whisper.cpp logs : 32 threads on 240 thread cloud cpu

[trholding]

time ./main -m models/ggml-base.en.bin -f elon.wav -t 32
whisper_model_load: loading model from 'models/ggml-base.en.bin'
whisper_model_load: n_vocab       = 51864
whisper_model_load: n_audio_ctx   = 1500
whisper_model_load: n_audio_state = 512
whisper_model_load: n_audio_head  = 8
whisper_model_load: n_audio_layer = 6
whisper_model_load: n_text_ctx    = 448
whisper_model_load: n_text_state  = 512
whisper_model_load: n_text_head   = 8
whisper_model_load: n_text_layer  = 6
whisper_model_load: n_mels        = 80
whisper_model_load: f16           = 1
whisper_model_load: type          = 2
whisper_model_load: mem_required  = 670.00 MB
whisper_model_load: adding 1607 extra tokens
whisper_model_load: ggml ctx size = 140.60 MB
whisper_model_load: memory size =    22.83 MB
whisper_model_load: model size  =   140.54 MB

system_info: n_threads = 32 / 240 | AVX2 = 1 | AVX512 = 0 | NEON = 0 | FP16_VA = 0 | WASM_SIMD = 0 | BLAS = 0 | 

main: processing 'elon.wav' (34466688 samples, 2154.2 sec), 32 threads, 1 processors, lang = en, task = transcribe, timestamps = 1 ...


[00:00:00.000 --> 00:00:03.960]   [MUSIC PLAYING]
[00:00:03.960 --> 00:00:18.240]   In life, we've seen within this part of the world
[00:00:18.240 --> 00:00:26.520]   great people like Al-Khalwarizmi, who invented Al-Garedan,
[00:00:26.520 --> 00:00:33.600]   globally Newton, Henry Ford, the Wright brother,
[00:00:33.600 --> 00:00:37.720]   Albert Einstein, and Elon Musk.
[00:00:37.720 --> 00:00:49.520]   See, you are in rush.
[00:00:49.520 --> 00:00:54.400]   You want to go to places that nobody been.
[00:00:54.400 --> 00:01:00.520]   You are reinventing a certain industry
[00:01:00.520 --> 00:01:04.480]   from the rocket industry with SpaceX to the car industry
[00:01:04.480 --> 00:01:06.440]   with Tesla.
[00:01:06.440 --> 00:01:07.680]   What's your mission life?
[00:01:07.680 --> 00:01:11.760]   Why you do whatever you do?
[00:01:11.760 --> 00:01:14.760]   Sure, well, first of all, thank you for having me.
[00:01:14.760 --> 00:01:16.920]   It's an honor to be here.
[00:01:16.920 --> 00:01:21.040]   And we're having a really, really great time with my kids
[00:01:21.040 --> 00:01:22.000]   in the device.
[00:01:22.000 --> 00:01:24.640]   We've really been fantastic at it, really encourage anyone
[00:01:24.640 --> 00:01:26.560]   who hasn't been to visit.
[00:01:26.560 --> 00:01:28.080]   What a great city.
[00:01:28.080 --> 00:01:28.580]   Thank you.
[00:01:28.580 --> 00:01:30.400]   [APPLAUSE]
[00:01:30.400 --> 00:01:36.880]   And in terms of the motivations, I
[00:01:36.880 --> 00:01:42.040]   guess the sort of a long version of the explanation.
[00:01:42.040 --> 00:01:48.200]   But essentially, when I was a kid,
[00:01:48.200 --> 00:01:51.520]   I was wondering what's the meaning of life?
[00:01:51.520 --> 00:01:52.880]   Like, why are we here?
[00:01:52.880 --> 00:01:54.480]   What's it all about?
[00:01:54.480 --> 00:02:00.640]   And I came to the conclusion that what really matters
[00:02:00.640 --> 00:02:04.320]   is trying to understand the right questions to ask.
[00:02:04.320 --> 00:02:08.680]   And the more that we can increase the scope and scale
[00:02:08.680 --> 00:02:11.760]   of human consciousness, the better
[00:02:11.760 --> 00:02:14.560]   we are able to ask these questions.
[00:02:14.560 --> 00:02:17.640]   And so I think that there are certain things that are
[00:02:17.640 --> 00:02:22.240]   necessary to ensure that the future is good.
[00:02:22.240 --> 00:02:28.000]   And some of those things are in the long term,
[00:02:28.000 --> 00:02:30.680]   having long term sustainable transport
[00:02:30.680 --> 00:02:34.200]   and sustainable energy generation,
[00:02:34.200 --> 00:02:39.480]   and to be a space-braying civilization
[00:02:39.480 --> 00:02:42.200]   and for humanity to be out there among the stars
[00:02:42.200 --> 00:02:46.960]   and be a multi-planetary species.
[00:02:46.960 --> 00:02:50.280]   I mean, I think being a multi-planetary species
[00:02:50.280 --> 00:02:51.520]   and being out there among the stars
[00:02:51.520 --> 00:02:57.040]   is important for the long term survival of humanity.
[00:02:57.040 --> 00:03:01.080]   And that's one reason, kind of like life insurance for life
[00:03:01.080 --> 00:03:02.000]   collectively.
[00:03:02.000 --> 00:03:03.840]   Life as we know it.
[00:03:03.840 --> 00:03:08.040]   But then the part that I find personally most motivating
[00:03:08.040 --> 00:03:12.000]   is that it creates a sense of adventure.
[00:03:12.000 --> 00:03:15.560]   And it makes people excited about the future.
[00:03:15.560 --> 00:03:19.080]   And if you consider two futures, one where we are forever
[00:03:19.080 --> 00:03:22.520]   confined to Earth until eventually something terrible
[00:03:22.520 --> 00:03:26.400]   happens, or another future where we are out there
[00:03:26.400 --> 00:03:28.400]   on many planets, maybe even going
[00:03:28.400 --> 00:03:32.040]   beyond the solar system, I think that second version is
[00:03:32.040 --> 00:03:34.720]   incredibly exciting and inspiring.
[00:03:34.720 --> 00:03:38.080]   And there need to be reasons to get up in the morning.
[00:03:38.080 --> 00:03:40.320]   Life can't just be about solving problems.
[00:03:40.320 --> 00:03:42.040]   Otherwise, what's the point?
[00:03:42.040 --> 00:03:45.000]   There's got to be things that people find inspiring
[00:03:45.000 --> 00:03:48.200]   and make life worth living.
[00:03:48.200 --> 00:03:50.720]   So what is life for you?
[00:03:50.720 --> 00:03:55.080]   I mean, you look at our life, and I
[00:03:55.080 --> 00:03:57.240]   heard you before speaking.
[00:03:57.240 --> 00:03:58.920]   Is it a dream?
[00:03:58.920 --> 00:04:00.640]   Is it a dream?
