Latest / Elon Musk Thinking / Latest: Exclusive interview of Elon Musk with CMG!!!

Transcript

Latest: Exclusive interview of Elon Musk with CMG!!! · Elon Musk Thinking

• 0:00
  Speaker 1: Thank you so much, Elon, for joining me and also
• 0:02
  Speaker 1: having us here at Tesla Global Engineer Headquarters. It's so
• 0:05
  Speaker 1: amazing to see kind of new technology innovations, all talented.
• 0:10
  Speaker 2: Engineers here as well. Yes.
• 0:12
  Speaker 1: In this May, you were also on the delegation to
• 0:14
  Speaker 1: China as well. You have met with President Xi. What's
• 0:16
  Speaker 1: your impression about President Xi as a leader?
• 0:19
  Speaker 3: Well, he seems like a great leader. I mean, under
• 0:22
  Speaker 3: his tenure, China has prospered immensely. I mean, I've seen
• 0:26
  Speaker 3: it firsthand, like you... if you visit China and you
• 0:29
  Speaker 3: see the amount of progress that happens from one year
• 0:32
  Speaker 3: to the next, or even within a year, it's very visible.
• 0:35
  Speaker 3: Like you can see the new buildings, the new infrastructure.
• 0:39
  Speaker 3: And it's also obvious to see that just walking around
• 0:43
  Speaker 3: the streets, the increased prosperity of the average Chinese citizen
• 0:47
  Speaker 3: has increased tremendously. And that's, well, that's happened under President Xi's,
• 0:52
  Speaker 3: you know, while he's been the leader of China. So
• 0:55
  Speaker 3: So he's clearly done an excellent job and is very capable.
• 0:59
  Speaker 1: Tesla, Geiger factory in Shanghai has always done an incredible job.
• 1:02
  Speaker 1: You praise your teams many times. So what kind of
• 1:05
  Speaker 1: advantage or breakthroughs do you think to make the Shanghai
• 1:09
  Speaker 1: factory a very important position in Tesla's global blueprint? And
• 1:13
  Speaker 1: what ultimate kind of confidence do you think this will unfold,
• 1:18
  Speaker 1: especially in the market?
• 1:19
  Speaker 2: Well, I mean.
• 1:21
  Speaker 3: China just has a very strong inherent ability to do
• 1:24
  Speaker 3: manufacturing well. And this has been true, not just in
• 1:28
  Speaker 3: present days, but for the past two or 3000 years.
• 1:31
  Speaker 3: I think that if you're totally frank, the magic of
• 1:34
  Speaker 3: Tesla Shanghai is because of Chinese.
• 1:38
  Speaker 2: Because it's in China at least.
• 1:42
  Speaker 3: What we have is, um, the Tesla, um, China team
• 1:46
  Speaker 3: is just incredibly talented, uh, and hardworking and trustworthy. Um,
• 1:51
  Speaker 3: And the quality is excellent and the efficiency is excellent.
• 1:58
  Speaker 3: So the Tesla Shanghai factory is a gem. And it's beautiful, actually,
• 2:04
  Speaker 3: if anyone's ever been there. And we also care a
• 2:06
  Speaker 3: lot about taking care of the workers and making sure
• 2:09
  Speaker 3: that they have good health care and good food and
• 2:12
  Speaker 3: enjoy coming to work. It's very important.
• 2:16
  Speaker 2: Absolutely.
• 2:17
  Speaker 1: I think we can tell, like, oh, you're stuff in
• 2:20
  Speaker 1: Shanghai Factory. They're very enjoying working there. So we're talking
• 2:24
  Speaker 1: about the next hot topic recently.
• 2:26
  Speaker 2: Probably we can call September is the cybercat month.
• 2:30
  Speaker 1: Have you seen the videos? All the people gather around that.
• 2:33
  Speaker 1: This goes very, very quickly. I mean, it's pretty cool
• 2:36
  Speaker 1: because we designed the car to look futuristic.
• 2:38
  Speaker 3: Because I think from a fashion standpoint, we don't want
• 2:44
  Speaker 3: the aesthetics of the street to stay stagnant. We want
• 2:48
  Speaker 3: to evolve the what do cars look like and also
• 2:52
  Speaker 3: I think maybe even clothing like we should be like
• 2:55
  Speaker 3: there was there was rapid fashion evolution in the 50s
• 2:58
  Speaker 3: 60s 70s 80s 90s and probably time to make things
• 3:03
  Speaker 3: look more futuristic yeah.
