deploy it as far as much as possible. So if you do, if you apply this to,",
            "tokens": [
                50400,
                293,
                550,
                436,
                7274,
                309,
                382,
                1400,
                382,
                709,
                382,
                1944,
                13,
                407,
                498,
                291,
                360,
                11,
                498,
                291,
                3079,
                341,
                281,
                11,
                50868
            ],
            "temperature": 0.0,
            "avg_logprob": -0.22639266336741654,
            "compression_ratio": 1.643835616438356,
            "no_speech_prob": 0.0226287841796875,
            "words": [
                {
                    "start": 1503.6,
                    "end": 1503.8,
                    "word": " and",
                    "probability": 0.39599609375
                },
                {
                    "start": 1503.8,
                    "end": 1503.92,
                    "word": " then",
                    "probability": 0.64990234375
                },
                {
                    "start": 1503.92,
                    "end": 1505.18,
                    "word": " they",
                    "probability": 0.4375
                },
                {
                    "start": 1505.18,
                    "end": 1505.88,
                    "word": " deploy",
                    "probability": 0.703125
                },
                {
                    "start": 1505.88,
                    "end": 1506.58,
                    "word": " it",
                    "probability": 0.8955078125
                },
                {
                    "start": 1506.58,
                    "end": 1507.74,
                    "word": " as",
                    "probability": 0.87451171875
                },
                {
                    "start": 1507.74,
                    "end": 1508.0,
                    "word": " far",
                    "probability": 0.482177734375
                },
                {
                    "start": 1508.0,
                    "end": 1508.16,
                    "word": " as",
                    "probability": 0.62158203125
                },
                {
                    "start": 1508.16,
                    "end": 1508.3,
                    "word": " much",
                    "probability": 0.787109375
                },
                {
                    "start": 1508.3,
                    "end": 1508.44,
                    "word": " as",
                    "probability": 0.9033203125
                },
                {
                    "start": 1508.44,
                    "end": 1508.8,
                    "word": " possible",
                    "probability": 0.84326171875
                },
                {
                    "start": 1508.8,
                    "end": 1509.34,
                    "word": ".",
                    "probability": 0.83740234375
                },
                {
                    "start": 1509.7,
                    "end": 1509.8,
                    "word": " So",
                    "probability": 0.74267578125
                },
                {
                    "start": 1509.8,
                    "end": 1511.44,
                    "word": " if",
                    "probability": 0.5966796875
                },
                {
                    "start": 1511.44,
                    "end": 1511.54,
                    "word": " you",
                    "probability": 0.8955078125
                },
                {
                    "start": 1511.54,
                    "end": 1511.7,
                    "word": " do",
                    "probability": 0.49609375
                },
                {
                    "start": 1511.7,
                    "end": 1511.98,
                    "word": ",",
                    "probability": 0.66650390625
                },
                {
                    "start": 1511.98,
                    "end": 1512.14,
                    "word": " if",
                    "probability": 0.89013671875
                },
                {
                    "start": 1512.14,
                    "end": 1512.26,
                    "word": " you",
                    "probability": 0.8984375
                },
                {
                    "start": 1512.26,
                    "end": 1512.54,
                    "word": " apply",
                    "probability": 0.84912109375
                },
                {
                    "start": 1512.54,
                    "end": 1512.9,
                    "word": " this",
                    "probability": 0.8359375
                },
                {
                    "start": 1512.9,
                    "end": 1513.6,
                    "word": " to",
                    "probability": 0.83935546875
                },
                {
                    "start": 1513.6,
                    "end": 1513.84,
                    "word": ",",
                    "probability": 0.3984375
                }
            ]
        },
        {
            "id": 228,
            "seek": 139264,
            "start": 1513.84,
            "end": 1520.42,
            "text": " you know, the fast drift calculators to that, you can attach a little drift calculator to each user.",
            "tokens": [
                50868,
                291,
                458,
                11,
                264,
                2370,
                19699,
                4322,
                3391,
                281,
                300,
                11,
                291,
                393,
                5085,
                257,
                707,
                19699,
                24993,
                281,
                1184,
                4195,
                13,
                51200
            ],
            "temperature": 0.0,
            "avg_logprob": -0.22639266336741654,
            "compression_ratio": 1.643835616438356,
            "no_speech_prob": 0.0226287841796875,
            "words": [
                {
                    "start": 1513.84,
                    "end": 1513.98,
                    "word": " you",
                    "probability": 0.80517578125
                },
                {
                    "start": 1513.98,
                    "end": 1514.08,
                    "word": " know",
                    "probability": 0.81103515625
                },
                {
                    "start": 1514.08,
                    "end": 1514.18,
                    "word": ",",
                    "probability": 0.89794921875
                },
                {
                    "start": 1514.18,
                    "end": 1514.62,
                    "word": " the",
                    "probability": 0.8134765625
                },
                {
                    "start": 1514.62,
                    "end": 1514.98,
                    "word": " fast",
                    "probability": 0.6220703125
                },
                {
                    "start": 1514.98,
                    "end": 1515.14,
                    "word": " drift",
                    "probability": 0.7880859375
                },
                {
                    "start": 1515.14,
                    "end": 1515.66,
                    "word": " calculators",
                    "probability": 0.88818359375
                },
                {
                    "start": 1515.66,
                    "end": 1515.88,
                    "word": " to",
                    "probability": 0.84423828125
                },
                {
                    "start": 1515.88,
                    "end": 1516.12,
                    "word": " that",
                    "probability": 0.8818359375
                },
                {
                    "start": 1516.12,
                    "end": 1516.54,
                    "word": ",",
                    "probability": 0.767578125
                },
                {
                    "start": 1516.54,
                    "end": 1516.72,
                    "word": " you",
                    "probability": 0.8916015625
                },
                {
                    "start": 1516.72,
                    "end": 1516.94,
                    "word": " can",
                    "probability": 0.89990234375
                },
                {
                    "start": 1516.94,
                    "end": 1517.78,
                    "word": " attach",
                    "probability": 0.79296875
                },
                {
                    "start": 1518.14,
                    "end": 1518.34,
                    "word": " a little",
                    "probability": 0.79833984375
                },
                {
                    "start": 1518.34,
                    "end": 1518.56,
                    "word": " drift",
                    "probability": 0.880859375
                },
                {
                    "start": 1518.56,
                    "end": 1518.98,
                    "word": " calculator",
                    "probability": 0.828125
                },
                {
                    "start": 1518.98,
                    "end": 1519.38,
                    "word": " to",
                    "probability": 0.90576171875
                },
                {
                    "start": 1519.38,
                    "end": 1519.6,
                    "word": " each",
                    "probability": 0.86083984375
                },
                {
                    "start": 1519.6,
                    "end": 1519.88,
                    "word": " user",
                    "probability": 0.81201171875
                },
                {
                    "start": 1519.88,
                    "end": 1520.42,
                    "word": ".",
                    "probability": 0.488037109375
                }
            ]
        },
        {
            "id": 229,
            "seek": 139264,
            "start": 1520.56,
            "end": 1528.5,
            "text": " And as samples come over the wire, you track it for that user. And if the user starts",
            "tokens": [
                51200,
                400,
                382,
                10938,
                808,
                670,
                264,
                6234,
                11,
                291,
                2837,
                309,
                337,
                300,
                4195,
                13,
                400,
                498,
                264,
                4195,
                3719,
                51616
            ],
            "temperature": 0.0,
            "avg_logprob": -0.22639266336741654,
            "compression_ratio": 1.643835616438356,
            "no_speech_prob": 0.0226287841796875,
            "words": [
                {
                    "start": 1520.56,
                    "end": 1520.62,
                    "word": " And",
                    "probability": 0.86279296875
                },
                {
                    "start": 1520.62,
                    "end": 1520.96,
                    "word": " as",
                    "probability": 0.904296875
                },
                {
                    "start": 1520.96,
                    "end": 1521.36,
                    "word": " samples",
                    "probability": 0.7431640625
                },
                {
                    "start": 1521.36,
                    "end": 1521.62,
                    "word": " come",
                    "probability": 0.794921875
                },
                {
                    "start": 1521.62,
                    "end": 1521.82,
                    "word": " over",
                    "probability": 0.8818359375
                },
                {
                    "start": 1521.82,
                    "end": 1521.98,
                    "word": " the",
                    "probability": 0.82177734375
                },
                {
                    "start": 1521.98,
                    "end": 1522.26,
                    "word": " wire",
                    "probability": 0.7939453125
                },
                {
                    "start": 1522.26,
                    "end": 1522.84,
                    "word": ",",
                    "probability": 0.89453125
                },
                {
                    "start": 1522.84,
                    "end": 1523.38,
                    "word": " you",
                    "probability": 0.89111328125
                },
                {
                    "start": 1523.38,
                    "end": 1524.3,
                    "word": " track",
                    "probability": 0.826171875
                },
                {
                    "start": 1524.3,
                    "end": 1524.46,
                    "word": " it",
                    "probability": 0.8095703125
                },
                {
                    "start": 1524.46,
                    "end": 1524.64,
                    "word": " for",
                    "probability": 0.89892578125
                },
                {
                    "start": 1524.64,
                    "end": 1524.86,
                    "word": " that",
                    "probability": 0.888671875
                },
                {
                    "start": 1524.86,
                    "end": 1525.18,
                    "word": " user",
                    "probability": 0.80517578125
                },
                {
                    "start": 1525.18,
                    "end": 1525.72,
                    "word": ".",
                    "probability": 0.744140625
                },
                {
                    "start": 1526.44,
                    "end": 1526.98,
                    "word": " And",
                    "probability": 0.8857421875
                },
                {
                    "start": 1526.98,
                    "end": 1527.12,
                    "word": " if",
                    "probability": 0.90234375
                },
                {
                    "start": 1527.12,
                    "end": 1527.24,
                    "word": " the",
                    "probability": 0.80712890625
                },
                {
                    "start": 1527.24,
                    "end": 1527.54,
                    "word": " user",
                    "probability": 0.80029296875
                },
                {
                    "start": 1527.54,
                    "end": 1528.5,
                    "word": " starts",
                    "probability": 0.76806640625
                }
            ]
        },
        {
            "id": 230,
            "seek": 139264,
            "start": 1528.5,
            "end": 1532.86,
            "text": " exploiting something, they're going to get a very boring distribution where it's just",
            "tokens": [
                51616,
                12382,
                1748,
                746,
                11,
                436,
                434,
                516,
                281,
                483,
                257,
                588,
                9989,
                7316,
                689,
                309,
                311,
                445,
                51824
            ],
            "temperature": 0.0,
            "avg_logprob": -0.22639266336741654,
            "compression_ratio": 1.643835616438356,
            "no_speech_prob": 0.0226287841796875,
            "words": [
                {
                    "start": 1528.5,
                    "end": 1529.28,
                    "word": " exploiting",
                    "probability": 0.891845703125
                },
                {
                    "start": 1529.28,
                    "end": 1529.62,
                    "word": " something",
                    "probability": 0.7392578125
                },
                {
                    "start": 1529.62,
                    "end": 1530.1,
                    "word": ",",
                    "probability": 0.92138671875
                },
                {
                    "start": 1530.1,
                    "end": 1530.4,
                    "word": " they're",
                    "probability": 0.83544921875
                },
                {
                    "start": 1530.4,
                    "end": 1530.62,
                    "word": " going",
                    "probability": 0.85205078125
                },
                {
                    "start": 1530.62,
                    "end": 1530.78,
                    "word": " to",
                    "probability": 0.9072265625
                },
                {
                    "start": 1530.78,
                    "end": 1530.9,
                    "word": " get",
                    "probability": 0.8916015625
                },
                {
                    "start": 1531.0,
                    "end": 1531.18,
                    "word": " a very",
                    "probability": 0.74169921875
                },
                {
                    "start": 1531.18,
                    "end": 1531.56,
                    "word": " boring",
                    "probability": 0.83251953125
                },
                {
                    "start": 1531.56,
                    "end": 1532.1,
                    "word": " distribution",
                    "probability": 0.736328125
                },
                {
                    "start": 1532.1,
                    "end": 1532.58,
                    "word": " where",
                    "probability": 0.626953125
                },
                {
                    "start": 1532.58,
                    "end": 1532.72,
                    "word": " it's",
                    "probability": 0.95068359375
                },
                {
                    "start": 1532.72,
                    "end": 1532.86,
                    "word": " just",
                    "probability": 0.89111328125
                }
            ]
        },
        {
            "id": 231,
            "seek": 142202,
            "start": 1532.86,
            "end": 1541.52,
            "text": " one point repeated over and over again. And you spike your distribution, your KL divergence",
            "tokens": [
                50364,
                472,
                935,
                10477,
                670,
                293,
                670,
                797,
                13,
                400,
                291,
                21053,
                428,
                7316,
                11,
                428,
                47991,
                47387,
                50804
            ],
            "temperature": 0.0,
            "avg_logprob": -0.23422181124196334,
            "compression_ratio": 1.8506224066390042,
            "no_speech_prob": 0.0128173828125,
            "words": [
                {
                    "start": 1532.86,
                    "end": 1533.1,
                    "word": " one",
                    "probability": 0.5888671875
                },
                {
                    "start": 1533.1,
                    "end": 1533.36,
                    "word": " point",
                    "probability": 0.86376953125
                },
                {
                    "start": 1533.36,