[00:04:00.640 --> 00:04:03.120]   Is it a million D?
[00:04:03.120 --> 00:04:04.640]   What is life for Elon Musk?
[00:04:04.640 --> 00:04:10.480]   I find as I get older, I find that question
[00:04:10.480 --> 00:04:17.400]   to be maybe more and more confusing or troubling or uncertain.
[00:04:17.400 --> 00:04:22.680]   I think particularly when you see the advancement
[00:04:22.680 --> 00:04:27.320]   of something like video games, say 40 years ago,
[00:04:27.320 --> 00:04:29.280]   you had video games, the most advanced video game
[00:04:29.280 --> 00:04:35.200]   would be like Pong, where you had two rectangles and a dot,
[00:04:35.200 --> 00:04:36.760]   and you're like batting it back and forth.
[00:04:36.760 --> 00:04:37.480]   I played it.
[00:04:37.480 --> 00:04:38.680]   Oh yeah, me too, exactly.
[00:04:38.680 --> 00:04:39.680]   That's all.
[00:04:39.680 --> 00:04:42.320]   It sort of dates you a little bit.
[00:04:42.320 --> 00:04:45.200]   Yeah, we both played the same game.
[00:04:45.200 --> 00:04:48.920]   And that was like, wow, that was a pretty fun game at the time.
[00:04:48.920 --> 00:04:52.480]   But now you can see a video game that's photorealistic,
[00:04:52.480 --> 00:04:54.600]   almost photorealistic, and millions of people
[00:04:54.600 --> 00:04:56.360]   playing simultaneously.
[00:04:56.360 --> 00:05:02.160]   And you see where things going with virtual reality
[00:05:02.160 --> 00:05:03.560]   and augmented reality.
[00:05:03.560 --> 00:05:07.200]   And if you extrapolate that out into the future
[00:05:07.200 --> 00:05:09.480]   with any rate of progress at all,
[00:05:09.480 --> 00:05:15.180]   like even 0.1% or something like that a year,
[00:05:15.180 --> 00:05:17.580]   then eventually those games will be indistinguishable
[00:05:17.580 --> 00:05:19.240]   from reality.
[00:05:19.240 --> 00:05:20.700]   If we're so realistic, you will not
[00:05:20.700 --> 00:05:23.540]   be able to tell the difference between that game
[00:05:23.540 --> 00:05:26.800]   and the reality as we know it.
[00:05:26.800 --> 00:05:31.320]   And then it seems like, well, how do we
[00:05:31.320 --> 00:05:33.680]   know that that didn't happen in the past?
[00:05:33.680 --> 00:05:36.760]   And that we're not in one of those games ourselves?
[00:05:36.760 --> 00:05:37.800]   Interesting.
[00:05:37.800 --> 00:05:39.640]   Interesting.
[00:05:39.640 --> 00:05:41.960]   I mean, could be.
[00:05:41.960 --> 00:05:44.000]   Everything is possible in life.
[00:05:44.000 --> 00:05:46.160]   I mean, there's--
[00:05:46.160 --> 00:05:47.720]   I mean, yeah, particularly things
[00:05:47.720 --> 00:05:51.400]   seem to be accelerating to something.
[00:05:51.400 --> 00:05:52.080]   Isn't it?
[00:05:52.080 --> 00:05:57.560]   I mean, if we look at our life, a seam in the past 100 years,
[00:05:57.560 --> 00:05:59.240]   life been exubating quite fast.
[00:05:59.240 --> 00:05:59.720]   Yeah.
[00:05:59.720 --> 00:06:01.600]   In the past 20, more faster.
[00:06:01.600 --> 00:06:03.200]   It's getting faster and faster.
[00:06:03.200 --> 00:06:04.240]   Is it more slow?
[00:06:04.240 --> 00:06:07.240]   So my question is really, how life
[00:06:07.240 --> 00:06:11.720]   will be in air 20, 30, 50 years from our education,
[00:06:11.720 --> 00:06:12.680]   our transport?
[00:06:12.680 --> 00:06:14.320]   How do you see it?
[00:06:14.320 --> 00:06:15.960]   Well, I think this is one of those things
[00:06:15.960 --> 00:06:19.920]   that's quite difficult to predict.
[00:06:19.920 --> 00:06:24.280]   When you think of, say, I mean, the first controlled powered
[00:06:24.280 --> 00:06:28.840]   flight was 1903 with the Wright Brothers.
[00:06:28.840 --> 00:06:34.040]   And then 66 years later, we put the first people on the moon.
[00:06:34.040 --> 00:06:39.920]   I mean, if you'd asked people, say, in 1900,
[00:06:39.920 --> 00:06:42.800]   what are the odds of man landing on the moon,
[00:06:42.800 --> 00:06:46.200]   they would have said that's ridiculous.
[00:06:46.200 --> 00:06:48.600]   And if you try to talk to them about the internet,
[00:06:48.600 --> 00:06:50.400]   they would not even know what the heck you even--
[00:06:50.400 --> 00:06:52.000]   What are you even talking about?
[00:06:52.000 --> 00:06:54.800]   This sounds so crazy.
[00:06:54.800 --> 00:06:59.600]   But today, with a $100 device, you
[00:06:59.600 --> 00:07:03.920]   can video conference with anyone in the world,
[00:07:03.920 --> 00:07:05.440]   on the other side of the world.
[00:07:05.440 --> 00:07:08.760]   And if you have a Wi-Fi connection,
[00:07:08.760 --> 00:07:11.120]   it's basically free.
[00:07:11.120 --> 00:07:15.240]   Free to have an instant visual communication with anyone,
[00:07:15.240 --> 00:07:17.320]   or even with millions of people.
[00:07:17.320 --> 00:07:19.360]   With social media, you can communicate to millions of people
[00:07:19.360 --> 00:07:20.280]   simultaneously.
[00:07:20.280 --> 00:07:26.560]   And you can Google something and ask any question.
[00:07:26.560 --> 00:07:29.360]   It's like an article of wisdom that you can ask almost
[00:07:29.360 --> 00:07:32.800]   any question, again, instant response.
[00:07:32.800 --> 00:07:35.000]   It would have been incredibly difficult to predict these things
[00:07:35.000 --> 00:07:39.120]   in the past, even the relatively recent past.
[00:07:39.120 --> 00:07:44.600]   So I think the one thing that we can be quite certain of is
[00:07:44.600 --> 00:07:47.920]   that any predictions we make today for the future
[00:07:47.920 --> 00:07:51.600]   will be like in 50 years will be wrong.
[00:07:51.600 --> 00:07:53.000]   That's for sure.
[00:07:53.000 --> 00:07:57.000]   I mean, I think directionally, I can tell you what I hope
[00:07:57.000 --> 00:08:00.160]   the future has, as opposed to maybe what it will be.