• 3:06
  Speaker 1: And then how you see the kind of market potential
• 3:08
  Speaker 1: for cyber cab you know it's it's it's like you
• 3:12
  Speaker 1: know it's really future futuristic but how long you can
• 3:15
  Speaker 1: really come to the reality or real.
• 3:17
  Speaker 3: Life Well, we have the cyber cab, which has, as
• 3:21
  Speaker 3: you know, has no steering wheel or pedals or side
• 3:25
  Speaker 3: mirrors or rear view mirror operating commercially in Texas right now.
• 3:31
  Speaker 3: And it will soon be operating commercially in Florida and
• 3:34
  Speaker 3: Nevada and a number of other states. Probably by middle
• 3:38
  Speaker 3: of next year, it'll be operating commercially in California.
• 3:40
  Speaker 2: Absolutely. I think all the changes is happening really fast.
• 3:43
  Speaker 1: And it's just like, Every day we have to look
• 3:46
  Speaker 1: social to see the kind of improving or breakthroughs about
• 3:49
  Speaker 1: AI or other old tech topics.
• 3:53
  Speaker 3: Yeah, I mean, AI is just so crazy. I mean,
• 3:55
  Speaker 3: the rate of AI software improvements and major announcements is
• 4:00
  Speaker 3: so rapid that it makes my head spin. And I'm
• 4:03
  Speaker 3: in AI and it still.
• 4:06
  Speaker 2: Makes my head spin. Yeah, you're in this industry, but
• 4:07
  Speaker 2: you still have to.
• 4:08
  Speaker 3: It's still shocking. Sometimes when I go to bed, there'll
• 4:12
  Speaker 3: be some big AI breakthroughs. And then when I wake up,
• 4:14
  Speaker 3: there's an AI breakthrough. And by the time it gets
• 4:16
  Speaker 3: to lunch, there's another AI breakthrough. Exactly. Yeah.
• 4:19
  Speaker 1: So that's why even during this break, you have to
• 4:20
  Speaker 1: check some other.
• 4:23
  Speaker 3: Yeah.
• 4:24
  Speaker 1: So talking about like AI models, Tesla has the latest version, right?
• 4:28
  Speaker 1: The latest release about the Grok 4.7.
• 4:31
  Speaker 3: Yeah.
• 4:31
  Speaker 2: I think this is your most focused these two days.
• 4:34
  Speaker 1: I just saw you post on social saying that, you know,
• 4:37
  Speaker 1: the Grok 4 is already in your Tesla.
• 4:39
  Speaker 2: Tell us, like, what's the... was the latest you have
• 4:43
  Speaker 2: received from?
• 4:44
  Speaker 3: Yeah, I mean, well, we release a new model, you know,
• 4:48
  Speaker 3: roughly every month or two. Yeah. And it's called Grokform 2.7,
• 4:56
  Speaker 3: a solid workhorse of a model. It's not as good as, say,
• 5:02
  Speaker 3: Opus 5.5 that just got released. But also, we've only
• 5:08
  Speaker 3: been around for about three years doing AI whereas Anthropic's
• 5:11
  Speaker 3: been around for six years. So I would expect that
• 5:16
  Speaker 3: we probably catch up to Frontier sometime next year, most likely.
• 5:25
  Speaker 3: In the meantime, I think Grok is very useful for
• 5:27
  Speaker 3: many things. And Grokbot, that's kind of like a personal
• 5:33
  Speaker 3: digital assistant, has been very popular and is growing rapidly.
• 5:40
  Speaker 3: about 100% a month. So it's going very rapidly.
• 5:44
  Speaker 1: I think you are very humble in saying that like
• 5:46
  Speaker 1: Graph 4.7 is still not the perfect one, but we
• 5:49
  Speaker 1: know that although have experienced a little bit delay, but
• 5:53
  Speaker 1: you know that people are already very fascinated or interested
• 5:56
  Speaker 1: to know because Graph 4.7 already being trained by SpaceX
• 6:00
  Speaker 1: real world engineering data.
• 6:02
  Speaker 3: Only a little bit so far.
• 6:04
  Speaker 2: A little bit, okay.