                    "end": 1533.72,
                    "word": " repeated",
                    "probability": 0.7705078125
                },
                {
                    "start": 1533.72,
                    "end": 1534.0,
                    "word": " over",
                    "probability": 0.88037109375
                },
                {
                    "start": 1534.0,
                    "end": 1534.14,
                    "word": " and",
                    "probability": 0.89111328125
                },
                {
                    "start": 1534.14,
                    "end": 1534.24,
                    "word": " over",
                    "probability": 0.87646484375
                },
                {
                    "start": 1534.24,
                    "end": 1534.5,
                    "word": " again",
                    "probability": 0.91552734375
                },
                {
                    "start": 1534.5,
                    "end": 1535.02,
                    "word": ".",
                    "probability": 0.458251953125
                },
                {
                    "start": 1537.42,
                    "end": 1537.94,
                    "word": " And",
                    "probability": 0.64892578125
                },
                {
                    "start": 1537.94,
                    "end": 1538.2,
                    "word": " you",
                    "probability": 0.84423828125
                },
                {
                    "start": 1538.2,
                    "end": 1539.0,
                    "word": " spike",
                    "probability": 0.8994140625
                },
                {
                    "start": 1539.0,
                    "end": 1539.4,
                    "word": " your",
                    "probability": 0.29736328125
                },
                {
                    "start": 1539.4,
                    "end": 1539.84,
                    "word": " distribution",
                    "probability": 0.490234375
                },
                {
                    "start": 1539.84,
                    "end": 1540.1,
                    "word": ",",
                    "probability": 0.505859375
                },
                {
                    "start": 1540.1,
                    "end": 1540.58,
                    "word": " your",
                    "probability": 0.8115234375
                },
                {
                    "start": 1540.58,
                    "end": 1541.12,
                    "word": " KL",
                    "probability": 0.95751953125
                },
                {
                    "start": 1541.12,
                    "end": 1541.52,
                    "word": " divergence",
                    "probability": 0.5859375
                }
            ]
        },
        {
            "id": 232,
            "seek": 142202,
            "start": 1541.52,
            "end": 1545.82,
            "text": " spikes up and then becomes very normal because it's essentially you're tracing the same path",
            "tokens": [
                50804,
                28997,
                493,
                293,
                550,
                3643,
                588,
                2710,
                570,
                309,
                311,
                4476,
                291,
                434,
                25262,
                264,
                912,
                3100,
                51016
            ],
            "temperature": 0.0,
            "avg_logprob": -0.23422181124196334,
            "compression_ratio": 1.8506224066390042,
            "no_speech_prob": 0.0128173828125,
            "words": [
                {
                    "start": 1541.52,
                    "end": 1542.02,
                    "word": " spikes",
                    "probability": 0.80419921875
                },
                {
                    "start": 1542.02,
                    "end": 1542.3,
                    "word": " up",
                    "probability": 0.908203125
                },
                {
                    "start": 1542.3,
                    "end": 1542.48,
                    "word": " and",
                    "probability": 0.74658203125
                },
                {
                    "start": 1542.48,
                    "end": 1542.6,
                    "word": " then",
                    "probability": 0.728515625
                },
                {
                    "start": 1542.6,
                    "end": 1542.84,
                    "word": " becomes",
                    "probability": 0.77685546875
                },
                {
                    "start": 1542.84,
                    "end": 1543.02,
                    "word": " very",
                    "probability": 0.73193359375
                },
                {
                    "start": 1543.02,
                    "end": 1543.28,
                    "word": " normal",
                    "probability": 0.8232421875
                },
                {
                    "start": 1543.28,
                    "end": 1543.6,
                    "word": " because",
                    "probability": 0.71337890625
                },
                {
                    "start": 1543.6,
                    "end": 1543.96,
                    "word": " it's",
                    "probability": 0.9033203125
                },
                {
                    "start": 1543.96,
                    "end": 1544.36,
                    "word": " essentially",
                    "probability": 0.6337890625
                },
                {
                    "start": 1544.36,
                    "end": 1544.76,
                    "word": " you're",
                    "probability": 0.853515625
                },
                {
                    "start": 1544.76,
                    "end": 1545.1,
                    "word": " tracing",
                    "probability": 0.56298828125
                },
                {
                    "start": 1545.1,
                    "end": 1545.32,
                    "word": " the",
                    "probability": 0.81396484375
                },
                {
                    "start": 1545.32,
                    "end": 1545.52,
                    "word": " same",
                    "probability": 0.80322265625
                },
                {
                    "start": 1545.52,
                    "end": 1545.82,
                    "word": " path",
                    "probability": 0.8564453125
                }
            ]
        },
        {
            "id": 233,
            "seek": 142202,
            "start": 1545.82,
            "end": 1552.24,
            "text": " over and over and over again. And you can see that there's an exploration phase where the",
            "tokens": [
                51016,
                670,
                293,
                670,
                293,
                670,
                797,
                13,
                400,
                291,
                393,
                536,
                300,
                456,
                311,
                364,
                16197,
                5574,
                689,
                264,
                51324
            ],
            "temperature": 0.0,
            "avg_logprob": -0.23422181124196334,
            "compression_ratio": 1.8506224066390042,
            "no_speech_prob": 0.0128173828125,
            "words": [
                {
                    "start": 1545.82,
                    "end": 1546.18,
                    "word": " over",
                    "probability": 0.873046875
                },
                {
                    "start": 1546.18,
                    "end": 1546.38,
                    "word": " and",
                    "probability": 0.88232421875
                },
                {
                    "start": 1546.38,
                    "end": 1546.56,
                    "word": " over",
                    "probability": 0.875
                },
                {
                    "start": 1546.56,
                    "end": 1546.7,
                    "word": " and",
                    "probability": 0.853515625
                },
                {
                    "start": 1546.7,
                    "end": 1546.84,
                    "word": " over",
                    "probability": 0.87109375
                },
                {
                    "start": 1546.84,
                    "end": 1547.1,
                    "word": " again",
                    "probability": 0.91748046875
                },
                {
                    "start": 1547.1,
                    "end": 1547.62,
                    "word": ".",
                    "probability": 0.9443359375
                },
                {
                    "start": 1549.16,
                    "end": 1549.68,
                    "word": " And",
                    "probability": 0.83935546875
                },
                {
                    "start": 1549.68,
                    "end": 1549.94,
                    "word": " you",
                    "probability": 0.8974609375
                },
                {
                    "start": 1549.94,
                    "end": 1550.08,
                    "word": " can",
                    "probability": 0.89892578125
                },
                {
                    "start": 1550.08,
                    "end": 1550.34,
                    "word": " see",
                    "probability": 0.818359375
                },
                {
                    "start": 1550.34,
                    "end": 1550.56,
                    "word": " that",
                    "probability": 0.86962890625
                },
                {
                    "start": 1550.56,
                    "end": 1550.74,
                    "word": " there's",
                    "probability": 0.925537109375
                },
                {
                    "start": 1550.74,
                    "end": 1550.84,
                    "word": " an",
                    "probability": 0.9091796875
                },
                {
                    "start": 1550.84,
                    "end": 1551.22,
                    "word": " exploration",
                    "probability": 0.7412109375
                },
                {
                    "start": 1551.22,
                    "end": 1551.7,
                    "word": " phase",
                    "probability": 0.85205078125
                },
                {
                    "start": 1551.7,
                    "end": 1552.04,
                    "word": " where",
                    "probability": 0.82421875
                },
                {
                    "start": 1552.04,
                    "end": 1552.24,
                    "word": " the",
                    "probability": 0.82275390625
                }
            ]
        },
        {
            "id": 234,
            "seek": 142202,
            "start": 1552.24,
            "end": 1557.64,
            "text": " malicious users try to look for a benign misclassified sample, and when they find one,",
            "tokens": [
                51324,
                33496,
                5022,
                853,
                281,
                574,
                337,
                257,
                3271,
                788,
                3346,
                11665,
                2587,
                6889,
                11,
                293,
                562,
                436,
                915,
                472,
                11,
                51596
            ],
            "temperature": 0.0,
            "avg_logprob": -0.23422181124196334,
            "compression_ratio": 1.8506224066390042,
            "no_speech_prob": 0.0128173828125,
            "words": [
                {
                    "start": 1552.24,
                    "end": 1552.48,
                    "word": " malicious",
                    "probability": 0.84423828125
                },
                {
                    "start": 1552.48,
                    "end": 1552.8,
                    "word": " users",
                    "probability": 0.72119140625
                },
                {
                    "start": 1552.8,
                    "end": 1553.0,
                    "word": " try",
                    "probability": 0.81298828125
                },
                {
                    "start": 1553.0,
                    "end": 1553.16,
                    "word": " to",
                    "probability": 0.90966796875
                },
                {
                    "start": 1553.16,
                    "end": 1553.36,
                    "word": " look",
                    "probability": 0.919921875
                },
                {
                    "start": 1553.36,
                    "end": 1553.6,
                    "word": " for",
                    "probability": 0.89892578125
                },
                {
                    "start": 1553.88,
                    "end": 1554.3,
                    "word": " a benign",
                    "probability": 0.900390625
                },
                {
                    "start": 1554.3,
                    "end": 1555.76,
                    "word": " misclassified",
                    "probability": 0.8155924479166666
                },
                {
                    "start": 1555.76,
                    "end": 1556.1,
                    "word": " sample",
                    "probability": 0.7001953125
                },
                {
                    "start": 1556.1,
                    "end": 1556.28,
                    "word": ",",
                    "probability": 0.16552734375
                },
                {
                    "start": 1556.28,
                    "end": 1556.34,
                    "word": " and",
                    "probability": 0.8916015625
                },
                {
                    "start": 1556.34,
                    "end": 1556.48,
                    "word": " when",
                    "probability": 0.1649169921875
                },
                {
                    "start": 1556.48,
                    "end": 1556.62,
                    "word": " they",
                    "probability": 0.740234375
                },
                {
                    "start": 1556.62,
                    "end": 1557.42,
                    "word": " find",
                    "probability": 0.82275390625
                },
                {
                    "start": 1557.42,
                    "end": 1557.6,
                    "word": " one",
                    "probability": 0.82666015625
                },
                {
                    "start": 1557.6,
                    "end": 1557.64,
                    "word": ",",
                    "probability": 0.3193359375
                }
            ]
        },
        {
            "id": 235,
            "seek": 142202,
            "start": 1557.64,
            "end": 1562.24,
            "text": " there's an exploitation phase. And the KL divergence rises rapidly. You can see that",
            "tokens": [
                51596,
                456,
                311,
                364,
                33122,
                5574,
                13,
                400,
                264,
                47991,
                47387,
                21373,
                12910,
                13,
                509,
                393,
                536,
                300,
                51828
            ],
            "temperature": 0.0,
            "avg_logprob": -0.23422181124196334,
            "compression_ratio": 1.8506224066390042,
            "no_speech_prob": 0.0128173828125,
            "words": [
                {
                    "start": 1557.64,
                    "end": 1557.78,
                    "word": " there's",
                    "probability": 0.92529296875
                },
                {
                    "start": 1557.78,
                    "end": 1557.88,
                    "word": " an",
                    "probability": 0.904296875
                },
                {
                    "start": 1557.88,
                    "end": 1558.2,
                    "word": " exploitation",
                    "probability": 0.80126953125
                },
                {
                    "start": 1558.2,
                    "end": 1558.66,
                    "word": " phase",
                    "probability": 0.81591796875
                },
                {
                    "start": 1558.66,
                    "end": 1559.04,
                    "word": ".",
                    "probability": 0.364990234375
                },
                {
                    "start": 1559.04,
                    "end": 1559.24,
                    "word": " And",
                    "probability": 0.8603515625
                },
                {
                    "start": 1559.24,
                    "end": 1559.38,
                    "word": " the",
                    "probability": 0.8173828125
                },
                {
                    "start": 1559.38,
                    "end": 1559.58,
                    "word": " KL",
                    "probability": 0.9453125
                },
                {
                    "start": 1559.58,
                    "end": 1560.0,
                    "word": " divergence",
                    "probability": 0.791015625
                },
                {
                    "start": 1560.0,
                    "end": 1560.58,
                    "word": " rises",
                    "probability": 0.79150390625
                },
                {
                    "start": 1560.58,
                    "end": 1560.96,
                    "word": " rapidly",
                    "probability": 0.716796875
                },
                {
                    "start": 1560.96,
                    "end": 1561.48,
                    "word": ".",
                    "probability": 0.96826171875
                },
                {
                    "start": 1561.48,
                    "end": 1561.72,
                    "word": " You",
                    "probability": 0.88720703125
                },
                {
                    "start": 1561.72,
                    "end": 1561.86,
                    "word": " can",
                    "probability": 0.89990234375
                },
                {
                    "start": 1561.86,
                    "end": 1562.06,
                    "word": " see",
                    "probability": 0.80859375
                },
                {
                    "start": 1562.06,
                    "end": 1562.24,
                    "word": " that",
                    "probability": 0.7861328125
                }
            ]
        },
        {
            "id": 236,
            "seek": 144400,
            "start": 1562.24,
            "end": 1570.32,
            "text": " that this is in log of scale. So this is literally hundreds of times more KL divergence than the",
            "tokens": [
                50380,
                300,
                341,
                307,
                294,
                3565,
                295,
                4373,
                13,
                407,
                341,
                307,