[00:08:00.160 --> 00:08:02.000]   Because this may just be wishful thinking.
[00:08:02.000 --> 00:08:10.400]   I mean, I hope we are out there on Mars, and maybe beyond Mars,
[00:08:10.400 --> 00:08:12.800]   the moons are Jupiter.
[00:08:12.800 --> 00:08:16.800]   I hope we're traveling frequently
[00:08:16.800 --> 00:08:20.840]   throughout the solar system, perhaps preparing for missions
[00:08:20.840 --> 00:08:24.320]   to nearby star systems.
[00:08:24.320 --> 00:08:28.800]   I think all of this is possible within 50 years.
[00:08:28.800 --> 00:08:33.440]   And I think that would be very exciting to do that.
[00:08:33.440 --> 00:08:39.920]   And I think we'll see autonomy and artificial intelligence
[00:08:39.920 --> 00:08:41.920]   advanced tremendously.
[00:08:41.920 --> 00:08:44.520]   That's actually quite near term.
[00:08:44.520 --> 00:08:49.200]   My guess is in probably 10 years, it
[00:08:49.200 --> 00:08:52.640]   will be very unusual for cars to be
[00:08:52.640 --> 00:08:55.280]   built that are not fully autonomous.
[00:08:55.280 --> 00:08:55.880]   Ten years.
[00:08:55.880 --> 00:08:56.800]   Ten years from now.
[00:08:56.800 --> 00:08:57.800]   Yeah.
[00:08:57.800 --> 00:09:02.120]   I think almost all cars built will be
[00:09:02.120 --> 00:09:06.160]   capable of full autonomy in about 10 years.
[00:09:06.160 --> 00:09:08.960]   As it is, the Tesla cars that are made today
[00:09:08.960 --> 00:09:12.520]   have the sensor system necessary for full autonomy.
[00:09:12.520 --> 00:09:15.480]   And we think probably enough compute power
[00:09:15.480 --> 00:09:17.320]   to be safer than a person.
[00:09:17.320 --> 00:09:19.760]   So it's mostly just a question of developing the software
[00:09:19.760 --> 00:09:21.760]   and uploading the software.
[00:09:21.760 --> 00:09:26.040]   And if it turns out that the more compute power is needed,
[00:09:26.040 --> 00:09:28.800]   we can easily upgrade the computer.
[00:09:28.800 --> 00:09:33.640]   And so that's all Tesla's built since October of last year.
[00:09:33.640 --> 00:09:36.920]   And other manufacturers will follow and do the same thing.
[00:09:36.920 --> 00:09:42.160]   So getting in a car will be like getting in an elevator.
[00:09:42.160 --> 00:09:43.920]   You just tell it where you want to go.
[00:09:43.920 --> 00:09:48.240]   And it takes you there with extreme levels of safety.
[00:09:48.240 --> 00:09:50.240]   And that'll be normal.
[00:09:50.240 --> 00:09:51.560]   It'll just be normal.
[00:09:51.560 --> 00:09:54.240]   For elevators, there used to be elevator operators.
[00:09:54.240 --> 00:09:57.120]   You get in, there'd be a guy moving a lever.
[00:09:57.120 --> 00:09:59.240]   Now you just get in, you press the button,
[00:09:59.240 --> 00:10:02.640]   and that's taken for granted.
[00:10:02.640 --> 00:10:04.040]   So autonomy will be widespread.
[00:10:04.040 --> 00:10:10.160]   I think one of the most troubling questions
[00:10:10.160 --> 00:10:12.320]   is artificial intelligence.
[00:10:12.320 --> 00:10:15.760]   And I don't mean narrow AI.
[00:10:15.760 --> 00:10:21.480]   Like a vehicle autonomy I would put in the narrow AI class.
[00:10:21.480 --> 00:10:26.240]   It's narrowly trying to achieve a certain function.
[00:10:26.240 --> 00:10:28.120]   But deep artificial intelligence,
[00:10:28.120 --> 00:10:32.160]   or what is sometimes called artificial general intelligence,
[00:10:32.160 --> 00:10:37.240]   where you could have AI that is much smarter than the smartest
[00:10:37.240 --> 00:10:39.000]   human on Earth.
[00:10:39.000 --> 00:10:45.280]   This, I think, is a dangerous situation.
[00:10:45.280 --> 00:10:46.640]   Why it is dangerous?
[00:10:46.640 --> 00:10:47.680]   I mean, there is two of you.
[00:10:47.680 --> 00:10:51.960]   One view is artificial intelligence will have humanity.
[00:10:51.960 --> 00:10:53.720]   There is another school of think of thought
[00:10:53.720 --> 00:10:57.880]   as artificial intelligence as a threat to humanity.
[00:10:57.880 --> 00:10:59.480]   Why is it?
[00:10:59.480 --> 00:11:01.000]   Well, I think it's both.
[00:11:01.000 --> 00:11:04.440]   You know, it's like--
[00:11:04.440 --> 00:11:07.320]   one way to think of it is imagine we're
[00:11:07.320 --> 00:11:09.520]   going to be visited--
[00:11:09.520 --> 00:11:11.480]   imagine you're very confident that we're
[00:11:11.480 --> 00:11:16.080]   going to be visited by super intelligent aliens
[00:11:16.080 --> 00:11:19.400]   in, let's say, 10 years or 20 years at the most.
[00:11:19.400 --> 00:11:20.200]   Super intelligent.
[00:11:20.200 --> 00:11:25.960]   So you think within 20 years you have alien and Earth?
[00:11:25.960 --> 00:11:30.120]   Well, digital super intelligence will be like an alien.
[00:11:30.120 --> 00:11:31.520]   It will be like an alien.
[00:11:31.520 --> 00:11:36.560]   But my question is, do you think there
[00:11:36.560 --> 00:11:41.240]   is other intelligent life outside the Earth?
[00:11:41.240 --> 00:11:43.120]   It seems probable.
[00:11:43.120 --> 00:11:45.600]   But I think this is one of the great questions
[00:11:45.600 --> 00:11:52.120]   in physics and philosophy is, where are the aliens?
[00:11:52.120 --> 00:11:53.000]   Maybe they're among us.
[00:11:53.000 --> 00:11:55.400]   I don't know.
[00:11:55.400 --> 00:11:58.600]   Some people think I'm an alien.
[00:11:58.600 --> 00:12:00.520]   Not true.
[00:12:00.520 --> 00:12:01.880]   Not true.
[00:12:01.880 --> 00:12:04.360]   But maybe we are aliens.
[00:12:04.360 --> 00:12:05.360]   Of course, I'd say that would not.
[00:12:05.360 --> 00:12:06.120]   Maybe we are aliens.
[00:12:06.120 --> 00:12:12.880]   And then, I mean, if you look at this part of the world,
[00:12:12.880 --> 00:12:16.680]   they believe that human beings are not from Earth.