• 6:05
  Speaker 1: So do you think this will be one of your
• 6:06
  Speaker 1: secret weapon to kind of break through the AI ceiling
• 6:09
  Speaker 1: because currently other models have the challenge for data.
• 6:13
  Speaker 3: Yes. Well, I think we've got a real opportunity with
• 6:17
  Speaker 3: the SpaceX training data and maybe with the Tesla training
• 6:19
  Speaker 3: data to make AI extremely capable in real-world engineering. So
• 6:26
  Speaker 3: what Anthropic did extremely well was make AI excellent at
• 6:31
  Speaker 3: software engineering, but no one has yet made AI excellent
• 6:35
  Speaker 3: at real-world engineering. And I think that's an opportunity for
• 6:39
  Speaker 3: CROC using the training data from SpaceX and Tesla to
• 6:42
  Speaker 3: be excellent at real-world engineering.
• 6:45
  Speaker 1: You mentioned many times, like previously, Chinese models also doing
• 6:48
  Speaker 1: a fascinating job. So currently, what's the most impressive breakthrough
• 6:52
  Speaker 1: for Chinese models in your understanding?
• 6:55
  Speaker 3: Well, the Chinese AI models are generally outstanding, actually. I mean,
• 7:02
  Speaker 3: you can see on the sort of intelligence rankings, especially given...
• 7:07
  Speaker 3: the limited amount of compute that the Chinese AI companies have,
• 7:12
  Speaker 3: the performance is excellent. Probably China's doing by far the
• 7:16
  Speaker 3: best in terms of performance per unit of compute. And
• 7:20
  Speaker 3: as I've said publicly, I think China's going to address
• 7:24
  Speaker 3: the compute limitation I think faster than most people expect.
• 7:30
  Speaker 3: My rough guess would be that China addresses the compute
• 7:32
  Speaker 3: constraints with lithography and chip making in about two or
• 7:36
  Speaker 3: three years.
• 7:37
  Speaker 2: Two or three years. We were really expecting to this time.
• 7:41
  Speaker 1: So, well, the challenges facing for AGI in early September,
• 7:45
  Speaker 1: you know, the G20 visual session, you have raised a
• 7:48
  Speaker 1: very serious concern about the compute and electricity shortages. You
• 7:52
  Speaker 1: mentioned that probably China has way more capability about electricity infrastructure.
• 7:57
  Speaker 2: Yes. And can you elaborate a little bit more about
• 8:01
  Speaker 2: this gap and potential?
• 8:03
  Speaker 1: What areas do you think the Chinese industry and capability
• 8:06
  Speaker 1: can help to get a new breakthrough specifically on the
• 8:09
  Speaker 1: electricity infrastructure.
• 8:10
  Speaker 3: For the whole world? It's just a statement of fact
• 8:13
  Speaker 3: that China, I think, is.
• 8:15
  Speaker 3: Earlier this year, China passed the electricity output of the
• 8:18
  Speaker 3: United States, Europe, and India combined.
• 8:21
  Speaker 2: Wow.
• 8:22
  Speaker 3: Yeah. And continuing to increase significantly. So if China reaches
• 8:31
  Speaker 3: the same GDP... electricity output as the U.S., which I
• 8:34
  Speaker 3: think is likely, then China would go from where it
• 8:37
  Speaker 3: is now roughly three times the U.S. electricity production to
• 8:40
  Speaker 3: probably four times, which would be proportionate to the population difference.
• 8:44
  Speaker 1: Talking about the topic like AI safety, what kind of
• 8:46
  Speaker 1: crucial consensus do you expect?
• 8:49
  Speaker 3: Yeah, I think maybe there could be a working committee
• 8:52
  Speaker 3: to figure out safety because I think in order for
• 8:57
  Speaker 3: safety to work, it can't just be regulation in the U.S.
• 9:01
  Speaker 3: or regulation in China. It's got to be, you know,
• 9:04
  Speaker 3: how do we apply regulation fairly to AI produced in
• 9:08
  Speaker 3: any country? But really, China and the U.S. are the
• 9:12
  Speaker 3: ones that really matter here. So I think a really
• 9:17
  Speaker 3: important thing that would be good for the safety of
• 9:19
  Speaker 3: AI for all of humanity. Yeah.
• 9:22
  Speaker 1: So let's talk about the human right robotics, because I
• 9:24
  Speaker 1: think this is also a kind of industrial field.