                3736,
                6779,
                295,
                1413,
                544,
                47991,
                47387,
                813,
                264,
                50760
            ],
            "temperature": 0.0,
            "avg_logprob": -0.26976103554753694,
            "compression_ratio": 1.6785714285714286,
            "no_speech_prob": 0.004680633544921875,
            "words": [
                {
                    "start": 1562.24,
                    "end": 1562.7,
                    "word": " that",
                    "probability": 0.054351806640625
                },
                {
                    "start": 1562.7,
                    "end": 1563.18,
                    "word": " this",
                    "probability": 0.83984375
                },
                {
                    "start": 1563.18,
                    "end": 1564.5,
                    "word": " is",
                    "probability": 0.8818359375
                },
                {
                    "start": 1564.5,
                    "end": 1565.08,
                    "word": " in",
                    "probability": 0.880859375
                },
                {
                    "start": 1565.08,
                    "end": 1565.46,
                    "word": " log",
                    "probability": 0.6669921875
                },
                {
                    "start": 1565.46,
                    "end": 1565.66,
                    "word": " of",
                    "probability": 0.5849609375
                },
                {
                    "start": 1565.66,
                    "end": 1565.98,
                    "word": " scale",
                    "probability": 0.78466796875
                },
                {
                    "start": 1565.98,
                    "end": 1566.28,
                    "word": ".",
                    "probability": 0.492919921875
                },
                {
                    "start": 1566.28,
                    "end": 1566.38,
                    "word": " So",
                    "probability": 0.814453125
                },
                {
                    "start": 1566.38,
                    "end": 1566.58,
                    "word": " this",
                    "probability": 0.82470703125
                },
                {
                    "start": 1566.58,
                    "end": 1566.76,
                    "word": " is",
                    "probability": 0.90625
                },
                {
                    "start": 1566.76,
                    "end": 1567.2,
                    "word": " literally",
                    "probability": 0.73193359375
                },
                {
                    "start": 1567.2,
                    "end": 1567.96,
                    "word": " hundreds",
                    "probability": 0.755859375
                },
                {
                    "start": 1567.96,
                    "end": 1568.22,
                    "word": " of",
                    "probability": 0.90283203125
                },
                {
                    "start": 1568.22,
                    "end": 1568.58,
                    "word": " times",
                    "probability": 0.81884765625
                },
                {
                    "start": 1568.58,
                    "end": 1569.12,
                    "word": " more",
                    "probability": 0.87548828125
                },
                {
                    "start": 1569.12,
                    "end": 1569.38,
                    "word": " KL",
                    "probability": 0.77490234375
                },
                {
                    "start": 1569.38,
                    "end": 1569.82,
                    "word": " divergence",
                    "probability": 0.60107421875
                },
                {
                    "start": 1569.82,
                    "end": 1570.16,
                    "word": " than",
                    "probability": 0.88525390625
                },
                {
                    "start": 1570.16,
                    "end": 1570.32,
                    "word": " the",
                    "probability": 0.7978515625
                }
            ]
        },
        {
            "id": 237,
            "seek": 144400,
            "start": 1570.32,
            "end": 1575.28,
            "text": " benign samples. And here, it's the same thing. Hundreds of times more KL divergence than",
            "tokens": [
                50760,
                3271,
                788,
                10938,
                13,
                400,
                510,
                11,
                309,
                311,
                264,
                912,
                551,
                13,
                45785,
                295,
                1413,
                544,
                47991,
                47387,
                813,
                51028
            ],
            "temperature": 0.0,
            "avg_logprob": -0.26976103554753694,
            "compression_ratio": 1.6785714285714286,
            "no_speech_prob": 0.004680633544921875,
            "words": [
                {
                    "start": 1570.32,
                    "end": 1570.64,
                    "word": " benign",
                    "probability": 0.909912109375
                },
                {
                    "start": 1570.64,
                    "end": 1571.04,
                    "word": " samples",
                    "probability": 0.7177734375
                },
                {
                    "start": 1571.04,
                    "end": 1571.22,
                    "word": ".",
                    "probability": 0.76171875
                },
                {
                    "start": 1571.22,
                    "end": 1571.3,
                    "word": " And",
                    "probability": 0.84765625
                },
                {
                    "start": 1571.3,
                    "end": 1571.64,
                    "word": " here",
                    "probability": 0.80322265625
                },
                {
                    "start": 1571.64,
                    "end": 1572.24,
                    "word": ",",
                    "probability": 0.39697265625
                },
                {
                    "start": 1572.24,
                    "end": 1572.62,
                    "word": " it's",
                    "probability": 0.837646484375
                },
                {
                    "start": 1572.62,
                    "end": 1572.84,
                    "word": " the",
                    "probability": 0.8095703125
                },
                {
                    "start": 1572.84,
                    "end": 1573.0,
                    "word": " same",
                    "probability": 0.8125
                },
                {
                    "start": 1573.0,
                    "end": 1573.24,
                    "word": " thing",
                    "probability": 0.80859375
                },
                {
                    "start": 1573.24,
                    "end": 1573.38,
                    "word": ".",
                    "probability": 0.424072265625
                },
                {
                    "start": 1573.38,
                    "end": 1573.64,
                    "word": " Hundreds",
                    "probability": 0.7529296875
                },
                {
                    "start": 1573.64,
                    "end": 1573.88,
                    "word": " of",
                    "probability": 0.9052734375
                },
                {
                    "start": 1573.88,
                    "end": 1574.14,
                    "word": " times",
                    "probability": 0.81591796875
                },
                {
                    "start": 1574.14,
                    "end": 1574.34,
                    "word": " more",
                    "probability": 0.86767578125
                },
                {
                    "start": 1574.34,
                    "end": 1574.56,
                    "word": " KL",
                    "probability": 0.96826171875
                },
                {
                    "start": 1574.56,
                    "end": 1574.92,
                    "word": " divergence",
                    "probability": 0.7353515625
                },
                {
                    "start": 1574.92,
                    "end": 1575.28,
                    "word": " than",
                    "probability": 0.849609375
                }
            ]
        },
        {
            "id": 238,
            "seek": 144400,
            "start": 1575.28,
            "end": 1584.22,
            "text": " the benign and the attackers were unsuccessful. Because of this, you can build, you can probably",
            "tokens": [
                51028,
                264,
                3271,
                788,
                293,
                264,
                45129,
                645,
                46258,
                13,
                1436,
                295,
                341,
                11,
                291,
                393,
                1322,
                11,
                291,
                393,
                1391,
                51464
            ],
            "temperature": 0.0,
            "avg_logprob": -0.26976103554753694,
            "compression_ratio": 1.6785714285714286,
            "no_speech_prob": 0.004680633544921875,
            "words": [
                {
                    "start": 1575.28,
                    "end": 1575.82,
                    "word": " the",
                    "probability": 0.6103515625
                },
                {
                    "start": 1575.82,
                    "end": 1576.46,
                    "word": " benign",
                    "probability": 0.947998046875
                },
                {
                    "start": 1576.46,
                    "end": 1576.9,
                    "word": " and",
                    "probability": 0.58154296875
                },
                {
                    "start": 1576.9,
                    "end": 1577.56,
                    "word": " the",
                    "probability": 0.7861328125
                },
                {
                    "start": 1577.56,
                    "end": 1578.08,
                    "word": " attackers",
                    "probability": 0.208251953125
                },
                {
                    "start": 1578.08,
                    "end": 1578.36,
                    "word": " were",
                    "probability": 0.32470703125
                },
                {
                    "start": 1578.36,
                    "end": 1578.96,
                    "word": " unsuccessful",
                    "probability": 0.8603515625
                },
                {
                    "start": 1578.96,
                    "end": 1579.64,
                    "word": ".",
                    "probability": 0.978515625
                },
                {
                    "start": 1580.78,
                    "end": 1581.34,
                    "word": " Because",
                    "probability": 0.76123046875
                },
                {
                    "start": 1581.34,
                    "end": 1581.54,
                    "word": " of",
                    "probability": 0.90673828125
                },
                {
                    "start": 1581.54,
                    "end": 1581.82,
                    "word": " this",
                    "probability": 0.87841796875
                },
                {
                    "start": 1581.82,
                    "end": 1582.32,
                    "word": ",",
                    "probability": 0.92138671875
                },
                {
                    "start": 1582.32,
                    "end": 1582.4,
                    "word": " you",
                    "probability": 0.89501953125
                },
                {
                    "start": 1582.4,
                    "end": 1582.58,
                    "word": " can",
                    "probability": 0.7607421875
                },
                {
                    "start": 1582.58,
                    "end": 1582.98,
                    "word": " build",
                    "probability": 0.73828125
                },
                {
                    "start": 1582.98,
                    "end": 1583.5,
                    "word": ",",
                    "probability": 0.681640625
                },
                {
                    "start": 1583.5,
                    "end": 1583.7,
                    "word": " you",
                    "probability": 0.892578125
                },
                {
                    "start": 1583.7,
                    "end": 1583.84,
                    "word": " can",
                    "probability": 0.461669921875
                },
                {
                    "start": 1583.84,
                    "end": 1584.22,
                    "word": " probably",
                    "probability": 0.9130859375
                }
            ]
        },
        {
            "id": 239,
            "seek": 147300,
            "start": 1584.22,
            "end": 1594.46,
            "text": " build a very good online defense of your system against a set attack. And here's where basically",
            "tokens": [
                50364,
                1322,
                257,
                588,
                665,
                2950,
                7654,
                295,
                428,
                1185,
                1970,
                257,
                992,
                2690,
                13,
                400,
                510,
                311,
                689,
                1936,
                50876
            ],
            "temperature": 0.0,
            "avg_logprob": -0.25181684729664827,
            "compression_ratio": 1.6788990825688073,
            "no_speech_prob": 0.484375,
            "words": [
                {
                    "start": 1584.22,
                    "end": 1584.62,
                    "word": " build",
                    "probability": 0.67333984375
                },
                {
                    "start": 1584.82,
                    "end": 1585.06,
                    "word": " a very",
                    "probability": 0.72705078125
                },
                {
                    "start": 1585.06,
                    "end": 1585.56,
                    "word": " good",
                    "probability": 0.87548828125
                },
                {
                    "start": 1585.56,
                    "end": 1586.56,
                    "word": " online",
                    "probability": 0.73193359375
                },
                {
                    "start": 1586.56,
                    "end": 1587.36,
                    "word": " defense",
                    "probability": 0.72607421875
                },
                {
                    "start": 1587.36,
                    "end": 1587.52,
                    "word": " of",
                    "probability": 0.88232421875
                },
                {
                    "start": 1587.52,
                    "end": 1587.84,
                    "word": " your",
                    "probability": 0.8154296875
                },
                {
                    "start": 1587.84,
                    "end": 1588.66,
                    "word": " system",
                    "probability": 0.86279296875
                },
                {
                    "start": 1588.66,
                    "end": 1589.06,
                    "word": " against",
                    "probability": 0.8662109375
                },
                {
                    "start": 1589.28,
                    "end": 1589.4,
                    "word": " a set",
                    "probability": 0.19873046875
                },
                {
                    "start": 1589.4,
                    "end": 1589.9,
                    "word": " attack",
                    "probability": 0.7685546875
                },
                {
                    "start": 1589.9,
                    "end": 1590.46,
                    "word": ".",
                    "probability": 0.75390625
                },
                {
                    "start": 1592.9,
                    "end": 1593.46,
                    "word": " And",
                    "probability": 0.411865234375
                },
                {
                    "start": 1593.46,
                    "end": 1593.88,
                    "word": " here's",
                    "probability": 0.65625
                },
                {
                    "start": 1593.88,
                    "end": 1594.06,
                    "word": " where",
                    "probability": 0.89697265625
                },
                {
                    "start": 1594.06,
                    "end": 1594.46,
                    "word": " basically",
                    "probability": 0.78564453125
                }
            ]
        },
        {
            "id": 240,
            "seek": 147300,
            "start": 1594.46,
            "end": 1601.88,
            "text": " the system was originally built sort of with this mind and other attacks in mind, because",
            "tokens": [
                50960,
                264,
                1185,
                390,
                7993,
                3094,
                1333,
                295,
                365,
                341,
                1575,
                293,
                661,
                8122,
                294,
                1575,
                11,
                570,
                51240
            ],
            "temperature": 0.0,
            "avg_logprob": -0.25181684729664827,
            "compression_ratio": 1.6788990825688073,
            "no_speech_prob": 0.484375,
            "words": [
                {
                    "start": 1594.46,
                    "end": 1595.06,
                    "word": " the",
                    "probability": 0.68310546875
                },
                {
                    "start": 1595.06,
                    "end": 1596.84,
                    "word": " system",
                    "probability": 0.64208984375
                },
                {
                    "start": 1596.84,
                    "end": 1597.04,
                    "word": " was",
                    "probability": 0.89306640625
                },
                {
                    "start": 1597.04,
                    "end": 1597.42,
                    "word": " originally",
                    "probability": 0.58544921875
                },
                {
                    "start": 1597.42,
                    "end": 1597.84,
                    "word": " built",
                    "probability": 0.92333984375
                },
                {
                    "start": 1597.84,
                    "end": 1598.42,
                    "word": " sort",
                    "probability": 0.45166015625
                },
                {
                    "start": 1598.42,
                    "end": 1598.52,
                    "word": " of",
                    "probability": 0.90673828125