[00:12:16.680 --> 00:12:19.120]   They came from somewhere else.
[00:12:19.120 --> 00:12:22.520]   Eve and Adam came from somewhere else to Earth.
[00:12:22.520 --> 00:12:29.440]   So in a way, human beings alien to this land.
[00:12:29.440 --> 00:12:31.960]   Do you think we'll make contact with aliens
[00:12:31.960 --> 00:12:35.680]   within the next 50 years?
[00:12:35.680 --> 00:12:37.280]   Well, that's a really tough one to say.
[00:12:40.160 --> 00:12:44.960]   If there are super intelligent aliens out there,
[00:12:44.960 --> 00:12:47.760]   they're probably already observing us.
[00:12:47.760 --> 00:12:49.280]   That would seem quite likely.
[00:12:49.280 --> 00:12:54.320]   And we just are not spot enough to realize it.
[00:12:54.320 --> 00:12:58.320]   But I can do some back-the-overlook calculations.
[00:12:58.320 --> 00:13:04.440]   And any advanced alien civilization
[00:13:04.440 --> 00:13:10.440]   that was at all interested in populating the galaxy,
[00:13:10.440 --> 00:13:14.960]   even without exceeding the speed of light,
[00:13:14.960 --> 00:13:18.480]   even if you're only moving at, say, 10% or 20%
[00:13:18.480 --> 00:13:23.840]   of the speed of light, you could populate the entire galaxy
[00:13:23.840 --> 00:13:27.120]   in, let's say, 10 million years.
[00:13:27.120 --> 00:13:29.560]   Maybe 20 million years max.
[00:13:29.560 --> 00:13:34.240]   This is nothing in the grand scheme of things.
[00:13:34.240 --> 00:13:36.320]   Once you said, do you want to die in Mars?
[00:13:36.320 --> 00:13:37.480]   Why?
[00:13:37.480 --> 00:13:40.480]   I don't-- declare, I don't want to die on Mars.
[00:13:40.480 --> 00:13:42.640]   [LAUGHTER]
[00:13:42.640 --> 00:13:46.960]   It's like, if-- I mean, we're all going to die someday.
[00:13:46.960 --> 00:13:50.360]   And if you're going to pick someplace to die,
[00:13:50.360 --> 00:13:52.240]   then why not Mars?
[00:13:52.240 --> 00:13:57.200]   You know, if we're born on Earth, why not die on Mars?
[00:13:57.200 --> 00:14:00.200]   Seems like maybe it'd be quite exciting.
[00:14:00.200 --> 00:14:03.480]   So I think given the choice of dying on Earth,
[00:14:03.480 --> 00:14:07.720]   so dying on Mars, I'd say, yeah, sure, I'll die on Mars.
[00:14:07.720 --> 00:14:11.160]   But it's not some kind of Mars death wish.
[00:14:11.160 --> 00:14:13.280]   And if I do die on Mars, I just don't
[00:14:13.280 --> 00:14:14.480]   want it to be on impact.
[00:14:14.480 --> 00:14:15.480]   OK.
[00:14:15.480 --> 00:14:16.280]   Got it.
[00:14:16.280 --> 00:14:17.280]   Here.
[00:14:17.280 --> 00:14:20.720]   Let's come back to Earth, actually.
[00:14:20.720 --> 00:14:27.480]   You tweeted that you are building a tunnel under Washington,
[00:14:27.480 --> 00:14:29.200]   DC.
[00:14:29.200 --> 00:14:30.280]   Why?
[00:14:30.280 --> 00:14:31.560]   What is it?
[00:14:31.560 --> 00:14:34.520]   Well, it's a secret plot.
[00:14:34.520 --> 00:14:35.760]   Just between us.
[00:14:35.760 --> 00:14:37.000]   Nobody had new?
[00:14:37.000 --> 00:14:38.240]   Yeah, exactly.
[00:14:38.240 --> 00:14:41.520]   Please keep that secret.
[00:14:41.520 --> 00:14:45.600]   Well, I think this is going to sound a little--
[00:14:45.600 --> 00:14:51.000]   I mean, it seems like somewhat trivial or silly.
[00:14:51.000 --> 00:14:53.600]   But I've been saying this for many years now.
[00:14:53.600 --> 00:14:57.600]   But I think that the solution to urban congestion
[00:14:57.600 --> 00:15:01.720]   is a network of tunnels under cities.
[00:15:01.720 --> 00:15:05.720]   And when I say, I don't mean a 2D plane of tunnels.
[00:15:05.720 --> 00:15:09.600]   I mean tunnels that go many levels deep.
[00:15:09.600 --> 00:15:14.720]   So you can always go deeper than you can go up.
[00:15:14.720 --> 00:15:18.960]   Like the deepest mines are taller than the tallest buildings.
[00:15:18.960 --> 00:15:22.480]   So you could have a network of tunnels
[00:15:22.480 --> 00:15:26.200]   that are 20, 30, 40, 50 levels.
[00:15:26.200 --> 00:15:28.840]   As many levels as you want, really.
[00:15:28.840 --> 00:15:33.600]   And so given that, you can overcome the congestion situation
[00:15:33.600 --> 00:15:37.680]   in any city in the world.
[00:15:37.680 --> 00:15:39.200]   The challenge is just figuring out
[00:15:39.200 --> 00:15:44.360]   how do you build tunnels quickly and at low cost
[00:15:44.360 --> 00:15:47.840]   and with high safety.
[00:15:47.840 --> 00:15:50.560]   So if tunneling technology can be improved to the point
[00:15:50.560 --> 00:15:55.400]   where you can build tunnels fast, cheap, and safe,
[00:15:55.400 --> 00:15:58.000]   then that would completely get rid
[00:15:58.000 --> 00:16:03.400]   of any traffic situations in cities.
[00:16:03.400 --> 00:16:07.000]   And so I think that's why I think it's an important technology.
[00:16:07.000 --> 00:16:11.920]   And Washington DC, LA, and most of the major American cities,
[00:16:11.920 --> 00:16:15.320]   most major cities in the world suffer from severe traffic
[00:16:15.320 --> 00:16:16.160]   issues.
[00:16:16.160 --> 00:16:17.960]   And it's mostly because you've got these buildings
[00:16:17.960 --> 00:16:20.120]   which are-- you have tall buildings that are 3D.
[00:16:20.120 --> 00:16:25.280]   And you have a road network that is one level.
[00:16:25.280 --> 00:16:27.080]   And then people generally want to go in and out
[00:16:27.080 --> 00:16:29.560]   of those buildings at the exact same time.
[00:16:29.560 --> 00:16:33.320]   So then, do you get the traffic jam?
[00:16:33.320 --> 00:16:39.880]   Let's come back to UAE and Dubai.
[00:16:39.880 --> 00:16:44.640]   The first time I met you, it was 4th of June 2015
[00:16:44.640 --> 00:16:47.880]   at your office in SpaceX.