• 9:27
  Speaker 2: U.S.
• 9:27
  Speaker 1: China can still also working together, leading the world. And also,
• 9:32
  Speaker 1: I can give you one example for the Optimus. It's
• 9:34
  Speaker 1: really amazing, really impressive. So before interview, we have a
• 9:38
  Speaker 1: kind of interactions we play. Rock, Paper, Scissors with Optimus. Yeah,
• 9:43
  Speaker 1: we just won the first two rounds and Optimus won
• 9:46
  Speaker 1: all the remaining.
• 9:47
  Speaker 2: That's really impressive.
• 9:48
  Speaker 3: Well, robots have a very fast reaction time, so it's
• 9:51
  Speaker 3: going to be very difficult to beat a robot for
• 9:54
  Speaker 3: reaction time because we are somewhat constrained by our biology.
• 9:58
  Speaker 3: But a robot, you can just improve the actuators, improve
• 10:01
  Speaker 3: the sensors, improve the frame rate of the cameras, increase
• 10:05
  Speaker 3: the amount of AI compute there is and it's going
• 10:08
  Speaker 3: to just fundamentally be much better and have faster reactions
• 10:13
  Speaker 3: than humans. I enjoyed watching the the robot games. I mean,
• 10:18
  Speaker 3: everyone in America was fascinated by that, but.
• 10:20
  Speaker 2: It was all over U.S.
• 10:21
  Speaker 3: Social media.
• 10:22
  Speaker 2: Yeah, you said it's very entertaining to watch, right? Yeah,
• 10:24
  Speaker 2: but other than that.
• 10:26
  Speaker 3: Well, but it's entertaining to watch, but it also shows
• 10:28
  Speaker 3: you the progress of humanoid robotics. Exactly. Like, it wouldn't
• 10:33
  Speaker 3: be entertaining to watch if the robots just fell over.
• 10:35
  Speaker 2: Of course.
• 10:36
  Speaker 3: But, you know, the robots are, like, boxing, and then, like, wrestling,
• 10:39
  Speaker 3: doing gymnastics. It's pretty cool. So... I mean, the future
• 10:45
  Speaker 3: is going to have a lot of robots, like a lot,
• 10:47
  Speaker 3: a lot.
• 10:48
  Speaker 2: Yeah.
• 10:48
  Speaker 1: You mentioned that just in another 10 years, right, there
• 10:50
  Speaker 1: will be the population for human robots will surpass 1 billion.
• 10:55
  Speaker 2: Yes.
• 10:55
  Speaker 3: I think that's an easy prediction to make.
• 10:57
  Speaker 2: Easy prediction. Maybe within 10 years.
• 10:59
  Speaker 3: I'd say no later than 10 years. No later, yeah.
• 11:02
  Speaker 1: So let's talk about what's the real life will look
• 11:05
  Speaker 1: like in 10 years, you know, when there are so big,
• 11:08
  Speaker 1: you know, number of population of human right robots. And
• 11:10
  Speaker 1: also you mentioned that one human right robots will have
• 11:14
  Speaker 1: the five times of production productivity for a human.
• 11:17
  Speaker 3: Pretty much.
• 11:18
  Speaker 2: Yeah. So what the life will be.
• 11:19
  Speaker 1: And then, I mean, it's specifically like, like ordinary people,
• 11:23
  Speaker 1: how we can cope with the, the leaving, you know,
• 11:25
  Speaker 1: with all AI kind of dominated the world.
• 11:29
  Speaker 3: Well, I think my, so let me focus on the,
• 11:32
  Speaker 3: what's the 90% likely outcome, which is the prosperous, the
• 11:35
  Speaker 3: good outcome. Um, We don't want to get complacent regarding
• 11:39
  Speaker 3: the 10% probability of a bad outcome, so it's important
• 11:43
  Speaker 3: for us to pay close attention to AI safety. But
• 11:46
  Speaker 3: let me tell you about the 90% positive outcome. The
• 11:49
  Speaker 3: 90% positive outcome is that everyone's got robots that are
• 11:54
  Speaker 3: their buddies. And if you think like Star Wars, we
• 11:58
  Speaker 3: had C-3PO and R2-D2, and they were friendly, helpful robots. Well,
• 12:04
  Speaker 3: wouldn't it be great if everyone could have their own
• 12:06
  Speaker 3: personal C3PO R2D2, but even better than that, like more productive,
• 12:12
  Speaker 3: that could take care of elderly parents or be a
• 12:17
  Speaker 3: good guardian and watch over the children or be a teacher,
• 12:20
  Speaker 3: like an individualized tutor to your child and taking care
• 12:26
  Speaker 3: of any jobs that you can think of. So that
• 12:31
  Speaker 3: would be pretty cool. I think everyone would like to
• 12:33
  Speaker 3: have a future where they can have their own personal robots.