                },
                {
                    "start": 1598.52,
                    "end": 1598.72,
                    "word": " with",
                    "probability": 0.81396484375
                },
                {
                    "start": 1598.72,
                    "end": 1599.18,
                    "word": " this",
                    "probability": 0.876953125
                },
                {
                    "start": 1599.18,
                    "end": 1599.88,
                    "word": " mind",
                    "probability": 0.387451171875
                },
                {
                    "start": 1599.88,
                    "end": 1600.66,
                    "word": " and",
                    "probability": 0.85986328125
                },
                {
                    "start": 1600.66,
                    "end": 1600.9,
                    "word": " other",
                    "probability": 0.79736328125
                },
                {
                    "start": 1600.9,
                    "end": 1601.26,
                    "word": " attacks",
                    "probability": 0.6923828125
                },
                {
                    "start": 1601.26,
                    "end": 1601.4,
                    "word": " in",
                    "probability": 0.904296875
                },
                {
                    "start": 1601.4,
                    "end": 1601.62,
                    "word": " mind",
                    "probability": 0.8017578125
                },
                {
                    "start": 1601.62,
                    "end": 1601.64,
                    "word": ",",
                    "probability": 0.277099609375
                },
                {
                    "start": 1601.64,
                    "end": 1601.88,
                    "word": " because",
                    "probability": 0.8310546875
                }
            ]
        },
        {
            "id": 241,
            "seek": 147300,
            "start": 1601.88,
            "end": 1608.44,
            "text": " I'm more interested in defending my models than calculating drift. But because it's so",
            "tokens": [
                51272,
                286,
                478,
                544,
                3102,
                294,
                21377,
                452,
                5245,
                813,
                28258,
                19699,
                13,
                583,
                570,
                309,
                311,
                370,
                51568
            ],
            "temperature": 0.0,
            "avg_logprob": -0.25181684729664827,
            "compression_ratio": 1.6788990825688073,
            "no_speech_prob": 0.484375,
            "words": [
                {
                    "start": 1601.88,
                    "end": 1602.72,
                    "word": " I'm",
                    "probability": 0.965087890625
                },
                {
                    "start": 1602.72,
                    "end": 1603.06,
                    "word": " more",
                    "probability": 0.8740234375
                },
                {
                    "start": 1603.06,
                    "end": 1603.54,
                    "word": " interested",
                    "probability": 0.6552734375
                },
                {
                    "start": 1603.54,
                    "end": 1603.7,
                    "word": " in",
                    "probability": 0.88720703125
                },
                {
                    "start": 1603.7,
                    "end": 1603.94,
                    "word": " defending",
                    "probability": 0.794921875
                },
                {
                    "start": 1603.94,
                    "end": 1604.18,
                    "word": " my",
                    "probability": 0.90087890625
                },
                {
                    "start": 1604.18,
                    "end": 1604.48,
                    "word": " models",
                    "probability": 0.794921875
                },
                {
                    "start": 1604.48,
                    "end": 1604.76,
                    "word": " than",
                    "probability": 0.6611328125
                },
                {
                    "start": 1604.76,
                    "end": 1605.66,
                    "word": " calculating",
                    "probability": 0.89794921875
                },
                {
                    "start": 1605.66,
                    "end": 1606.0,
                    "word": " drift",
                    "probability": 0.82958984375
                },
                {
                    "start": 1606.0,
                    "end": 1606.56,
                    "word": ".",
                    "probability": 0.880859375
                },
                {
                    "start": 1607.34,
                    "end": 1607.66,
                    "word": " But",
                    "probability": 0.87939453125
                },
                {
                    "start": 1607.66,
                    "end": 1608.08,
                    "word": " because",
                    "probability": 0.7626953125
                },
                {
                    "start": 1608.08,
                    "end": 1608.26,
                    "word": " it's",
                    "probability": 0.952392578125
                },
                {
                    "start": 1608.26,
                    "end": 1608.44,
                    "word": " so",
                    "probability": 0.900390625
                }
            ]
        },
        {
            "id": 242,
            "seek": 147300,
            "start": 1608.44,
            "end": 1613.22,
            "text": " bloody fast, it does this really well. And because it's so bloody fast, you can retrofit it.",
            "tokens": [
                51568,
                18938,
                2370,
                11,
                309,
                775,
                341,
                534,
                731,
                13,
                400,
                570,
                309,
                311,
                370,
                18938,
                2370,
                11,
                291,
                393,
                18820,
                6845,
                309,
                13,
                51780
            ],
            "temperature": 0.0,
            "avg_logprob": -0.25181684729664827,
            "compression_ratio": 1.6788990825688073,
            "no_speech_prob": 0.484375,
            "words": [
                {
                    "start": 1608.44,
                    "end": 1608.7,
                    "word": " bloody",
                    "probability": 0.9072265625
                },
                {
                    "start": 1608.7,
                    "end": 1609.1,
                    "word": " fast",
                    "probability": 0.81103515625
                },
                {
                    "start": 1609.1,
                    "end": 1609.16,
                    "word": ",",
                    "probability": 0.70654296875
                },
                {
                    "start": 1609.16,
                    "end": 1609.32,
                    "word": " it",
                    "probability": 0.90234375
                },
                {
                    "start": 1609.32,
                    "end": 1609.5,
                    "word": " does",
                    "probability": 0.9306640625
                },
                {
                    "start": 1609.5,
                    "end": 1609.66,
                    "word": " this",
                    "probability": 0.87841796875
                },
                {
                    "start": 1609.66,
                    "end": 1609.96,
                    "word": " really",
                    "probability": 0.78515625
                },
                {
                    "start": 1609.96,
                    "end": 1610.28,
                    "word": " well",
                    "probability": 0.8837890625
                },
                {
                    "start": 1610.28,
                    "end": 1610.42,
                    "word": ".",
                    "probability": 0.442138671875
                },
                {
                    "start": 1610.42,
                    "end": 1610.6,
                    "word": " And",
                    "probability": 0.8974609375
                },
                {
                    "start": 1610.6,
                    "end": 1610.8,
                    "word": " because",
                    "probability": 0.8310546875
                },
                {
                    "start": 1610.8,
                    "end": 1610.94,
                    "word": " it's",
                    "probability": 0.941162109375
                },
                {
                    "start": 1610.94,
                    "end": 1611.06,
                    "word": " so",
                    "probability": 0.90478515625
                },
                {
                    "start": 1611.06,
                    "end": 1611.26,
                    "word": " bloody",
                    "probability": 0.88232421875
                },
                {
                    "start": 1611.26,
                    "end": 1611.54,
                    "word": " fast",
                    "probability": 0.818359375
                },
                {
                    "start": 1611.54,
                    "end": 1611.56,
                    "word": ",",
                    "probability": 0.9208984375
                },
                {
                    "start": 1611.56,
                    "end": 1611.66,
                    "word": " you",
                    "probability": 0.8955078125
                },
                {
                    "start": 1611.66,
                    "end": 1611.84,
                    "word": " can",
                    "probability": 0.89501953125
                },
                {
                    "start": 1611.84,
                    "end": 1612.46,
                    "word": " retrofit",
                    "probability": 0.882080078125
                },
                {
                    "start": 1612.46,
                    "end": 1612.66,
                    "word": " it",
                    "probability": 0.88720703125
                },
                {
                    "start": 1612.66,
                    "end": 1613.22,
                    "word": ".",
                    "probability": 0.9912109375
                }
            ]
        },
        {
            "id": 243,
            "seek": 149964,
            "start": 1613.22,
            "end": 1621.08,
            "text": " So here's where I'll talk about some next steps. So as I mentioned previously,",
            "tokens": [
                50364,
                407,
                510,
                311,
                689,
                286,
                603,
                751,
                466,
                512,
                958,
                4439,
                13,
                407,
                382,
                286,
                2835,
                8046,
                11,
                50752
            ],
            "temperature": 0.0,
            "avg_logprob": -0.22694360028679778,
            "compression_ratio": 1.558252427184466,
            "no_speech_prob": 0.01690673828125,
            "words": [
                {
                    "start": 1613.22,
                    "end": 1613.52,
                    "word": " So",
                    "probability": 0.226318359375
                },
                {
                    "start": 1613.52,
                    "end": 1615.68,
                    "word": " here's",
                    "probability": 0.682861328125
                },
                {
                    "start": 1615.68,
                    "end": 1617.52,
                    "word": " where",
                    "probability": 0.53564453125
                },
                {
                    "start": 1617.52,
                    "end": 1617.66,
                    "word": " I'll",
                    "probability": 0.93505859375
                },
                {
                    "start": 1617.66,
                    "end": 1617.82,
                    "word": " talk",
                    "probability": 0.8017578125
                },
                {
                    "start": 1617.82,
                    "end": 1617.98,
                    "word": " about",
                    "probability": 0.81689453125
                },
                {
                    "start": 1617.98,
                    "end": 1618.16,
                    "word": " some",
                    "probability": 0.62255859375
                },
                {
                    "start": 1618.16,
                    "end": 1618.34,
                    "word": " next",
                    "probability": 0.849609375
                },
                {
                    "start": 1618.34,
                    "end": 1618.64,
                    "word": " steps",
                    "probability": 0.73193359375
                },
                {
                    "start": 1618.64,
                    "end": 1619.16,
                    "word": ".",
                    "probability": 0.83349609375
                },
                {
                    "start": 1619.18,
                    "end": 1619.32,
                    "word": " So",
                    "probability": 0.65625
                },
                {
                    "start": 1619.32,
                    "end": 1619.7,
                    "word": " as",
                    "probability": 0.69580078125
                },
                {
                    "start": 1619.7,
                    "end": 1619.8,
                    "word": " I",
                    "probability": 0.9853515625
                },
                {
                    "start": 1619.8,
                    "end": 1620.22,
                    "word": " mentioned",
                    "probability": 0.422607421875
                },
                {
                    "start": 1620.22,
                    "end": 1620.68,
                    "word": " previously",
                    "probability": 0.7021484375
                },
                {
                    "start": 1620.68,
                    "end": 1621.08,
                    "word": ",",
                    "probability": 0.77099609375
                }
            ]
        },
        {
            "id": 244,
            "seek": 149964,
            "start": 1621.08,
            "end": 1627.0,
            "text": " this thing is a bit noisy. There are ways to fix it. In this case,",
            "tokens": [
                50752,
                341,
                551,
                307,
                257,
                857,
                24518,
                13,
                821,
                366,
                2098,
                281,
                3191,
                309,
                13,
                682,
                341,
                1389,
                11,
                50960
            ],
            "temperature": 0.0,
            "avg_logprob": -0.22694360028679778,
            "compression_ratio": 1.558252427184466,
            "no_speech_prob": 0.01690673828125,
            "words": [
                {
                    "start": 1621.08,
                    "end": 1621.32,
                    "word": " this",
                    "probability": 0.8740234375
                },
                {
                    "start": 1621.32,
                    "end": 1621.5,
                    "word": " thing",
                    "probability": 0.80859375
                },
                {
                    "start": 1621.5,
                    "end": 1621.66,
                    "word": " is",
                    "probability": 0.90283203125
                },
                {
                    "start": 1621.96,
                    "end": 1622.08,
                    "word": " a bit",
                    "probability": 0.90234375
                },
                {
                    "start": 1622.08,
                    "end": 1622.4,
                    "word": " noisy",
                    "probability": 0.7880859375
                },
                {
                    "start": 1622.4,
                    "end": 1622.74,
                    "word": ".",
                    "probability": 0.83740234375
                },
                {
                    "start": 1622.74,
                    "end": 1623.0,
                    "word": " There",
                    "probability": 0.8017578125
                },
                {
                    "start": 1623.0,
                    "end": 1623.14,
                    "word": " are",
                    "probability": 0.89599609375
                },
                {
                    "start": 1623.14,
                    "end": 1623.38,
                    "word": " ways",
                    "probability": 0.8076171875
                },
                {
                    "start": 1623.38,
                    "end": 1623.58,
                    "word": " to",
                    "probability": 0.904296875
                },
                {
                    "start": 1623.58,
                    "end": 1623.78,
                    "word": " fix",
                    "probability": 0.904296875
                },
                {
                    "start": 1623.78,
                    "end": 1623.98,
                    "word": " it",
                    "probability": 0.90283203125
                },
                {
                    "start": 1623.98,
                    "end": 1624.5,
                    "word": ".",
                    "probability": 0.6552734375
                },
                {
                    "start": 1624.5,
                    "end": 1624.74,
                    "word": " In",
                    "probability": 0.908203125
                },
                {
                    "start": 1624.74,
                    "end": 1624.9,
                    "word": " this",
                    "probability": 0.88037109375
                },
                {
                    "start": 1624.9,
                    "end": 1625.32,
                    "word": " case",
                    "probability": 0.8154296875
                },
                {
                    "start": 1625.32,
                    "end": 1627.0,
                    "word": ",",
                    "probability": 0.810546875
                }
            ]
        },
        {
            "id": 245,
            "seek": 149964,
            "start": 1627.0,
            "end": 1633.7,
            "text": " window sizes and some tuning things that I didn't go into in these slides because I've",
            "tokens": [
                51072,
                4910,
                11602,
                293,
                512,
                15164,
                721,
                300,
                286,
                994,
                380,
                352,
                666,
                294,
                613,
                9788,
                570,
                286,
                600,
                51380
            ],