[00:16:47.880 --> 00:16:53.040]   And I asked you, would you have a presence in UAE?
[00:16:53.040 --> 00:16:55.840]   A new answer was, I'm busy with China.
[00:16:55.840 --> 00:16:57.800]   Maybe not in the near future.
[00:16:57.800 --> 00:17:00.680]   And almost a year and a half later, we are here.
[00:17:00.680 --> 00:17:02.720]   Same time goes quite fast.
[00:17:02.720 --> 00:17:05.400]   Why now?
[00:17:05.400 --> 00:17:09.200]   Well, I think actually things are going reasonably well in China.
[00:17:09.200 --> 00:17:12.640]   So we had some initial challenges
[00:17:12.640 --> 00:17:16.360]   figuring out charging and service infrastructure
[00:17:16.360 --> 00:17:17.440]   and various other things.
[00:17:17.440 --> 00:17:19.760]   But now it's actually going fairly well.
[00:17:19.760 --> 00:17:26.120]   And so the timing seemed to be good to really make
[00:17:26.120 --> 00:17:30.520]   a significant debut in this region starting in Dubai.
[00:17:30.520 --> 00:17:34.400]   A new opinion.
[00:17:34.400 --> 00:17:40.120]   What is the new disturbing thing that will come to technology
[00:17:40.120 --> 00:17:40.480]   in X?
[00:17:40.480 --> 00:17:41.840]   What's X in technology?
[00:17:41.840 --> 00:17:45.600]   What's X in technology?
[00:17:45.600 --> 00:17:49.280]   That will disturb the way we live, the way we think,
[00:17:49.280 --> 00:17:51.320]   the way we do business.
[00:17:51.320 --> 00:17:56.760]   Well, the most near term impact from a technology standpoint
[00:17:56.760 --> 00:18:01.480]   is autonomous cars, like fully self-driving cars.
[00:18:01.480 --> 00:18:04.400]   Like I said, that's going to happen much faster
[00:18:04.400 --> 00:18:05.640]   than people realize.
[00:18:05.640 --> 00:18:12.440]   It's going to be a great convenience to be an autonomous
[00:18:12.440 --> 00:18:17.840]   car, but there are many people whose jobs it is to drive.
[00:18:17.840 --> 00:18:23.240]   So in fact, I think it might be the single largest
[00:18:23.240 --> 00:18:30.320]   employer of people driving in various forms.
[00:18:30.320 --> 00:18:34.880]   And so then we need to figure out new roles for what
[00:18:34.880 --> 00:18:37.760]   do those people do.
[00:18:37.760 --> 00:18:40.880]   But it will be very disruptive and very quick.
[00:18:40.880 --> 00:18:43.800]   Now, I should characterize what I mean by quick.
[00:18:47.280 --> 00:18:51.040]   There are quick means, different things to different people.
[00:18:51.040 --> 00:18:54.800]   There are over 2 billion vehicles in the world,
[00:18:54.800 --> 00:18:58.480]   approaching, in fact, approaching 2 and 1/2 billion cars
[00:18:58.480 --> 00:19:01.360]   and trucks in the world.
[00:19:01.360 --> 00:19:04.400]   The total new vehicle production capacity
[00:19:04.400 --> 00:19:08.760]   is about 100 million, which makes sense
[00:19:08.760 --> 00:19:11.000]   because the life of a car or truck before it's finally
[00:19:11.000 --> 00:19:13.640]   scrapped is about 20, 25 years.
[00:19:13.640 --> 00:19:21.120]   So the point at which we see full autonomy up here
[00:19:21.120 --> 00:19:24.280]   will not be the point at which there is massive societal upheaval
[00:19:24.280 --> 00:19:28.040]   because it will take a long time to make enough autonomous
[00:19:28.040 --> 00:19:34.120]   vehicles to disrupt employment.
[00:19:34.120 --> 00:19:36.520]   So that disruption I'm talking about
[00:19:36.520 --> 00:19:40.360]   will take place over about 20 years.
[00:19:40.360 --> 00:19:45.680]   But still 20 years is a short period of time to have,
[00:19:45.680 --> 00:19:51.640]   I think, something like 12% to 15% of the workforce be unemployed.
[00:19:51.640 --> 00:19:54.080]   Thank you.
[00:19:54.080 --> 00:19:57.760]   This is the largest global government summit.
[00:19:57.760 --> 00:20:01.920]   We have over 139 government here.
[00:20:01.920 --> 00:20:06.840]   If you want to advise government official to be ready for the future,
[00:20:06.840 --> 00:20:11.000]   what is three things or three advice, you'll give them?
[00:20:11.000 --> 00:20:18.240]   Well, I think the first bit of advice would be to really pay close
[00:20:18.240 --> 00:20:22.600]   attention to the development of artificial intelligence.
[00:20:22.600 --> 00:20:29.760]   I think this is-- we need to just be very careful in how
[00:20:29.760 --> 00:20:32.880]   we adopt artificial intelligence and to make sure
[00:20:32.880 --> 00:20:36.440]   that researchers don't get carried away,
[00:20:36.440 --> 00:20:38.840]   because sometimes what happens is a scientist can get so
[00:20:38.840 --> 00:20:42.320]   engrossed in their work that they don't necessarily realize
[00:20:42.320 --> 00:20:46.520]   the ramifications of what they're doing.
[00:20:46.520 --> 00:20:50.520]   So I think it's important for public safety that we--
[00:20:50.520 --> 00:20:53.440]   governments keep a close eye on artificial intelligence
[00:20:53.440 --> 00:20:56.760]   and make sure that it does not represent a danger to the public.
[00:21:00.720 --> 00:21:10.360]   Secondly, I would say we do need to think about transport in general.
[00:21:10.360 --> 00:21:20.600]   There's the movement towards electric vehicles,
[00:21:20.600 --> 00:21:21.800]   sustainable transport.
[00:21:21.800 --> 00:21:27.520]   I think that's going to be good for many reasons,
[00:21:27.520 --> 00:21:29.840]   but again, not something that happens immediately.
[00:21:29.840 --> 00:21:34.440]   That'll happen slower than the self-driving vehicles.
[00:21:34.440 --> 00:21:38.440]   So that's probably something that happens over 30 or 40 years,
[00:21:38.440 --> 00:21:40.880]   the transition to electric vehicles.
[00:21:40.880 --> 00:21:49.280]   So thinking about that in context,
[00:21:49.280 --> 00:21:54.040]   the demand for electricity will increase dramatically.
[00:21:54.040 --> 00:21:57.560]   So currently, in terms of total energy usage in the world,
[00:21:57.560 --> 00:22:00.720]   it's about one-third electricity, about one-third transport,
[00:22:00.720 --> 00:22:03.880]   about one-third heating.