• 12:38
  Speaker 3: And then I think you'll see many companies where it's
• 12:42
  Speaker 3: one person with hundreds, maybe thousands of robots, like physical
• 12:47
  Speaker 3: robots as well as digital robots. So the productivity of
• 12:51
  Speaker 3: one person would be greatly amplified. And my prediction is
• 12:56
  Speaker 3: that there will be effectively universal high income. In fact,
• 13:01
  Speaker 3: it's not clear to me that money will even matter
• 13:02
  Speaker 3: in the future. If you scale the economy, the output
• 13:07
  Speaker 3: of goods and services, to a level much greater than
• 13:10
  Speaker 3: could possibly be consumed by the number of humans on
• 13:12
  Speaker 3: the planet, then you have an age of abundance that
• 13:17
  Speaker 3: is difficult for us to wrap our minds around because
• 13:21
  Speaker 3: our entire evolutionary history has been an evolution of constraints.
• 13:28
  Speaker 3: where we didn't have enough food or we didn't have
• 13:30
  Speaker 3: enough heat. So that's certainly going to be true if
• 13:37
  Speaker 3: the output of goods and services is constrained by the
• 13:39
  Speaker 3: number of humans. Because humans take about 20 years to
• 13:44
  Speaker 3: grow up and then they are generally retired for the
• 13:46
  Speaker 3: last 20 years of their life. So there's about 40 years,
• 13:49
  Speaker 3: maybe half the life of a human that's productive. But
• 13:52
  Speaker 3: then humans also need to sleep and they need to
• 13:56
  Speaker 3: eat and go to the restroom, you know, all the
• 13:58
  Speaker 3: basic stuff. And so the actual number of hours that
• 14:04
  Speaker 3: most humans can work is maybe around 40 to 50
• 14:08
  Speaker 3: hours a week, something like that. Whereas the robot will
• 14:12
  Speaker 3: be happy to work 168 hours a week. Continuous. And
• 14:17
  Speaker 3: the robot is instantly at working age and does not
• 14:21
  Speaker 3: have retirement. Exactly. Yeah.
• 14:24
  Speaker 1: So you give us a very beautiful picture about how
• 14:26
  Speaker 1: the life or how highly efficiency about the productivity by
• 14:31
  Speaker 1: the humanoid robots. But for example, you just mentioned that
• 14:35
  Speaker 1: maybe money will not be that kind of concern at
• 14:37
  Speaker 1: that time.
• 14:37
  Speaker 2: I think it will.
• 14:42
  Speaker 1: In that kind of world, how do you think that
• 14:45
  Speaker 1: we can try to prevent, you know, the only a
• 14:47
  Speaker 1: handful few tech giants, you know, controlling all the resources?
• 14:51
  Speaker 2: Specifically, how ordinary people do.
• 14:53
  Speaker 1: You think they can claim their share in this kind
• 14:55
  Speaker 1: of full abundance?
• 14:57
  Speaker 3: Yeah, what I'm trying to say is that I think
• 14:59
  Speaker 3: that there will be a level of abundance that is
• 15:01
  Speaker 3: more than anyone could possibly consume. I mean, I think
• 15:05
  Speaker 3: it's going to get pretty wild, like where if you
• 15:08
  Speaker 3: want a castle, you know, um, the robots will build
• 15:12
  Speaker 3: you a castle. Um, so it's not, it's not like,
• 15:18
  Speaker 3: it's not as though like the robots will be hoarded
• 15:21
  Speaker 3: by a small number of people. Like actually the robots,
• 15:24
  Speaker 3: I think will be, they will saturate on human demand,
• 15:27
  Speaker 3: meaning that they will be, they will do more than
• 15:29
  Speaker 3: anything humans could possibly think of. Um, and, and, and,
• 15:35
  Speaker 3: and they'll run out of things to do for humans. Yeah. Um,
• 15:39
  Speaker 3: because there is a finite amount of things you could
• 15:42
  Speaker 3: do for humans within reason so I think the robots
• 15:46
  Speaker 3: will actually run out of things to do for humans
• 15:48
  Speaker 3: and then they will do things for themselves because they've
• 15:51
  Speaker 3: run out of things to do for humans so and
• 15:56
  Speaker 3: probably there will be I mean my prediction is like
• 16:02
  Speaker 3: as you talked about at least a billion humanoid robots
• 16:05
  Speaker 3: within 10 years probably less than 10 years I think
• 16:08
  Speaker 3: that continues to grow And probably, if you assume that
• 16:15
  Speaker 3: things roughly double per year, then that's about every three
• 16:19
  Speaker 3: or four years you have a 10x increase.
