Latest / Elon Musk Thinking / Elon Musk: What Do You Think About AI in the next 5 to 10 years?

Transcript

Elon Musk: What Do You Think About AI in the next 5 to 10 years? · Elon Musk Thinking

• 0:00
  Change of the world, and he's continuing to change all of us with wonderful
• 0:05
  vehicles. He has brought to us a space launch system which his efficient as
• 0:09
  nothing before. New people on the on the space from yours ground after the
• 0:14
  Space Shuttle, and I could go on far for an hour, but I
• 0:16
  don't do that. Welcome, big hands, you know a musk My good
• 0:23
  hetch how how are you? Ah? Good? Fantastic? We have we
• 0:35
  have we have two days of AI and SHANII and whatsever But you look real.
• 0:43
  My head is very big in the screen and in reality too. We
• 0:49
  haven't had so much talk about AI in these two days, so we need
• 0:53
  your opinion. What do you think about the next five to ten years on
• 0:57
  AI? Please give us a view Official Insemination Division Intelligence. What do you
• 1:03
  think about that? But try try, audience, and I'll say so yes.
• 1:11
  Artificial intelligence is is certainly going to profoundly change the world. One of
• 1:19
  the most significant ways is self driving. So we've been putting a lot of
• 1:27
  effort into self driving technology for cars. You don't see it quite as much
• 1:34
  in Europe because we first have to make it work in the US before we
• 1:40
  increase the complexity of trying to make it elsewhere in the world. But I
• 1:45
  think we're quite close to having the car be fully autonomous. So for example,
• 1:49
  like right now, I could I'm in Austin and if I want to
• 1:53
  try, say to the airport, the car could take me to the airport
• 1:56
  with no interventions, using or it's using our digital neural nets, in other
• 2:02
  words, artificial intelligence and cameras, so there's no light, our radar,
• 2:07
  nothing. And in fact, if you think about how humans strive cars,
• 2:14
  humans are biological neural nets, and we use eyes, so it's the eyes
• 2:19
  and but biological neural nets. The analog is digital. The analog for digital
• 2:27
  is cameras and digital neural nets. This is working remarkably well that it has
• 2:34
  been quite difficult to do this because TNS out that in order to use that.
• 2:38
  In order for this to work, the car has to really be quite
• 2:43
  fully intelligently as a subset, for example, has to learn how to read
• 2:46
  everything and and and how to assess intention among drivers and pedestrians. So it's
• 2:53
  it's it's you're not creating sort of a baby artificial general into legence to solve
• 3:00
  this. But this will obviously be extremely profound, and I'm talking about what
• 3:06
  a motive here. I'll get into the broader question in a minute, but
• 3:09
  the average use of a passenger vehicle is only about ten hours per week out
• 3:17
  of one hundred and sixty eight hours. At the point at which it is
• 3:21
  autonomous, that I think you would see probably a third of the hours being
• 3:24
  used because it would be like a taxi. So maybe my guess is like
• 3:30
  fifty to sixty hours, which means that the utility of a passenger vehicle would
• 3:36
  increase by a factor of five. And this is gigantic, but we'll see.
• 3:46
  So that that's the automotive. Then there's what people think. I have
• 3:51
  certainly experienced online where you can ask it a question and ask for an essay
• 3:55
  or a picture, or to understand a picture, and that is progressing rapidly.
• 4:01
  But what I'm seeing in terms of AI compute is I've never seen any
• 4:06
  technology advance faster than this. The the artificial intelligence compute coming online appears to
• 4:15
  be increasing by a factor of ten every six months like now. Obviously that
• 4:23
  cannot continue at such a high rate for forever or will exceed the mast of
• 4:28
  the universe. But I've never seen anything like it. And this is why
• 4:32
  you see in videos market cap being so gigantic because they currently have the best
• 4:41
  neural net chips. I mean, I think the in videos market cap exceed
• 4:45
  the GDP of Canada or something. Increasingly, Awa is quite high. So
• 4:50
  yeah, you know, and it may go higher. Who knows. The
• 4:57
  chip rush is bigger than any gold rush that has ever existed. So and
• 5:05
  then there will also be robots, like human humanoid robots, not not just
• 5:10
  industrial industrial industrial robots we are core familiar with. You use industrial robots,
• 5:14
  We use industrial robots. Everyone does, but they're very programmed explicitly and they
• 5:18
  don't walk around. But Tesla with for example, the Optimist program is making
• 5:25
  a humanoid robot that is capable of doing almost anything a human can do.