[00:22:03.880 --> 00:22:09.760]   So over time, that will transition to almost all--
[00:22:09.760 --> 00:22:13.480]   not all, but predominantly electricity,
[00:22:13.480 --> 00:22:17.120]   which means that the demand for electricity will probably
[00:22:17.120 --> 00:22:19.600]   triple.
[00:22:19.600 --> 00:22:21.160]   So it's going to be very important to think about,
[00:22:21.160 --> 00:22:23.160]   how do you make so much more electricity?
[00:22:26.000 --> 00:22:32.320]   And it seems they have an easy job.
[00:22:32.320 --> 00:22:33.800]   That's it.
[00:22:33.800 --> 00:22:36.240]   There's no more challenges for that.
[00:22:36.240 --> 00:22:41.480]   No, well, I think maybe these things
[00:22:41.480 --> 00:22:43.160]   do play into each other a little bit.
[00:22:43.160 --> 00:22:46.360]   But what to do about mass unemployment?
[00:22:46.360 --> 00:22:50.760]   This is going to be a massive social challenge.
[00:22:50.760 --> 00:22:52.880]   And I think ultimately, we will have
[00:22:52.880 --> 00:22:56.160]   to have some kind of universal basic income.
[00:22:56.160 --> 00:22:57.720]   I don't think we're going to have a choice.
[00:22:57.720 --> 00:22:59.440]   Universal basic income.
[00:22:59.440 --> 00:23:02.680]   Universal basic income, I think it's going to be necessary.
[00:23:02.680 --> 00:23:05.000]   So it's being that unemployed people
[00:23:05.000 --> 00:23:08.000]   will be paid across the globe.
[00:23:08.000 --> 00:23:11.440]   Because there is no job, machine robots is taking over.
[00:23:11.440 --> 00:23:16.240]   There will be fewer and fewer jobs that a robot cannot do better.
[00:23:16.240 --> 00:23:21.800]   But I don't want to be clear that these are not things
[00:23:21.800 --> 00:23:23.720]   that I wish would happen.
[00:23:23.720 --> 00:23:26.840]   These are simply things that I think probably will happen.
[00:23:26.840 --> 00:23:34.760]   If my assessment is correct and they probably will happen,
[00:23:34.760 --> 00:23:37.520]   then we need to say, what are we going to do about it?
[00:23:37.520 --> 00:23:40.000]   And I think some kind of universal basic income
[00:23:40.000 --> 00:23:41.240]   is going to be necessary.
[00:23:41.240 --> 00:23:48.320]   The output of goods and services will be extremely high.
[00:23:48.320 --> 00:23:52.320]   So with automation, they will come abundance.
[00:23:52.320 --> 00:23:59.200]   Almost everything will get very cheap.
[00:23:59.200 --> 00:24:09.880]   I think we'll just end up doing universal basic income.
[00:24:09.880 --> 00:24:13.160]   It's going to be necessary.
[00:24:13.160 --> 00:24:18.080]   The much harder challenge is how do people then have meaning?
[00:24:18.080 --> 00:24:22.200]   Like a lot of people, they derive their meaning
[00:24:22.200 --> 00:24:24.400]   from their employment.
[00:24:24.400 --> 00:24:31.800]   So if you're not needed, if there's not a need for your labor,
[00:24:31.800 --> 00:24:34.520]   what's the meaning?
[00:24:34.520 --> 00:24:35.640]   Do you have meaning?
[00:24:35.640 --> 00:24:37.760]   Do you feel useless?
[00:24:37.760 --> 00:24:39.560]   That's a much harder problem to deal with.
[00:24:39.560 --> 00:24:44.200]   And then how do we ensure that the future is going
[00:24:44.200 --> 00:24:46.560]   to be the future that we want, that we still like?
[00:24:46.560 --> 00:24:56.080]   No, I mean, I do think that there's a potential path here,
[00:24:56.080 --> 00:25:00.760]   which is-- and we're really getting into science fiction
[00:25:00.760 --> 00:25:05.600]   or create advanced science stuff.
[00:25:05.600 --> 00:25:11.800]   But having some sort of merger with biological intelligence
[00:25:11.800 --> 00:25:16.640]   and machine intelligence, to some degree,
[00:25:16.640 --> 00:25:18.240]   we are already a cyborg.
[00:25:18.240 --> 00:25:25.800]   You think of the digital tools that you have--
[00:25:25.800 --> 00:25:28.480]   your phone, your computer, the applications
[00:25:28.480 --> 00:25:31.280]   that you have-- the fact that as I was mentioning earlier,
[00:25:31.280 --> 00:25:35.440]   you can ask a question and instantly get an answer
[00:25:35.440 --> 00:25:38.720]   from Google or from other things.
[00:25:38.720 --> 00:25:43.200]   And so you already have a digital tertiary layer.
[00:25:43.200 --> 00:25:45.560]   I say tertiary because you can think of the limbic system,
[00:25:45.560 --> 00:25:48.480]   kind of the animal brain or the primal brain,
[00:25:48.480 --> 00:25:51.000]   and then the cortex, kind of the thinking,
[00:25:51.000 --> 00:25:54.000]   planning part of the brain, and then your digital self
[00:25:54.000 --> 00:25:55.280]   as a third layer.
[00:25:55.280 --> 00:26:00.200]   So you already have that.
[00:26:00.200 --> 00:26:04.240]   And if somebody dies, their digital ghost is still around.
[00:26:04.240 --> 00:26:06.120]   All of their emails and the pictures
[00:26:06.120 --> 00:26:08.280]   that they posted in the social media,
[00:26:08.280 --> 00:26:11.640]   that still lives, even if they died.
[00:26:11.640 --> 00:26:14.120]   So over time, I think we'll probably
[00:26:14.120 --> 00:26:20.920]   see a closer merger of biological intelligence
[00:26:20.920 --> 00:26:22.400]   and digital intelligence.
[00:26:22.400 --> 00:26:25.120]   And it's mostly about the bandwidth,
[00:26:25.120 --> 00:26:28.520]   the speed of the connection between your brain
[00:26:28.520 --> 00:26:34.080]   and your digital extension of yourself,
[00:26:34.080 --> 00:26:37.080]   particularly output.
[00:26:37.080 --> 00:26:39.840]   And if anything is getting worse,
[00:26:39.840 --> 00:26:42.800]   we used to have keyboards that would use a lot.
[00:26:42.800 --> 00:26:47.880]   Now we do most of our input through our thumbs on a phone.
[00:26:47.880 --> 00:26:51.040]   And that's just very slow.
[00:26:51.040 --> 00:26:54.720]   A computer can communicate at a trillion bits per second,
[00:26:54.720 --> 00:26:59.440]   but your thumb can maybe do, under 10 bits per second
[00:26:59.440 --> 00:27:02.760]   or 100 if you're being generous.