            "temperature": 0.0,
            "avg_logprob": -0.22694360028679778,
            "compression_ratio": 1.558252427184466,
            "no_speech_prob": 0.01690673828125,
            "words": [
                {
                    "start": 1627.0,
                    "end": 1627.76,
                    "word": " window",
                    "probability": 0.7119140625
                },
                {
                    "start": 1627.76,
                    "end": 1628.36,
                    "word": " sizes",
                    "probability": 0.8193359375
                },
                {
                    "start": 1628.36,
                    "end": 1628.76,
                    "word": " and",
                    "probability": 0.8154296875
                },
                {
                    "start": 1628.76,
                    "end": 1629.22,
                    "word": " some",
                    "probability": 0.6103515625
                },
                {
                    "start": 1629.22,
                    "end": 1629.76,
                    "word": " tuning",
                    "probability": 0.78369140625
                },
                {
                    "start": 1629.76,
                    "end": 1630.32,
                    "word": " things",
                    "probability": 0.71875
                },
                {
                    "start": 1630.32,
                    "end": 1630.62,
                    "word": " that",
                    "probability": 0.83251953125
                },
                {
                    "start": 1630.62,
                    "end": 1631.54,
                    "word": " I",
                    "probability": 0.98828125
                },
                {
                    "start": 1631.54,
                    "end": 1631.78,
                    "word": " didn't",
                    "probability": 0.928466796875
                },
                {
                    "start": 1631.78,
                    "end": 1631.98,
                    "word": " go",
                    "probability": 0.8955078125
                },
                {
                    "start": 1631.98,
                    "end": 1632.3,
                    "word": " into",
                    "probability": 0.7177734375
                },
                {
                    "start": 1632.3,
                    "end": 1632.88,
                    "word": " in",
                    "probability": 0.51708984375
                },
                {
                    "start": 1632.88,
                    "end": 1633.04,
                    "word": " these",
                    "probability": 0.53271484375
                },
                {
                    "start": 1633.04,
                    "end": 1633.28,
                    "word": " slides",
                    "probability": 0.94384765625
                },
                {
                    "start": 1633.28,
                    "end": 1633.5,
                    "word": " because",
                    "probability": 0.60986328125
                },
                {
                    "start": 1633.5,
                    "end": 1633.7,
                    "word": " I've",
                    "probability": 0.608154296875
                }
            ]
        },
        {
            "id": 246,
            "seek": 149964,
            "start": 1633.7,
            "end": 1639.86,
            "text": " already got 20 minutes. There are some tuning systems you can use to really clean up and",
            "tokens": [
                51380,
                1217,
                658,
                945,
                2077,
                13,
                821,
                366,
                512,
                15164,
                3652,
                291,
                393,
                764,
                281,
                534,
                2541,
                493,
                293,
                51692
            ],
            "temperature": 0.0,
            "avg_logprob": -0.22694360028679778,
            "compression_ratio": 1.558252427184466,
            "no_speech_prob": 0.01690673828125,
            "words": [
                {
                    "start": 1633.7,
                    "end": 1633.84,
                    "word": " already",
                    "probability": 0.68603515625
                },
                {
                    "start": 1633.84,
                    "end": 1633.98,
                    "word": " got",
                    "probability": 0.8466796875
                },
                {
                    "start": 1633.98,
                    "end": 1634.18,
                    "word": " 20",
                    "probability": 0.85400390625
                },
                {
                    "start": 1634.18,
                    "end": 1634.42,
                    "word": " minutes",
                    "probability": 0.8681640625
                },
                {
                    "start": 1634.42,
                    "end": 1634.94,
                    "word": ".",
                    "probability": 0.8798828125
                },
                {
                    "start": 1635.06,
                    "end": 1635.48,
                    "word": " There",
                    "probability": 0.79296875
                },
                {
                    "start": 1635.48,
                    "end": 1635.58,
                    "word": " are",
                    "probability": 0.7958984375
                },
                {
                    "start": 1635.58,
                    "end": 1635.76,
                    "word": " some",
                    "probability": 0.72900390625
                },
                {
                    "start": 1635.76,
                    "end": 1636.04,
                    "word": " tuning",
                    "probability": 0.8076171875
                },
                {
                    "start": 1636.04,
                    "end": 1636.52,
                    "word": " systems",
                    "probability": 0.798828125
                },
                {
                    "start": 1636.52,
                    "end": 1636.78,
                    "word": " you",
                    "probability": 0.8662109375
                },
                {
                    "start": 1636.78,
                    "end": 1636.9,
                    "word": " can",
                    "probability": 0.89599609375
                },
                {
                    "start": 1636.9,
                    "end": 1637.12,
                    "word": " use",
                    "probability": 0.8251953125
                },
                {
                    "start": 1637.12,
                    "end": 1637.38,
                    "word": " to",
                    "probability": 0.90576171875
                },
                {
                    "start": 1637.38,
                    "end": 1637.74,
                    "word": " really",
                    "probability": 0.81005859375
                },
                {
                    "start": 1637.74,
                    "end": 1638.24,
                    "word": " clean",
                    "probability": 0.83935546875
                },
                {
                    "start": 1638.24,
                    "end": 1638.54,
                    "word": " up",
                    "probability": 0.9013671875
                },
                {
                    "start": 1638.54,
                    "end": 1639.86,
                    "word": " and",
                    "probability": 0.87890625
                }
            ]
        },
        {
            "id": 247,
            "seek": 152514,
            "start": 1640.1,
            "end": 1644.9,
            "text": " de-noise. There's also some interesting waking functions you can do inside the tree because",
            "tokens": [
                50364,
                368,
                12,
                42641,
                13,
                821,
                311,
                611,
                512,
                1880,
                20447,
                6828,
                291,
                393,
                360,
                1854,
                264,
                4230,
                570,
                50612
            ],
            "temperature": 0.0,
            "avg_logprob": -0.2682291546038219,
            "compression_ratio": 1.6200873362445414,
            "no_speech_prob": 0.0275726318359375,
            "words": [
                {
                    "start": 1640.1,
                    "end": 1640.5,
                    "word": " de-noise",
                    "probability": 0.796630859375
                },
                {
                    "start": 1640.5,
                    "end": 1640.94,
                    "word": ".",
                    "probability": 0.69921875
                },
                {
                    "start": 1640.94,
                    "end": 1641.42,
                    "word": " There's",
                    "probability": 0.8369140625
                },
                {
                    "start": 1641.42,
                    "end": 1641.72,
                    "word": " also",
                    "probability": 0.8203125
                },
                {
                    "start": 1641.72,
                    "end": 1642.28,
                    "word": " some",
                    "probability": 0.8173828125
                },
                {
                    "start": 1642.28,
                    "end": 1642.68,
                    "word": " interesting",
                    "probability": 0.5947265625
                },
                {
                    "start": 1642.68,
                    "end": 1643.02,
                    "word": " waking",
                    "probability": 0.68505859375
                },
                {
                    "start": 1643.02,
                    "end": 1643.5,
                    "word": " functions",
                    "probability": 0.84228515625
                },
                {
                    "start": 1643.5,
                    "end": 1643.72,
                    "word": " you",
                    "probability": 0.8330078125
                },
                {
                    "start": 1643.72,
                    "end": 1643.8,
                    "word": " can",
                    "probability": 0.89990234375
                },
                {
                    "start": 1643.8,
                    "end": 1643.92,
                    "word": " do",
                    "probability": 0.90283203125
                },
                {
                    "start": 1643.92,
                    "end": 1644.16,
                    "word": " inside",
                    "probability": 0.89111328125
                },
                {
                    "start": 1644.16,
                    "end": 1644.32,
                    "word": " the",
                    "probability": 0.8173828125
                },
                {
                    "start": 1644.32,
                    "end": 1644.54,
                    "word": " tree",
                    "probability": 0.80712890625
                },
                {
                    "start": 1644.54,
                    "end": 1644.9,
                    "word": " because",
                    "probability": 0.56591796875
                }
            ]
        },
        {
            "id": 248,
            "seek": 152514,
            "start": 1645.5,
            "end": 1656.9,
            "text": " there's some noise from an overactive leaf node might end up just building a very high KL",
            "tokens": [
                50668,
                456,
                311,
                512,
                5658,
                490,
                364,
                670,
                12596,
                10871,
                9984,
                1062,
                917,
                493,
                445,
                2390,
                257,
                588,
                1090,
                47991,
                51212
            ],
            "temperature": 0.0,
            "avg_logprob": -0.2682291546038219,
            "compression_ratio": 1.6200873362445414,
            "no_speech_prob": 0.0275726318359375,
            "words": [
                {
                    "start": 1645.5,
                    "end": 1645.98,
                    "word": " there's",
                    "probability": 0.5294189453125
                },
                {
                    "start": 1645.98,
                    "end": 1646.34,
                    "word": " some",
                    "probability": 0.8046875
                },
                {
                    "start": 1646.34,
                    "end": 1646.6,
                    "word": " noise",
                    "probability": 0.88818359375
                },
                {
                    "start": 1646.6,
                    "end": 1646.84,
                    "word": " from",
                    "probability": 0.8037109375
                },
                {
                    "start": 1646.84,
                    "end": 1647.2,
                    "word": " an",
                    "probability": 0.587890625
                },
                {
                    "start": 1647.2,
                    "end": 1648.06,
                    "word": " overactive",
                    "probability": 0.850341796875
                },
                {
                    "start": 1648.06,
                    "end": 1653.02,
                    "word": " leaf",
                    "probability": 0.7021484375
                },
                {
                    "start": 1653.02,
                    "end": 1653.28,
                    "word": " node",
                    "probability": 0.83154296875
                },
                {
                    "start": 1653.28,
                    "end": 1653.74,
                    "word": " might",
                    "probability": 0.295654296875
                },
                {
                    "start": 1653.74,
                    "end": 1654.74,
                    "word": " end",
                    "probability": 0.75927734375
                },
                {
                    "start": 1654.74,
                    "end": 1655.18,
                    "word": " up",
                    "probability": 0.9033203125
                },
                {
                    "start": 1655.18,
                    "end": 1655.56,
                    "word": " just",
                    "probability": 0.5478515625
                },
                {
                    "start": 1655.56,
                    "end": 1656.12,
                    "word": " building",
                    "probability": 0.8125
                },
                {
                    "start": 1656.24,
                    "end": 1656.38,
                    "word": " a very",
                    "probability": 0.72412109375
                },
                {
                    "start": 1656.38,
                    "end": 1656.68,
                    "word": " high",
                    "probability": 0.85693359375
                },
                {
                    "start": 1656.68,
                    "end": 1656.9,
                    "word": " KL",
                    "probability": 0.65673828125
                }
            ]
        },
        {
            "id": 249,
            "seek": 152514,
            "start": 1656.9,
            "end": 1661.38,
            "text": " divergence for just that one node and it throws everything off. And there's a couple other little",
            "tokens": [
                51212,
                47387,
                337,
                445,
                300,
                472,
                9984,
                293,
                309,
                19251,
                1203,
                766,
                13,
                400,
                456,
                311,
                257,
                1916,
                661,
                707,
                51436
            ],
            "temperature": 0.0,
            "avg_logprob": -0.2682291546038219,
            "compression_ratio": 1.6200873362445414,
            "no_speech_prob": 0.0275726318359375,
            "words": [
                {
                    "start": 1656.9,
                    "end": 1657.26,
                    "word": " divergence",
                    "probability": 0.53271484375
                },
                {
                    "start": 1657.26,
                    "end": 1657.46,
                    "word": " for",
                    "probability": 0.398193359375
                },
                {
                    "start": 1657.46,
                    "end": 1657.72,
                    "word": " just",
                    "probability": 0.517578125
                },
                {
                    "start": 1657.72,
                    "end": 1657.88,
                    "word": " that",
                    "probability": 0.87646484375
                },
                {
                    "start": 1657.88,
                    "end": 1658.08,
                    "word": " one",
                    "probability": 0.80322265625
                },
                {
                    "start": 1658.08,
                    "end": 1658.3,
                    "word": " node",
                    "probability": 0.814453125
                },
                {
                    "start": 1658.3,
                    "end": 1658.44,
                    "word": " and",
                    "probability": 0.74609375
                },
                {
                    "start": 1658.44,
                    "end": 1658.56,
                    "word": " it",
                    "probability": 0.86328125
                },
                {
                    "start": 1658.56,
                    "end": 1658.76,
                    "word": " throws",
                    "probability": 0.77294921875
                },
                {
                    "start": 1658.76,
                    "end": 1659.1,
                    "word": " everything",
                    "probability": 0.8798828125
                },
                {
                    "start": 1659.1,
                    "end": 1659.4,
                    "word": " off",
                    "probability": 0.81884765625
                },
                {
                    "start": 1659.4,
                    "end": 1659.88,
                    "word": ".",
                    "probability": 0.8662109375
                },
                {
                    "start": 1660.3,
                    "end": 1660.48,
                    "word": " And",
                    "probability": 0.67529296875
                },
                {
                    "start": 1660.48,
                    "end": 1660.6,
                    "word": " there's",
                    "probability": 0.916259765625
                },
                {
                    "start": 1660.7,
                    "end": 1660.82,
                    "word": " a couple",
                    "probability": 0.787109375
                },
                {
                    "start": 1660.82,
                    "end": 1661.04,
                    "word": " other",
                    "probability": 0.6083984375
                },
                {
                    "start": 1661.04,