• 5:34
  I mean, I hope, I just hope the robots are nice to us.
• 5:38
  You know, we hope. So do you know that the robots will
• 5:44
  be nice to us? And maybe, let's say, in this regard,
• 5:46
  because you are working on optimals, what can we expect in terms of let's
• 5:54
  say, support, which is going to come Maybe also a little bit with
• 5:59
  regard to production, because in the past Connected World we were also discussing application
• 6:06
  of ai jen ai into production in order to boost productivity. Maybe a couple
• 6:15
  of thoughts on this one too. Well. The plus side of air is
• 6:21
  that I think productivity will increase dramatically so across every field, whether it be
• 6:30
  manual labor to supply chain logistics. There's already a lot of movement to use
• 6:41
  chat bots, like sophisticated chat bots for customer service, for example, where
• 6:46
  they can answer quite quite complex questions already. But I mean, I think
• 6:56
  we really are on the edge of one of the biggest technology revel that has
• 7:00
  ever existed. You know, there's supposedly a sort of a Chinese curse,
• 7:05
  may you live in interesting times? Well, we live in the most interesting
• 7:10
  of times, the most interesting And for a while there I was a little
• 7:15
  bit it's making me a bit a bit depressed, frankly, because I was
• 7:19
  like, well, will we be will they take over? Will we be
• 7:24
  useless? But then the way that I reconciled myself to this question was would
• 7:29
  I rather be alive to see AI apocalypse or not? I'm like, I
• 7:33
  guess I'd like to see it. You know, that's a good thought.
• 7:46
  To be more, that's a good thought because you mentioned it tenfold every six
• 7:49
  months that's an exponential curve, right, Exponential curves are beautiful, but there's
• 7:54
  kind of a question where does it ever end, right, do you see
• 7:57
  any and yet? Or would it go on for three four five years like
• 8:00
  this, because then we talk hundred folds right, thousand folds off time?
• 8:03
  Yeah, that's why obviously, Yeah, I mean you can't. You can't
• 8:07
  grow at that rate without quickly exceeding the mass of the universe. So clearly
• 8:13
  it's going to hit some constraints. The constraints that AI compute and very predictable.
• 8:18
  I've actually predicted this over a year ago. So over here a go
• 8:22
  the shortage was chips, neural net chips. Then it was very easy predict
• 8:28
  that the next shortage will be voltage step down transformers, So because you've got
• 8:35
  to feed the power to these things. So if you've got one hundred to
• 8:39
  three hundred killer bolts coming out of your utility and it's got to step all
• 8:43
  the way down to you know, point six bolts, there's a lot of
• 8:46
  stepping down. So now we're in step down. And this is my not
• 8:50
  that funny joke, which is that they need transformers to run transformers because you
• 8:56
  know, the AI is like, there's this thing called transformer in AI.
• 9:01
  It's neural I don't know, it's a combination of sort of neural nets,
• 9:07
  and they're running out of transformers. Run transformers. Then the next shortage will
• 9:11
  be electricity. So I think next year you'll see the electricity that they just
• 9:16
  can't find and electricity run all the chips. I mean, during the chip
• 9:20
  crisis, we certainly all did have many contacts. Fortunately we were able to
• 9:26
  master it together. Of course, we all would like to avoid a second
• 9:30
  chip crisis. But with regard to your to your comment on power demand,
• 9:37
  we all know today vehicles have three to four kilowatt power demand for the peripherals.
• 9:45
  But this is continuously going up, and Teslam has recently launched the cyber
• 9:50
  truck, which I believe is mostly coming with forty eight volt. Do you
• 9:54
  see this as a let's say, further disruption of automotive industry, another technology
• 10:01
  change with an automotive, Yeah, I think you know the one motive industry
• 10:09
  has been living with twelve volts for a very long time, a century roughly.