[00:27:02.760 --> 00:27:05.280]   So some high bandwidth interface to the brain,
[00:27:05.280 --> 00:27:10.800]   I think, will be something that helps achieve
[00:27:10.800 --> 00:27:15.080]   a symbiosis between human and machine intelligence
[00:27:15.080 --> 00:27:18.480]   and maybe solves the control problem
[00:27:18.480 --> 00:27:21.440]   and the usefulness problem.
[00:27:21.440 --> 00:27:26.720]   Getting pretty esoteric here.
[00:27:26.720 --> 00:27:28.000]   I don't know if this is--
[00:27:28.000 --> 00:27:29.000]   It's closed.
[00:27:29.000 --> 00:27:29.680]   We got it.
[00:27:32.840 --> 00:27:37.600]   Always you think, out of the box, your idea is so huge.
[00:27:37.600 --> 00:27:42.200]   You want to go to space.
[00:27:42.200 --> 00:27:43.480]   You decided to go to space.
[00:27:43.480 --> 00:27:44.560]   You did it.
[00:27:44.560 --> 00:27:48.280]   You decided that you want to land your rocket back.
[00:27:48.280 --> 00:27:51.200]   You fail seven time at a time.
[00:27:51.200 --> 00:27:51.720]   Yeah.
[00:27:51.720 --> 00:27:52.720]   So work that.
[00:27:52.720 --> 00:27:53.760]   Then landed--
[00:27:53.760 --> 00:27:55.000]   More time is an accurate account.
[00:27:55.000 --> 00:27:57.400]   Yeah.
[00:27:57.400 --> 00:27:59.520]   How do you come with this idea, actually?
[00:27:59.520 --> 00:28:02.520]   Sometimes they're pushing the human limit.
[00:28:02.520 --> 00:28:05.520]   You are always pushing the human limit.
[00:28:05.520 --> 00:28:06.520]   Why?
[00:28:06.520 --> 00:28:16.040]   Well, I think about what technology solution
[00:28:16.040 --> 00:28:19.400]   is necessary in order to achieve the particular goal
[00:28:19.400 --> 00:28:22.240]   and then try to make as much progress in that direction
[00:28:22.240 --> 00:28:22.880]   as possible.
[00:28:22.880 --> 00:28:26.520]   So in the case of space flight, the critical breakthrough
[00:28:26.520 --> 00:28:28.440]   that's necessary in space flight
[00:28:28.440 --> 00:28:31.800]   is rapid and complete reusability of rockets.
[00:28:31.800 --> 00:28:34.600]   Just as we have for aircraft.
[00:28:34.600 --> 00:28:38.640]   You can imagine that if an aircraft was single use,
[00:28:38.640 --> 00:28:42.240]   almost no one would fly.
[00:28:42.240 --> 00:28:49.320]   Because you can buy, let's say, a 747 might be $250 million,
[00:28:49.320 --> 00:28:50.840]   $300 million, or something like that.
[00:28:50.840 --> 00:28:54.880]   You need two of them for a round trip.
[00:28:54.880 --> 00:28:57.960]   But nobody's going to pay millions of dollars per ticket
[00:28:57.960 --> 00:29:03.280]   to fly to do air travel.
[00:29:03.280 --> 00:29:06.360]   But because you can reuse the aircraft tens of thousands
[00:29:06.360 --> 00:29:14.000]   of times, the air travel becomes much more affordable.
[00:29:14.000 --> 00:29:15.560]   And the same is true of rockets.
[00:29:15.560 --> 00:29:21.800]   So our rocket costs $60 million roughly.
[00:29:21.800 --> 00:29:27.600]   So a capital cost, if it can be used once, is $60 million.
[00:29:27.600 --> 00:29:32.400]   But the capital cost, if it can be used 1,000 times,
[00:29:32.400 --> 00:29:35.720]   is $60,000.
[00:29:35.720 --> 00:29:40.520]   So then if you can carry a lot of people per flight,
[00:29:40.520 --> 00:29:43.080]   then you can get the cost of space flight
[00:29:43.080 --> 00:29:48.400]   to be something not far from the cost of air flight.
[00:29:48.400 --> 00:29:50.920]   So it's extremely fundamental.
[00:29:50.920 --> 00:29:54.080]   Because Earth's gravity well is quite deep,
[00:29:54.080 --> 00:29:57.440]   Earth has a fairly high gravity.
[00:29:57.440 --> 00:30:00.840]   The difficulty of making a rocket reusable
[00:30:00.840 --> 00:30:02.840]   is much greater than the difficulty of making
[00:30:02.840 --> 00:30:04.600]   an aircraft reusable.
[00:30:04.600 --> 00:30:07.680]   And that's why a fully reusable rocket has never
[00:30:07.680 --> 00:30:09.240]   been developed thus far.
[00:30:09.240 --> 00:30:14.000]   But if you use the most advanced materials,
[00:30:14.000 --> 00:30:17.160]   most advanced design techniques, and you get everything
[00:30:17.160 --> 00:30:19.400]   just right, then I'm confident that you
[00:30:19.400 --> 00:30:22.720]   can do a fully reusable rocket.
[00:30:22.720 --> 00:30:27.480]   Fortunately, if Earth's gravity was even 10% stronger,
[00:30:27.480 --> 00:30:28.800]   I would say it would be impossible.
[00:30:28.800 --> 00:30:35.840]   You need a team around you to deliver a lot of ideas.
[00:30:35.840 --> 00:30:38.280]   How do you choose your team, be someone?
[00:30:38.280 --> 00:30:47.000]   Well, I suppose honestly that it tends to be gut
[00:30:47.000 --> 00:30:49.600]   feel more than anything else.
[00:30:49.600 --> 00:30:54.320]   So when I interview somebody--
[00:30:54.320 --> 00:30:56.440]   my interview question is always the same.
[00:30:56.440 --> 00:30:57.480]   It's just--
[00:30:57.480 --> 00:30:57.960]   What do you ask?
[00:30:57.960 --> 00:31:01.320]   I said, tell me the story of your life,
[00:31:01.320 --> 00:31:03.680]   and the decisions that you made along the way
[00:31:03.680 --> 00:31:05.800]   and why you made them.
[00:31:05.800 --> 00:31:11.960]   And then-- and also tell me about some of the most difficult
[00:31:11.960 --> 00:31:15.240]   problems you worked on and how you solved them.
[00:31:15.240 --> 00:31:21.280]   And that question I think is very important because the people
[00:31:21.280 --> 00:31:25.640]   that really solved the problem, they know exactly how they
[00:31:25.640 --> 00:31:27.160]   solved it.
[00:31:27.160 --> 00:31:29.200]   They know the little details.
[00:31:29.200 --> 00:31:32.120]   And the people that pretended to solve the problem, they can
[00:31:32.120 --> 00:31:35.440]   maybe go one level, and then they get stuck.