                    "end": 1661.38,
                    "word": " little",
                    "probability": 0.8046875
                }
            ]
        },
        {
            "id": 250,
            "seek": 152514,
            "start": 1661.38,
            "end": 1665.36,
            "text": " things like that that still need to be cleaned up before this goes into production. But the",
            "tokens": [
                51436,
                721,
                411,
                300,
                300,
                920,
                643,
                281,
                312,
                16146,
                493,
                949,
                341,
                1709,
                666,
                4265,
                13,
                583,
                264,
                51636
            ],
            "temperature": 0.0,
            "avg_logprob": -0.2682291546038219,
            "compression_ratio": 1.6200873362445414,
            "no_speech_prob": 0.0275726318359375,
            "words": [
                {
                    "start": 1661.38,
                    "end": 1661.62,
                    "word": " things",
                    "probability": 0.73681640625
                },
                {
                    "start": 1661.62,
                    "end": 1661.78,
                    "word": " like",
                    "probability": 0.865234375
                },
                {
                    "start": 1661.78,
                    "end": 1662.02,
                    "word": " that",
                    "probability": 0.8818359375
                },
                {
                    "start": 1662.02,
                    "end": 1662.24,
                    "word": " that",
                    "probability": 0.77978515625
                },
                {
                    "start": 1662.24,
                    "end": 1662.46,
                    "word": " still",
                    "probability": 0.9111328125
                },
                {
                    "start": 1662.46,
                    "end": 1662.62,
                    "word": " need",
                    "probability": 0.79833984375
                },
                {
                    "start": 1662.62,
                    "end": 1662.76,
                    "word": " to",
                    "probability": 0.90966796875
                },
                {
                    "start": 1662.76,
                    "end": 1662.98,
                    "word": " be",
                    "probability": 0.90283203125
                },
                {
                    "start": 1662.98,
                    "end": 1663.24,
                    "word": " cleaned",
                    "probability": 0.755859375
                },
                {
                    "start": 1663.24,
                    "end": 1663.46,
                    "word": " up",
                    "probability": 0.9052734375
                },
                {
                    "start": 1663.46,
                    "end": 1663.68,
                    "word": " before",
                    "probability": 0.763671875
                },
                {
                    "start": 1663.68,
                    "end": 1663.84,
                    "word": " this",
                    "probability": 0.87646484375
                },
                {
                    "start": 1663.84,
                    "end": 1664.0,
                    "word": " goes",
                    "probability": 0.8544921875
                },
                {
                    "start": 1664.0,
                    "end": 1664.16,
                    "word": " into",
                    "probability": 0.6435546875
                },
                {
                    "start": 1664.16,
                    "end": 1664.5,
                    "word": " production",
                    "probability": 0.84375
                },
                {
                    "start": 1664.5,
                    "end": 1664.98,
                    "word": ".",
                    "probability": 0.313232421875
                },
                {
                    "start": 1665.0,
                    "end": 1665.12,
                    "word": " But",
                    "probability": 0.78515625
                },
                {
                    "start": 1665.12,
                    "end": 1665.36,
                    "word": " the",
                    "probability": 0.81982421875
                }
            ]
        },
        {
            "id": 251,
            "seek": 154600,
            "start": 1665.36,
            "end": 1671.56,
            "text": " system itself is looking very promising and is honestly the fastest of its kind",
            "tokens": [
                50364,
                1185,
                2564,
                307,
                1237,
                588,
                20257,
                293,
                307,
                6095,
                264,
                14573,
                295,
                1080,
                733,
                50668
            ],
            "temperature": 0.0,
            "avg_logprob": -0.2769097222222222,
            "compression_ratio": 1.3587786259541985,
            "no_speech_prob": 0.04534912109375,
            "words": [
                {
                    "start": 1665.36,
                    "end": 1665.68,
                    "word": " system",
                    "probability": 0.2568359375
                },
                {
                    "start": 1665.68,
                    "end": 1666.02,
                    "word": " itself",
                    "probability": 0.72705078125
                },
                {
                    "start": 1666.02,
                    "end": 1666.16,
                    "word": " is",
                    "probability": 0.88623046875
                },
                {
                    "start": 1666.16,
                    "end": 1666.44,
                    "word": " looking",
                    "probability": 0.80810546875
                },
                {
                    "start": 1666.44,
                    "end": 1666.74,
                    "word": " very",
                    "probability": 0.74365234375
                },
                {
                    "start": 1666.74,
                    "end": 1667.08,
                    "word": " promising",
                    "probability": 0.85546875
                },
                {
                    "start": 1667.08,
                    "end": 1668.52,
                    "word": " and",
                    "probability": 0.428955078125
                },
                {
                    "start": 1668.52,
                    "end": 1669.22,
                    "word": " is",
                    "probability": 0.79833984375
                },
                {
                    "start": 1669.22,
                    "end": 1669.54,
                    "word": " honestly",
                    "probability": 0.75390625
                },
                {
                    "start": 1669.54,
                    "end": 1669.86,
                    "word": " the",
                    "probability": 0.798828125
                },
                {
                    "start": 1669.86,
                    "end": 1670.64,
                    "word": " fastest",
                    "probability": 0.720703125
                },
                {
                    "start": 1670.64,
                    "end": 1670.96,
                    "word": " of",
                    "probability": 0.89599609375
                },
                {
                    "start": 1670.96,
                    "end": 1671.12,
                    "word": " its",
                    "probability": 0.7861328125
                },
                {
                    "start": 1671.12,
                    "end": 1671.56,
                    "word": " kind",
                    "probability": 0.8134765625
                }
            ]
        },
        {
            "id": 252,
            "seek": 154600,
            "start": 1672.32,
            "end": 1686.22,
            "text": " that I can see in the world. And that's basically it for Gokul's ability to detect benign test set",
            "tokens": [
                50716,
                300,
                286,
                393,
                536,
                294,
                264,
                1002,
                13,
                400,
                300,
                311,
                1936,
                309,
                337,
                460,
                453,
                425,
                311,
                3485,
                281,
                5531,
                3271,
                788,
                1500,
                992,
                51408
            ],
            "temperature": 0.0,
            "avg_logprob": -0.2769097222222222,
            "compression_ratio": 1.3587786259541985,
            "no_speech_prob": 0.04534912109375,
            "words": [
                {
                    "start": 1672.32,
                    "end": 1672.66,
                    "word": " that",
                    "probability": 0.80078125
                },
                {
                    "start": 1672.66,
                    "end": 1672.78,
                    "word": " I",
                    "probability": 0.943359375
                },
                {
                    "start": 1672.78,
                    "end": 1673.22,
                    "word": " can",
                    "probability": 0.6396484375
                },
                {
                    "start": 1673.22,
                    "end": 1673.38,
                    "word": " see",
                    "probability": 0.82421875
                },
                {
                    "start": 1673.38,
                    "end": 1673.48,
                    "word": " in",
                    "probability": 0.90673828125
                },
                {
                    "start": 1673.48,
                    "end": 1673.54,
                    "word": " the",
                    "probability": 0.8134765625
                },
                {
                    "start": 1673.54,
                    "end": 1673.76,
                    "word": " world",
                    "probability": 0.88916015625
                },
                {
                    "start": 1673.76,
                    "end": 1674.44,
                    "word": ".",
                    "probability": 0.83203125
                },
                {
                    "start": 1675.9,
                    "end": 1676.46,
                    "word": " And",
                    "probability": 0.67822265625
                },
                {
                    "start": 1676.46,
                    "end": 1677.0,
                    "word": " that's",
                    "probability": 0.925537109375
                },
                {
                    "start": 1677.0,
                    "end": 1677.38,
                    "word": " basically",
                    "probability": 0.88623046875
                },
                {
                    "start": 1677.38,
                    "end": 1677.66,
                    "word": " it",
                    "probability": 0.90283203125
                },
                {
                    "start": 1677.66,
                    "end": 1678.72,
                    "word": " for",
                    "probability": 0.83642578125
                },
                {
                    "start": 1678.72,
                    "end": 1680.22,
                    "word": " Gokul's",
                    "probability": 0.736083984375
                },
                {
                    "start": 1680.22,
                    "end": 1683.38,
                    "word": " ability",
                    "probability": 0.84912109375
                },
                {
                    "start": 1683.38,
                    "end": 1683.6,
                    "word": " to",
                    "probability": 0.908203125
                },
                {
                    "start": 1683.6,
                    "end": 1684.0,
                    "word": " detect",
                    "probability": 0.93505859375
                },
                {
                    "start": 1684.0,
                    "end": 1685.42,
                    "word": " benign",
                    "probability": 0.83154296875
                },
                {
                    "start": 1685.42,
                    "end": 1686.08,
                    "word": " test",
                    "probability": 0.58349609375
                },
                {
                    "start": 1686.08,
                    "end": 1686.22,
                    "word": " set",
                    "probability": 0.794921875
                }
            ]
        },
        {
            "id": 253,
            "seek": 157072,
            "start": 1686.22,
            "end": 1693.48,
            "text": " and other drift things. The library is at elastic.com, oh sorry, github slash elastic",
            "tokens": [
                50412,
                293,
                661,
                19699,
                721,
                13,
                440,
                6405,
                307,
                412,
                806,
                525,
                299,
                13,
                1112,
                11,
                1954,
                2597,
                11,
                290,
                355,
                836,
                17330,
                17115,
                50728
            ],
            "temperature": 0.0,
            "avg_logprob": -0.3323384885037883,
            "compression_ratio": 1.5323383084577114,
            "no_speech_prob": 0.8076171875,
            "words": [
                {
                    "start": 1686.22,
                    "end": 1686.8,
                    "word": " and",
                    "probability": 0.019927978515625
                },
                {
                    "start": 1686.8,
                    "end": 1687.9,
                    "word": " other",
                    "probability": 0.465087890625
                },
                {
                    "start": 1687.9,
                    "end": 1688.24,
                    "word": " drift",
                    "probability": 0.255859375
                },
                {
                    "start": 1688.24,
                    "end": 1688.6,
                    "word": " things",
                    "probability": 0.70849609375
                },
                {
                    "start": 1688.6,
                    "end": 1689.18,
                    "word": ".",
                    "probability": 0.68505859375
                },
                {
                    "start": 1689.18,
                    "end": 1689.58,
                    "word": " The",
                    "probability": 0.72314453125
                },
                {
                    "start": 1689.58,
                    "end": 1689.86,
                    "word": " library",
                    "probability": 0.7392578125
                },
                {
                    "start": 1689.86,
                    "end": 1690.04,
                    "word": " is",
                    "probability": 0.65966796875
                },
                {
                    "start": 1690.04,
                    "end": 1690.28,
                    "word": " at",
                    "probability": 0.91015625
                },
                {
                    "start": 1690.28,
                    "end": 1690.8,
                    "word": " elastic",
                    "probability": 0.7063802083333334
                },
                {
                    "start": 1690.8,
                    "end": 1691.4,
                    "word": ".com",
                    "probability": 0.891357421875
                },
                {
                    "start": 1691.4,
                    "end": 1691.54,
                    "word": ",",
                    "probability": 0.3427734375
                },
                {
                    "start": 1691.54,
                    "end": 1691.66,
                    "word": " oh",
                    "probability": 0.57421875
                },
                {
                    "start": 1691.66,
                    "end": 1691.82,
                    "word": " sorry",
                    "probability": 0.59130859375
                },
                {
                    "start": 1691.82,
                    "end": 1692.26,
                    "word": ",",
                    "probability": 0.787109375
                },
                {
                    "start": 1692.26,
                    "end": 1692.56,
                    "word": " github",
                    "probability": 0.8971354166666666
                },
                {
                    "start": 1692.56,
                    "end": 1693.06,
                    "word": " slash",
                    "probability": 0.060882568359375
                },
                {
                    "start": 1693.06,
                    "end": 1693.48,
                    "word": " elastic",
                    "probability": 0.6044921875
                }
            ]
        },
        {
            "id": 254,
            "seek": 157072,
            "start": 1694.06,
            "end": 1701.04,
            "text": " slash Gokul. It's named after my grandmother because the base algorithm for this thing is",
            "tokens": [
                50756,
                17330,
                460,
                453,
                425,
                13,
                467,
                311,
                4926,
                934,
                452,
                14317,
                570,
                264,
                3096,
                9284,
                337,
                341,
                551,
                307,
                51100
            ],
            "temperature": 0.0,
            "avg_logprob": -0.3323384885037883,
            "compression_ratio": 1.5323383084577114,
            "no_speech_prob": 0.8076171875,
            "words": [
                {
                    "start": 1694.06,
                    "end": 1694.56,
                    "word": " slash",
                    "probability": 0.74755859375
                },
                {
                    "start": 1694.56,
                    "end": 1694.94,
                    "word": " Gokul",
                    "probability": 0.4149576822916667
                },
                {
                    "start": 1694.94,
                    "end": 1695.54,
                    "word": ".",
                    "probability": 0.82568359375
                },
                {
                    "start": 1695.88,
                    "end": 1696.48,
                    "word": " It's",
                    "probability": 0.94091796875
                },
                {
                    "start": 1696.48,
                    "end": 1696.66,
                    "word": " named",
                    "probability": 0.783203125
                },
                {
                    "start": 1696.66,
                    "end": 1696.86,
                    "word": " after",
                    "probability": 0.720703125
                },
                {
                    "start": 1696.86,