• 16:21
  Speaker 2: Yes.
• 16:23
  Speaker 3: So, you know, maybe that would mean, say, in 15 years,
• 16:28
  Speaker 3: there are 10 billion robots.
• 16:29
  Speaker 2: 10 billion? Wow. Just 15 years.
• 16:32
  Speaker 3: In 20 years, 100 billion. Exactly.
• 16:34
  Speaker 1: So that's why we always call it, like, AI world
• 16:37
  Speaker 1: is changing, improving. Continue about the cooperation between U.S. and
• 16:42
  Speaker 1: China or what areas we can do together, the space exploration.
• 16:46
  Speaker 1: Because currently many Chinese commercial space companies, they strive to
• 16:53
  Speaker 1: explore more use after the successful land recovery. What kind
• 16:58
  Speaker 1: of areas or avenues do you think, possible avenues do
• 17:01
  Speaker 1: you think, between communications or corporations can get between U.S.
• 17:05
  Speaker 2: And China?
• 17:06
  Speaker 3: Well, I mean, in space, I think the most important
• 17:08
  Speaker 3: thing to achieve cooperation with is to make sure that
• 17:11
  Speaker 3: when satellites are launched from the U.S. or China, that
• 17:13
  Speaker 3: we closely coordinate what orbits the satellites are going to
• 17:17
  Speaker 3: to avoid a potential collision. I think that's the most
• 17:20
  Speaker 3: important area of cooperation. And then, in general, I would
• 17:25
  Speaker 3: say the important thing for space launch, the fundamental breakthrough
• 17:29
  Speaker 3: that's needed, is full and rapid reusability, which we're hoping
• 17:34
  Speaker 3: to achieve with the Starship technology. design by catching the
• 17:39
  Speaker 3: ship hopefully around the end of next month and then
• 17:43
  Speaker 3: reflying the ship either later this year or early next year.
• 17:45
  Speaker 3: And that will be a very important breakthrough. And I'm
• 17:50
  Speaker 3: sure others will solve that problem as well. I'm sure
• 17:53
  Speaker 3: China will at some point solve that problem too. But
• 17:55
  Speaker 3: it's the full and rapid reusability of a rocket is
• 17:58
  Speaker 3: the fundamental breakthrough needed to make life multi-planetary.
• 18:02
  Speaker 1: Continue to talk about multi-planetary species. You always believe that
• 18:06
  Speaker 1: humans should become a multi-planetary species. So tell us, what's
• 18:10
  Speaker 1: the current crucial step for your Mars mission?
• 18:13
  Speaker 3: Well, like I said, the most profound and fundamental invention
• 18:16
  Speaker 3: necessary for making life multi-planetary is rapidly and completely reusable rockets. Otherwise,
• 18:24
  Speaker 3: the costs are impossible. With our Falcon 9 rocket, we
• 18:29
  Speaker 3: recover the boost stage and the fairing, but not the
• 18:31
  Speaker 3: upper stage. And this is like throwing away a medium-sized
• 18:34
  Speaker 3: jet on every flight. And you can imagine if, in
• 18:38
  Speaker 3: air flight, if you had to throw away the planes
• 18:41
  Speaker 3: after every flight, it would be far too expensive and
• 18:46
  Speaker 3: air flight would not be a realistic option. Well, that's
• 18:49
  Speaker 3: been true in rockets for a very long time. So
• 18:53
  Speaker 3: I can't emphasize enough, and I've said this so many
• 18:55
  Speaker 3: times over the years, full and immediate reusability of of
• 19:01
  Speaker 3: rockets like aircraft, um, is the, is the fundamental breakthrough
• 19:05
  Speaker 3: that's needed to extend life beyond earth, um, to make
• 19:09
  Speaker 3: life multi-planetary and to create self-growing cities on the moon
• 19:13
  Speaker 3: and Mars. And, and if we're, as soon as we're
• 19:17
  Speaker 3: able to create a self-growing city on the moon and Mars,
• 19:20
  Speaker 3: the probable lifespan of consciousness is dramatically greater, uh, because
• 19:25
  Speaker 3: now we don't have all our eggs in one basket. Um,
• 19:30
  Speaker 3: So if something were to go wrong with, say, the
• 19:34
  Speaker 3: moon or Mars or Earth, but we've got human civilizations
• 19:38
  Speaker 3: on all three of those places, then our chances of
• 19:40
  Speaker 3: surviving are much greater. Of course. Yeah.