[00:31:35.440 --> 00:31:38.800]   So what was your biggest challenge in life?
[00:31:41.840 --> 00:31:46.120]   Biggest challenge in life?
[00:31:46.120 --> 00:31:47.120]   Hmm.
[00:31:47.120 --> 00:31:53.160]   Not yet.
[00:31:53.160 --> 00:31:56.080]   Well, no, there's a lot of them trying to say which is the
[00:31:56.080 --> 00:31:56.640]   worst.
[00:31:56.640 --> 00:32:04.800]   I think just thinking about how to spend time.
[00:32:04.800 --> 00:32:14.680]   One of the biggest challenges I think is making sure you
[00:32:14.680 --> 00:32:18.720]   have a correct to feedback loop, and then maintaining that
[00:32:18.720 --> 00:32:22.640]   correct to feedback loop over time, even when people want
[00:32:22.640 --> 00:32:25.240]   to tell you exactly what you want to hear.
[00:32:25.240 --> 00:32:26.480]   That's very difficult.
[00:32:26.480 --> 00:32:27.000]   Yes.
[00:32:27.000 --> 00:32:30.640]   Time's over.
[00:32:30.640 --> 00:32:33.000]   I'll ask you just one last question.
[00:32:33.000 --> 00:32:35.120]   If you allowed me.
[00:32:35.120 --> 00:32:40.800]   In the word government, somewhat we have so many people from
[00:32:40.800 --> 00:32:44.200]   so many young people, actually, from across the globe.
[00:32:44.200 --> 00:32:47.880]   If you have an advice to them, young people globally want to
[00:32:47.880 --> 00:32:52.840]   be like, "Elam Mosk, what's your advice today?"
[00:32:52.840 --> 00:32:58.720]   I think that probably they shouldn't want to be you.
[00:32:58.720 --> 00:33:00.200]   I think it sounds better than it is.
[00:33:00.200 --> 00:33:00.720]   OK.
[00:33:00.720 --> 00:33:04.040]   Yeah.
[00:33:04.040 --> 00:33:08.160]   It's not as much fun being me as you think.
[00:33:08.160 --> 00:33:08.640]   I don't know.
[00:33:08.640 --> 00:33:10.800]   You don't think so?
[00:33:10.800 --> 00:33:12.920]   There's definitely-- it could be worse for sure.
[00:33:12.920 --> 00:33:18.680]   But it's-- I'm not sure I would--
[00:33:18.680 --> 00:33:20.400]   I'm not sure I want to be me.
[00:33:20.400 --> 00:33:21.080]   OK.
[00:33:21.080 --> 00:33:23.040]   [LAUGHTER]
[00:33:23.040 --> 00:33:24.040]   But--
[00:33:24.040 --> 00:33:25.040]   [APPLAUSE]
[00:33:28.480 --> 00:33:31.560]   You know, I think advice-- I mean, if you want to make
[00:33:31.560 --> 00:33:39.240]   progress in things, I think that the best analytical framework
[00:33:39.240 --> 00:33:43.120]   for understanding the future is physics.
[00:33:43.120 --> 00:33:50.080]   I'd recommend studying the thinking process around physics.
[00:33:50.080 --> 00:33:52.120]   Not just-- not the equations.
[00:33:52.120 --> 00:33:54.280]   I mean, the equations certainly they're helpful.
[00:33:54.280 --> 00:33:58.880]   But the way of thinking in physics is--
[00:33:58.880 --> 00:34:01.120]   it's the best framework for understanding things
[00:34:01.120 --> 00:34:02.400]   that are counterintuitive.
[00:34:02.400 --> 00:34:09.400]   And always taking the position that you're
[00:34:09.400 --> 00:34:12.720]   some degree wrong and your goal is to be less wrong over time.
[00:34:12.720 --> 00:34:18.520]   One of the biggest mistakes people generally make--
[00:34:18.520 --> 00:34:22.480]   and I'm guilty of it too-- is wishful thinking.
[00:34:22.480 --> 00:34:26.800]   You want something to be true even if it isn't true.
[00:34:26.800 --> 00:34:29.840]   And so you ignore the things that--
[00:34:29.840 --> 00:34:35.040]   you ignore the real truth because of what you want to be true.
[00:34:35.040 --> 00:34:38.120]   This is a very difficult trap to avoid.
[00:34:38.120 --> 00:34:41.240]   And like I said, it's certainly one that I find myself
[00:34:41.240 --> 00:34:44.320]   in having problems with.
[00:34:44.320 --> 00:34:47.200]   But if you just take that approach of you're
[00:34:47.200 --> 00:34:51.640]   always to some degree wrong and your goal is to be less wrong.
[00:34:51.640 --> 00:34:56.120]   And solicit critical feedback, particularly from friends.
[00:34:56.120 --> 00:35:00.640]   Friends-- if somebody loves you, they want the best for you.
[00:35:00.640 --> 00:35:04.080]   They don't want to tell you the bad things.
[00:35:04.080 --> 00:35:06.280]   So you have to ask them--
[00:35:06.280 --> 00:35:10.520]   you know, and say, really, I really do want to know.
[00:35:10.520 --> 00:35:12.360]   And then they'll tell you.
[00:35:12.360 --> 00:35:15.040]   Thank you very much.
[00:35:15.040 --> 00:35:16.040]   It's been--
[00:35:16.040 --> 00:35:19.040]   [APPLAUSE]
[00:35:19.040 --> 00:35:27.280]   It is great for the world government summit
[00:35:27.280 --> 00:35:33.040]   to have a legend who's creating the future for humanity
[00:35:33.040 --> 00:35:37.960]   to share his thought, his idea, his vision, his challenges,
[00:35:37.960 --> 00:35:39.560]   and his hope for life.
[00:35:39.560 --> 00:35:40.320]   Thank you very much.
[00:35:40.320 --> 00:35:41.920]   Thank you, and have a great day.
[00:35:41.920 --> 00:35:43.920]   [APPLAUSE]
[00:35:43.920 --> 00:35:45.920]   [MUSIC PLAYING]
[00:35:45.920 --> 00:35:56.240]   [VIDEO PLAYBACK]


whisper_print_timings:     load time =   250.23 ms
whisper_print_timings:      mel time =  1918.48 ms
whisper_print_timings:   sample time =  1206.74 ms
whisper_print_timings:   encode time = 62509.04 ms / 10418.17 ms per layer
whisper_print_timings:   decode time = 231712.16 ms / 38618.69 ms per layer
whisper_print_timings:    total time = 297698.94 ms

real    4m57.746s
user    124m2.294s
sys     7m33.988s

[stevevaius2015]

Could you try it for large model also? Just wondered where we can access this cloud compute?

[silvacarl2]

question: does it default to automatically figuring out how many threads it can use, or does this always need to be passed from thec ommand line?