                    "end": 1697.06,
                    "word": " my",
                    "probability": 0.90869140625
                },
                {
                    "start": 1697.06,
                    "end": 1697.34,
                    "word": " grandmother",
                    "probability": 0.58984375
                },
                {
                    "start": 1697.34,
                    "end": 1698.18,
                    "word": " because",
                    "probability": 0.57421875
                },
                {
                    "start": 1698.18,
                    "end": 1699.0,
                    "word": " the",
                    "probability": 0.8037109375
                },
                {
                    "start": 1699.0,
                    "end": 1699.96,
                    "word": " base",
                    "probability": 0.8046875
                },
                {
                    "start": 1699.96,
                    "end": 1700.36,
                    "word": " algorithm",
                    "probability": 0.81591796875
                },
                {
                    "start": 1700.36,
                    "end": 1700.58,
                    "word": " for",
                    "probability": 0.892578125
                },
                {
                    "start": 1700.58,
                    "end": 1700.74,
                    "word": " this",
                    "probability": 0.876953125
                },
                {
                    "start": 1700.74,
                    "end": 1700.9,
                    "word": " thing",
                    "probability": 0.80859375
                },
                {
                    "start": 1700.9,
                    "end": 1701.04,
                    "word": " is",
                    "probability": 0.9033203125
                }
            ]
        },
        {
            "id": 255,
            "seek": 157072,
            "start": 1701.04,
            "end": 1704.82,
            "text": " called GRMA. And if you say that while drunk, you can kind of get grandma.",
            "tokens": [
                51100,
                1219,
                10903,
                9998,
                13,
                400,
                498,
                291,
                584,
                300,
                1339,
                11192,
                11,
                291,
                393,
                733,
                295,
                483,
                15766,
                13,
                51264
            ],
            "temperature": 0.0,
            "avg_logprob": -0.3323384885037883,
            "compression_ratio": 1.5323383084577114,
            "no_speech_prob": 0.8076171875,
            "words": [
                {
                    "start": 1701.04,
                    "end": 1701.36,
                    "word": " called",
                    "probability": 0.78857421875
                },
                {
                    "start": 1701.36,
                    "end": 1701.92,
                    "word": " GRMA",
                    "probability": 0.776123046875
                },
                {
                    "start": 1701.92,
                    "end": 1702.18,
                    "word": ".",
                    "probability": 0.2381591796875
                },
                {
                    "start": 1702.18,
                    "end": 1702.26,
                    "word": " And",
                    "probability": 0.74658203125
                },
                {
                    "start": 1702.26,
                    "end": 1702.36,
                    "word": " if",
                    "probability": 0.892578125
                },
                {
                    "start": 1702.36,
                    "end": 1702.44,
                    "word": " you",
                    "probability": 0.89404296875
                },
                {
                    "start": 1702.44,
                    "end": 1702.58,
                    "word": " say",
                    "probability": 0.88916015625
                },
                {
                    "start": 1702.58,
                    "end": 1702.76,
                    "word": " that",
                    "probability": 0.8798828125
                },
                {
                    "start": 1702.76,
                    "end": 1702.88,
                    "word": " while",
                    "probability": 0.87890625
                },
                {
                    "start": 1702.88,
                    "end": 1703.16,
                    "word": " drunk",
                    "probability": 0.85302734375
                },
                {
                    "start": 1703.16,
                    "end": 1703.24,
                    "word": ",",
                    "probability": 0.6875
                },
                {
                    "start": 1703.24,
                    "end": 1703.38,
                    "word": " you",
                    "probability": 0.8994140625
                },
                {
                    "start": 1703.38,
                    "end": 1703.54,
                    "word": " can",
                    "probability": 0.89306640625
                },
                {
                    "start": 1703.54,
                    "end": 1703.68,
                    "word": " kind",
                    "probability": 0.311767578125
                },
                {
                    "start": 1703.68,
                    "end": 1703.74,
                    "word": " of",
                    "probability": 0.90966796875
                },
                {
                    "start": 1703.74,
                    "end": 1703.88,
                    "word": " get",
                    "probability": 0.89208984375
                },
                {
                    "start": 1703.88,
                    "end": 1704.22,
                    "word": " grandma",
                    "probability": 0.7041015625
                },
                {
                    "start": 1704.22,
                    "end": 1704.82,
                    "word": ".",
                    "probability": 0.65185546875
                }
            ]
        },
        {
            "id": 256,
            "seek": 157072,
            "start": 1705.16,
            "end": 1710.94,
            "text": " And grandma's a bit too on the nose, so I named it Gokul.",
            "tokens": [
                51320,
                400,
                15766,
                311,
                257,
                857,
                886,
                322,
                264,
                6690,
                11,
                370,
                286,
                4926,
                309,
                460,
                453,
                425,
                13,
                51568
            ],
            "temperature": 0.0,
            "avg_logprob": -0.3323384885037883,
            "compression_ratio": 1.5323383084577114,
            "no_speech_prob": 0.8076171875,
            "words": [
                {
                    "start": 1705.16,
                    "end": 1705.64,
                    "word": " And",
                    "probability": 0.8720703125
                },
                {
                    "start": 1705.64,
                    "end": 1706.06,
                    "word": " grandma's",
                    "probability": 0.651611328125
                },
                {
                    "start": 1706.14,
                    "end": 1706.8,
                    "word": " a bit",
                    "probability": 0.88916015625
                },
                {
                    "start": 1706.8,
                    "end": 1707.28,
                    "word": " too",
                    "probability": 0.37255859375
                },
                {
                    "start": 1707.28,
                    "end": 1707.46,
                    "word": " on",
                    "probability": 0.88720703125
                },
                {
                    "start": 1707.46,
                    "end": 1707.54,
                    "word": " the",
                    "probability": 0.80322265625
                },
                {
                    "start": 1707.54,
                    "end": 1707.84,
                    "word": " nose",
                    "probability": 0.83740234375
                },
                {
                    "start": 1707.84,
                    "end": 1708.16,
                    "word": ",",
                    "probability": 0.53515625
                },
                {
                    "start": 1708.16,
                    "end": 1708.16,
                    "word": " so",
                    "probability": 0.85986328125
                },
                {
                    "start": 1708.16,
                    "end": 1708.52,
                    "word": " I",
                    "probability": 0.7373046875
                },
                {
                    "start": 1708.52,
                    "end": 1709.9,
                    "word": " named",
                    "probability": 0.79052734375
                },
                {
                    "start": 1709.9,
                    "end": 1710.02,
                    "word": " it",
                    "probability": 0.529296875
                },
                {
                    "start": 1710.02,
                    "end": 1710.34,
                    "word": " Gokul",
                    "probability": 0.9646809895833334
                },
                {
                    "start": 1710.34,
                    "end": 1710.94,
                    "word": ".",
                    "probability": 0.99267578125
                }
            ]
        },
        {
            "id": 257,
            "seek": 158177,
            "start": 1712.68,
            "end": 1715.42,
            "text": " Anyway, well, looking forward to hearing you guys on Twitter",
            "tokens": [
                50412,
                5684,
                11,
                731,
                11,
                1237,
                2128,
                281,
                4763,
                291,
                1074,
                322,
                5794,
                50600
            ],
            "temperature": 0.0,
            "avg_logprob": -0.3712121212121212,
            "compression_ratio": 1.2547169811320755,
            "no_speech_prob": 0.053375244140625,
            "words": [
                {
                    "start": 1712.68,
                    "end": 1713.14,
                    "word": " Anyway",
                    "probability": 0.1943359375
                },
                {
                    "start": 1713.14,
                    "end": 1713.26,
                    "word": ",",
                    "probability": 0.76171875
                },
                {
                    "start": 1713.26,
                    "end": 1713.38,
                    "word": " well",
                    "probability": 0.7998046875
                },
                {
                    "start": 1713.38,
                    "end": 1713.6,
                    "word": ",",
                    "probability": 0.68994140625
                },
                {
                    "start": 1713.6,
                    "end": 1713.94,
                    "word": " looking",
                    "probability": 0.72705078125
                },
                {
                    "start": 1713.94,
                    "end": 1714.28,
                    "word": " forward",
                    "probability": 0.84375
                },
                {
                    "start": 1714.28,
                    "end": 1714.48,
                    "word": " to",
                    "probability": 0.904296875
                },
                {
                    "start": 1714.48,
                    "end": 1714.72,
                    "word": " hearing",
                    "probability": 0.802734375
                },
                {
                    "start": 1714.72,
                    "end": 1714.88,
                    "word": " you",
                    "probability": 0.8408203125
                },
                {
                    "start": 1714.88,
                    "end": 1715.06,
                    "word": " guys",
                    "probability": 0.8828125
                },
                {
                    "start": 1715.06,
                    "end": 1715.2,
                    "word": " on",
                    "probability": 0.89599609375
                },
                {
                    "start": 1715.2,
                    "end": 1715.42,
                    "word": " Twitter",
                    "probability": 0.802734375
                }
            ]
        },
        {
            "id": 258,
            "seek": 158177,
            "start": 1715.42,
            "end": 1720.98,
            "text": " and Twitch and Slack or whatever. So thank you very much and good night.",
            "tokens": [
                50600,
                293,
                22222,
                293,
                37211,
                420,
                2035,
                13,
                407,
                1309,
                291,
                588,
                709,
                293,
                665,
                1818,
                13,
                50868
            ],
            "temperature": 0.0,
            "avg_logprob": -0.3712121212121212,
            "compression_ratio": 1.2547169811320755,
            "no_speech_prob": 0.053375244140625,
            "words": [
                {
                    "start": 1715.42,
                    "end": 1716.12,
                    "word": " and",
                    "probability": 0.55078125
                },
                {
                    "start": 1716.12,
                    "end": 1716.42,
                    "word": " Twitch",
                    "probability": 0.90283203125
                },
                {
                    "start": 1716.42,
                    "end": 1717.0,
                    "word": " and",
                    "probability": 0.826171875
                },
                {
                    "start": 1717.0,
                    "end": 1717.58,
                    "word": " Slack",
                    "probability": 0.82470703125
                },
                {
                    "start": 1717.58,
                    "end": 1717.88,
                    "word": " or",
                    "probability": 0.55078125
                },
                {
                    "start": 1717.88,
                    "end": 1718.18,
                    "word": " whatever",
                    "probability": 0.76318359375
                },
                {
                    "start": 1718.18,
                    "end": 1718.64,
                    "word": ".",
                    "probability": 0.6455078125
                },
                {
                    "start": 1718.98,
                    "end": 1719.1,
                    "word": " So",
                    "probability": 0.79736328125
                },
                {
                    "start": 1719.1,
                    "end": 1719.44,
                    "word": " thank",
                    "probability": 0.55517578125
                },
                {
                    "start": 1719.44,
                    "end": 1719.52,
                    "word": " you",
                    "probability": 0.89794921875
                },
                {
                    "start": 1719.52,
                    "end": 1719.62,
                    "word": " very",
                    "probability": 0.7275390625
                },
                {
                    "start": 1719.62,
                    "end": 1719.8,
                    "word": " much",
                    "probability": 0.81396484375
                },
                {
                    "start": 1719.8,
                    "end": 1719.98,
                    "word": " and",
                    "probability": 0.7470703125
                },
                {
                    "start": 1719.98,
                    "end": 1720.3,
                    "word": " good",
                    "probability": 0.91259765625
                },
                {
                    "start": 1720.3,
                    "end": 1720.52,
                    "word": " night",
                    "probability": 0.58447265625
                },
                {
                    "start": 1720.52,
                    "end": 1720.98,
                    "word": ".",
                    "probability": 0.87451171875
                }
            ]
        }
    ],