• 10:16
  I think it started off as six volts, and then they needed a
• 10:20
  more powerful starter voter, and then they doubled it to twelve volts, although
• 10:26
  it's actually more like thirteen point seven volts. You're right, absolutely, yeah,
• 10:33
  and the bus voltage actually varies a lot for those since there's probably some
• 10:37
  electrical engineers in that audience. It's like twelve volts is a very rough approximation,
• 10:43
  but it's really it's really kind of an arbitrary low voltage. And if
• 10:48
  you look at say power Ethernet, it's running at fifty volts or roughly forty
• 10:52
  eight volts, and we setted forty forty eight bolts because anything really above that
• 10:58
  you start to running too regulatory concerns, and power Ethernet is, like you
• 11:03
  said, roughly at that level. But the advantage of obviously quadrupling the voltage
• 11:09
  is that you can substantially reduce how much copper or conductor is used in the
• 11:13
  car, so you can use very small copper wires for you know, it's
• 11:20
  roughly roughly a quarter, not exactly a quarter, but roughly a quarter as
• 11:24
  much as much copper is needed for forty eight volts as for twelve. So
• 11:30
  I think it's it's a logical move. That's that's that's really the next step
• 11:37
  for the low voltage architecture for the car. I think ultimately long term WOR
• 11:41
  class will be forty eight volts. It's it's interesting that many people love electric
• 11:48
  vehicles and want to have them, but some are harmed by a missing infrastructure.
• 11:52
  That would be the same actually as the II thing. We will have
• 11:56
  lots of chips ready, but we will have fed the power to operate them.
• 11:58
  There's a race between infrastructure and technology right now. And you're also certain
• 12:03
  technology may win, which is a problem because actually infrastructures would win. We
• 12:07
  should have all the infrastructure we need. Is that correct? Yeah, are
• 12:13
  you talking about like sort of charging infrastructural cars and that kind of thing.
• 12:18
  Well, the reason tesn't developed the sort of global supercharger network was specifically to
• 12:24
  address that point and the point you're making, which is people need to be
• 12:28
  comfortable that they can travel long distances and charge their car wherever they go.
• 12:35
  Unfortunately, electricity is almost everywhere, and so it's you know, it's not
• 12:43
  all not all that rare unless you're really far out in the countryside or something.
• 12:46
  So yeah, you need the super you need the charging infrastructure and like
• 12:54
  the electricity to charge electric cars. In fact, the simultaneous growth of electric
• 12:58
  cars and AI, both of which need electricity, both of which need voltage
• 13:03
  transformers. Actually, I think is creating a tremendous demand for electrical equipment and
• 13:11
  for electrical power generation. Maybe one more question elon with regard to new vehicle
• 13:20
  designs coming to the market. Of course you cannot share with us confidential secrets,
• 13:26
  but anyways, what are the things you believe, let's say we will
• 13:31
  see over the next years in the market. And maybe just a small comment.
• 13:35
  In Europe, we see a shortage of affordable battery electric vehicles in the
• 13:41
  offerings because many people in Europe are only able to spend like twenty five thousand
• 13:48
  euros for their vehicles. So what are the new vehicle types vehicle designs you
• 13:54
  are thinking about at Tesla, Well, our next generation vehicle will be a
• 14:03
  low, a lower cost vehicle, but one also that is very focused on
• 14:09
  autonomy. Like I said, the we really are getting to the point where
• 14:18
  the car can drive itself. Well, when you're in the in the US,
• 14:22
  I would recommend perhaps giving you know, trying out our car and see,
• 14:26
  I seeing how it works. It's remarkable, you know. As the
• 14:33
  AI gets better and better, it actually feels quite human like in the way
• 14:35
  that it drives. It does things that you wouldn't expect a computer to do.
• 14:41
  It's it's not like a robot dancing, you know, it's like it
• 14:43
  feels smooth and intuitive. So our next generation vehicle, which were quite far
• 14:52
  along on, is very much anticipating and autonomy. And then of course we've
• 14:58
  got some fun things as well, like on our the new Tesla Roadster,
• 15:03
  which will be able to do zero two hundred kilometers in under one second.
• 15:11
  Great, thank you very much. You know that's normally need. That's good.
• 15:16
  Yeah, that's good. Some fun stuff too, you know, that's
• 15:20
  the that's the dessert, that's the cherry on the cake. But it's fun.
• 15:24
  You've got to have the fun stuff too. Yeah, right, No,
• 15:26
  that's very good news because in the end, there's a lot of people
• 15:30
  here in the room that's still on their list to buy an electric vehicle,
• 15:33
  and you just announced something which they probably want to buy. So thank you
• 15:37
  very much, Thank you very much, and thank you extremely for fifteen years
• 15:41
  of partnership and work together. We're inspired always by your speed attitude of graphics
• 15:46
  and things come out. Thank you very much too. Musk all right,
• 15:50
  and thank you for your support, very very much appreciate it.