    "text": " Hello everyone and welcome to our next talk which is mine. To introduce myself, which is wrong, I did a postdoc in machine learning after getting my PhD in algebraic topology with some of the people who did some of the math behind the concepts in this talk and I've just extended it for this thing. So that's enough about me. Let me play my talk back for you. So, here we go. I welcome everyone to my talk on calculating concepts very quickly with this library called Google which is based on... I hear no audio. Oopsie daisy. Um, there's no... do you hear audio from the talk? Okay, well that's the problem I'm trying to... Uh, sorry about this guys. Uh, sorry, I don't know what's going on. Um, I am... I'll apologize. Key takeaway for this talk is this is a novel way of calculating drift. Welcome everyone to my talk on calculating concept drift very quickly with this library called Google which is based on cover trees. My name is Sven Kittel. I am a senior data scientist at Elastic and my Twitter handle is co-mathematician. So, the key outline... key takeaway for this talk is this is a novel way of calculating drift. It's done using a new technique using cover trees. And because it's so fast, it's log-in time to actually calculate drift for a new sample. We can use it for on... in new ways that concept drift couldn't have been applied before now. So, we're going to go over the objective, why we care about concept drift, mathematically how it's done, then a couple results, and then finishing up with next steps. So, why do we care about data set drifts? Well, our data drifts. So, I work with malware and there are new malware families coming out every month. There are new benign pieces, examples of software coming out every month. Every time that Adobe patches Photoshop, it changes the pattern slightly and it changes its location slightly in our data set. Every single time a new piece of malware comes up, it's a new... it changes the way our... the malware data is distributed slightly. Another problem... well, what those do, basically, which gets into the second problem, they lower the efficacy of our models that we deploy. So, our models expect data in a sort of the same sort of general location as our training data. And every time we deploy, we are kind of fixing our model's view of the world at a certain day, and then we're training it and we're testing it on essentially future data. So, it doesn't have a perspective for the... if it's trained in March, it doesn't have... it might not work very well at the end... data coming in at the end of April. So, one of the methods for dealing with this is you build a dashboard that basically tracks the error. And when the error gets too high, you discard the model and go on for a new one. But the problem is, I don't trust VirusTotalLabels, and I don't think anyone really should trust VirusTotalLabels. And that's the best way we have because of the sheer volume of data we have to kind of trust VirusTotalLabels. So, and for other things, I wouldn't trust early labels on data because we have to move very quickly and sometimes with whitelisting and stuff. And, you know, well, and sometimes we only get like a label when our customer complains and we'd rather get ahead of that and figure out like, hey, where's the drift happening? What's going on? And then additionally, aside from the fact that things are just changing all the time, our models are also under attack. People are trying to bypass them by doing new and weird things to their data, like packing a spam filter and seeing if that works, or like posting a different pattern to see if that works, to see if they can get past like the Facebook spam filter. They're constantly innovating with a specific goal in mind of bypassing our models. This may not result be an adversarial example per se from the literature, but it sure kind of acts like that sometimes. And additionally, we could be under attack via a poisoning system. Maybe if the detector is fast enough, we can get ahead of those things beforehand. So here's a trivial example. So we have these two data sets. One is our training data set. That's we've got 1, 1, 2, 1, 1. And so our training data set has 70% 1s, 20% 2s, and 10% 3s. And this is what we trained on. Realistically, there's no point in training a machine learning model on a sequence of numbers, but suppose we do. But then we deploy this model, and we get this other sequence of data. 2, 2, 1, 2, so on, so on. And when we actually deploy it, we see data coming over the wire. We have 30% 1s, 60% 2s, and 10% 3s. So that's a little bit different from our training set. So the real world is different from our training set. So in this case, this is a... we can model this distribution very easily because it's a discrete space. There are only three things that our data could be. There's only... it isn't continuous like a lot of the stuff we have to deal with in data science. So we can actually compute the distribution fairly easily. And this is a pretty good... we can feel confident in our estimate of the distribution. We have to do some Bayesian statistics to do this properly, which we're not getting into this talk. But this is basically how it works. And then once we have like a concept of our training distribution, and our real world distribution, or our test distribution, as I'll be calling it today, we can take the Kublai-Klepch divergence. So this is a sort of a measure of the distance between this distribution and this distribution. And it's very easy to calculate on this categorical discrete data because it's just going to be, well, 0.7 times log of 0.7 over 0.3. So that's the category for 1s over the 1s. Take the log and then multiply it for the category for 1s. And that gives you the first term. Second term is the same thing for 2s. And the third term is the same thing for 3s. And then we add those all up and you get the KL divergence between these two distributions is 0.16. So that's all great. But the problem is our data looks like this. And I don't know whether this has any difference. Are my blue points sampled from the same distribution as my orange points? I don't know. It's really hard to tell by just looking at this picture. And now, in this case, they are. So my orange points are sampled from a two-dimensional Gaussian, two-dimensional normal distribution, where it's just, it's a nice sphere. And they're both, the covariance matrix is 1, 1. And the covariance matrix with blues is 1, 1. And it gives you this nice distribution. And it looks okay. Now, what I'm going to do is I'm going to drift the blues slightly over. And can you tell, is this correct? Well, you know, every single time that I put a blue, it's right next to some orange points. So this looks like it could be from the distribution. I can't really tell with my eyes. And one of the techniques we might do is like, oh, well, let me take the k-nearest neighbors and see if the distance from my k-nearest neighbors over time goes up, then I'm out of distribution. But if you tell here, that's not going to tell you much. If I go here, well, oh, I've got some outliers here now. So maybe that will work now. If I go here, well, this is really distributed. So this center, the center of this distribution should be around here. And the center for this distribution is around here. So that's really drifted. But you've got some outliers over here. Now, what happens when there's like millions of orange points and there's only a few hundred blue points? Well, you might not be able to tell the difference with that in that case. So, and also like, there's all this, what I've described to you is kind of a feeling. And to actually get at the math of this thing, I have to like model the distributions that these came from. And that's quite a hard problem. And especially in high dimensions. So things get kind of complicated. So the solution to this that I came up with is to use a cover tree. Now, why do I want to use a cover tree? So the cover tree, basically, the key takeaway for cover trees is it's a k-nearest neighbors data structure. There are many like it. There's kd trees, there's b trees, there's kd trees, you know, there's k-means trees, a whole bunch. And, well, the cover tree has this wonderful thing where in 2016, Mario Maggiorni and Wenjing Lau proved that it can arbitrarily well approximate the underlying distribution of the data. There's some caveats to that statement, and mathematically, it will probably, to properly statement, will take it, take the whole page. But basically, the whole concept of their proof is for a nice data set, and to know what a nice data set is, you need to know what a low dimensional manifold is. And you have to have enough data from that. But for a nice data set, which most of our data sets are pretty nice, sort of, kind of, a cover tree will arbitrarily well approximate the underlying distribution of the data. There's some caveats, and you can go read their paper if you want on exactly what that means. But for us, what this means is, if I use a cover tree to build a model of my data for k-nearest neighbors, I can take k-nearest neighbors, or I can just fiddle around with the properties of the cover tree to, like, get information about my data set. So, I can tell things like the local dimensionality, I can tell things like how things glue together, and I can tell, like, sort of, the shape of my data fairly well. I can tell clusters and things. I can infer what clusters are. Let's go build a very simple cover tree in two-dimensional data. So, here's a, you know, infinity sign, a bow tie, whatever you want to call it, and I start my cover tree by picking a point at random and then building a sphere. You know, in this case, it's a circle that covers everything. So, this is the start. And now, this doesn't help me that much. I haven't split up my data geometrically to, like, divide and conquer the k-nearest neighbors structure, because the naive way would take a linear time, big O of n, but we want to divide and conquer, so it takes log of n. So, here's what I do. I shrink my sphere down, add another one, and I cover it again. You can kind of tell that this object is longer than it is wide. It's kind of, if you really squint, it's kind of a one-dimensional thing in this direction. So, that's great. But now, if I go, okay, well, I'm going to split it up further, so really divide and conquer, and I, so I've got these two nodes, and they're both children of the previous node on the previous slide, and I'm going to build their children. So, here's their children. So, I divide and conquer their stuff, and I can kind of tell that each of their children have four, and if I, if you squint a little bit, it's kind of square-shaped in this dimension. At the scale, the top scale, it was too blurry. I could only see a line here. At the scale, I can kind of see that it's up in, it's got a nice shape, square shape, and so there's four children, and the reason, you know, that that dimensionality is relevant to the number of children you have. And then I can go further, and you know, it's got some more, it's kind of split up nicely, and I've got like a one-dimensional object here again, because all these children are kind of one-dimensional-ish, and I can split up some more, and you can see that it's kind of conforming to your data nicely, and at every single step of the way, I can make some inferences about what the shape of my data is. They're very cheap, because it's basically counting. So how do we build a probability distribution? So on slide two, we built a very simple probability distribution, where I had a discrete space, and now I've got a tree, and essentially, when I have a tree, at every single node, I can either go left, or right, or down, or you know, in this case, my number of children I have is variable, but I will be at a parent node, and I have to go to one of its children. So I always have discrete choices, a discrete choice of where to go. So how does that apply? So here is a very, like, simple cover tree, and I've colored it by the total, like, the population of this node. So this color here is the relative population of this node. So about 66% of the data is underneath this node. About 33% of the data is underneath this node. And you can see that colors are kind of changing. I get a kind of a picture of how dense my data is, but that doesn't help me, because I need to make choices. It doesn't help me make the choices at all steps. So if I go here, this helps me make my choices. So here, I've got 66% of my data over here, and I've got 33% of my data over here. And then once I've taken... I have to take one of those two choices. If I take this choice, I'm now at this node, and I have two choices again. And the relative probability of taking this choice versus this one... Well, this is basically a 0.1 probability of taking this path, and this is a 0.9 probability of taking this path. And once I get to this node, I have a probability of one of going straight down, because I have no other choices. And every single point, I have a discrete distribution, like we had on page one, which tells me where I'm going. And because my tree is geometrically motivated, these choices, which were all discrete and easy to compute, and easy to count, and easy to keep track of, now I can compute the KL divergence, because it's trivial. It's all... it's just counting, taking logs, and summing things up. So this is my prior distribution. As I said, we're going to use some Bayesian statistics. And so here is my prior. How... this is built with my training set. How do I include my test set? So when I see a piece of data coming over the wire, I can make an inference of where it should belong in this tree. So if it was already... if it was in my training set and didn't get involved in building the tree, where would have it ended up? So I have a point here. So this is my new point that came over the wire. And I queried and said, oh well, this point is covered by this parent, and it's covered by that parent, and that parent, and all the way up. So the path for this point goes up this way. And now for each of those elements in those paths, I can increment my probabilistic distribution and get a posterior one that involves... that includes the new information that I got, that there's a new point over here. And now I can do that again with another point. Oh yep, cool. I have a posterior... this point's been added, and you can see that the distribution changes. And then I can add some more points, and I can add some more points. So these are all my posterior. And I can update my original thing by taking the path of each of the points and updating all the little discrete distributions that I have. So this enables me to sort of quickly make a posterior distribution out of my prior distribution, and everything is discrete. And then I can take the KL divergence between my posterior discrete distributions and my prior discrete distributions. So in this case, this one has a probability of 0.33, and here I have a probability of 0.66. And afterwards, after I've included all my training set, this is a probability of like 0.4. And this, you know, this is a probability of 0.6, and this is a probability of 0.4. And those are very different. And I can plug those into the KL divergence thing and get a positive number for the KL divergence at this node. And then I can do it for this node, and this node, and this node, and get all the KL divergences and add them up. And then I can get the total KL divergence with respect to my original tree of the posterior distribution with respect to the prior distribution, which is really useful. And so this basically, yeah, does it. And the code for doing this enables you to do it very quickly, because it's basically just counting a couple logs and some other statistical equations. And the slide just covers how you do it. So let's go back to that original test set. So I've got basically a blown up version of those slides of the Gaussian distributions where I had them, where a large number of points from my training distribution, which was a single Gaussian sampled at the center of the thing. And then I have a small number of points from a test distribution. And in this case, I have 100,000 points from my training distribution and only 500 points from my test distribution. And these both have a covariance matrix of all ones. So it's actually quite, and it's 20-dimensional. This is actually harder than most datasets. MNIST is about 10-dimensional, if you really get down to it. If you want to know what that is, you can contact me on Twitter and ask me why MNIST might be lower dimensional. And between the training distribution, I build the tree, fix it, and then I make a posterior distribution for a bunch of test distributions. And I take the KL divergence. And you can see it's a bit noisy. The reason being is I generate a lot of the exact position of your points. It's fairly sensitive to these. So this is one of the problems. And one of the things I'll talk about at the end is like, there are ways to normalize this and resolve this, because I've designed it for speed, not really for accuracy. And I'll get into why I did that afterwards, because honestly, it came to the drift stuff backwards. But you can see, the further you get, the further you separate the two distributions, the higher the number gets. So the sanity test, the unit test of does this work is correct. And I'll have to compare this to other methods. This takes big O of, sorry, Wasserstein takes big O of N to the fourth. Wasserstein is the only other drift calculator that I really know works. And is model independent and doesn't have some funny business with building an estimated distribution initially. This takes K log N. Another big difference between Wasserstein is the traditional way of doing Wasserstein requires a relatively equal size test and training set. You can't have the test set be 0.5% of the training set. That would just not work so great on the traditional method I know of. And this one, honestly, it's online. You can do, the test set can be a fraction of the size of the training set. And because it's online, it can do it in real time. And this is stupidly fast. You can track for the Ember open source, Ember malware data set, you build a reasonable couple tree, you can track 16,000 new samples per second on my laptop. So that's fast enough that inference wouldn't be, this would not be the bottleneck. Inference would be the bottleneck if you would build a cloud distribution, a cloud model, and put this as a guard in front of that cloud model. So we built it so that it can go fast enough to defend. And now here is where the thing originally came from, and then we backtracked onto Drift. So here's the test set attack. So this was originally categorized by Ian Goodfellow, and some defenses have been proposed by Nicholas Carlini and others. So, but here is the gist of the attack. Normal users just query. They don't care about whether they got a false positive or a false negative, they just keep querying. And for us, we think of a normal user as being like an enterprise user, trying to figure out whether all the reports coming over the wire of new binaries are malicious or benign, or they'd be submitting every single binary that they see on their network to a service that will tell them whether it was malicious or benign for an ADV product. But a malicious user would be submitting things with a specific goal in mind. They don't care about what they care deeply what the label is, and they care deeply about defining a false negative. They want to find a malicious sample we classify as benign. So what they'll do is they'll try a bunch of stuff until they find something that is misclassified as benign and is actually malicious and bypasses our model, and then they deploy it as far as much as possible. So if you do, if you apply this to, you know, the fast drift calculators to that, you can attach a little drift calculator to each user. And as samples come over the wire, you track it for that user. And if the user starts exploiting something, they're going to get a very boring distribution where it's just one point repeated over and over again. And you spike your distribution, your KL divergence spikes up and then becomes very normal because it's essentially you're tracing the same path over and over and over again. And you can see that there's an exploration phase where the malicious users try to look for a benign misclassified sample, and when they find one, there's an exploitation phase. And the KL divergence rises rapidly. You can see that that this is in log of scale. So this is literally hundreds of times more KL divergence than the benign samples. And here, it's the same thing. Hundreds of times more KL divergence than the benign and the attackers were unsuccessful. Because of this, you can build, you can probably build a very good online defense of your system against a set attack. And here's where basically the system was originally built sort of with this mind and other attacks in mind, because I'm more interested in defending my models than calculating drift. But because it's so bloody fast, it does this really well. And because it's so bloody fast, you can retrofit it. So here's where I'll talk about some next steps. So as I mentioned previously, this thing is a bit noisy. There are ways to fix it. In this case, window sizes and some tuning things that I didn't go into in these slides because I've already got 20 minutes. There are some tuning systems you can use to really clean up and de-noise. There's also some interesting waking functions you can do inside the tree because there's some noise from an overactive leaf node might end up just building a very high KL divergence for just that one node and it throws everything off. And there's a couple other little things like that that still need to be cleaned up before this goes into production. But the system itself is looking very promising and is honestly the fastest of its kind that I can see in the world. And that's basically it for Gokul's ability to detect benign test set and other drift things. The library is at elastic.com, oh sorry, github slash elastic slash Gokul. It's named after my grandmother because the base algorithm for this thing is called GRMA. And if you say that while drunk, you can kind of get grandma. And grandma's a bit too on the nose, so I named it Gokul. Anyway, well, looking forward to hearing you guys on Twitter and Twitch and Slack or whatever. So thank you very much and good night."
}