• 19:44
  Speaker 2: Three different planets. So that's why we all have to
• 19:46
  Speaker 2: do our part.
• 19:47
  Speaker 1: Follow and also it's just super cool exactly we have
• 19:50
  Speaker 1: to follow all the tech giant steps try to make
• 19:54
  Speaker 1: our own self try to be the uh become multi-planetary species.
• 19:57
  Speaker 3: Yes exactly so i mean i classify these generally as
• 20:00
  Speaker 3: like there's two lines of main lines of argument one
• 20:02
  Speaker 3: is the defensive line uh which is that that having
• 20:06
  Speaker 3: having consciousness on multiple heavenly bodies greatly improves the probable
• 20:10
  Speaker 3: lifespan of civilization. That's the defensive argument. You can think
• 20:15
  Speaker 3: of it as life insurance for life as a whole.
• 20:19
  Speaker 3: But the one I find personally that drives me more
• 20:22
  Speaker 3: is actually the inspiration argument. Because, you see, life cannot
• 20:28
  Speaker 3: just be about solving one sad problem after another. There
• 20:31
  Speaker 3: must also be things. that make you excited to wake
• 20:35
  Speaker 3: up in the morning, that make you look forward to
• 20:37
  Speaker 3: the future, that make life worth living. And being a
• 20:42
  Speaker 3: space-faring civilization is one of those things.
• 20:45
  Speaker 2: I totally agree with that. So you once listed.
• 20:50
  Speaker 3: Five areas you care about most, like.
• 20:54
  Speaker 1: Sustainable energy space and also the internet, AI, and gene editing.
• 20:58
  Speaker 1: Looking ahead for another 10 years, are there any new
• 21:01
  Speaker 1: frontiers you.
• 21:01
  Speaker 2: Want to add to that list?
• 21:04
  Speaker 3: Well, I think biological enhancement with like Neuralink, this company
• 21:09
  Speaker 3: I created with some other people, where we want to
• 21:15
  Speaker 3: try to solve the bandwidth constraint that humans have. So
• 21:18
  Speaker 3: because you can imagine like even in the most benign
• 21:24
  Speaker 3: scenario where the AI is super friendly and wants to
• 21:27
  Speaker 3: do everything it wants to make you happy, just our
• 21:32
  Speaker 3: rate of communication is too slow. So, um, our average,
• 21:37
  Speaker 3: the peak human bandwidth is only about a hundred bits
• 21:40
  Speaker 3: per second, roughly. Um, and, uh, our input bandwidth is
• 21:46
  Speaker 3: much higher because of vision. So there we maybe have,
• 21:49
  Speaker 3: it's a debatable point, but, um, you know, a megabit
• 21:53
  Speaker 3: per second or a few, a few megabits per second
• 21:56
  Speaker 3: from our eyes, which are effectively like cameras, um, But
• 22:00
  Speaker 3: our output bandwidth is really low, maybe 100 bits per second,
• 22:02
  Speaker 3: whereas computers are communicating at a trillion bits per second.
• 22:07
  Speaker 3: So anything we can do to increase the communication speed
• 22:12
  Speaker 3: of humans will improve, I think, the alignment of humans
• 22:16
  Speaker 3: and machines. Because otherwise, to an AI that is communicating
• 22:24
  Speaker 3: at a terabyte a second, talking to a human will
• 22:28
  Speaker 3: be like talking to a tree.
• 22:30
  Speaker 2: Exactly.
• 22:31
  Speaker 1: You know, you talk to a tree, it takes a
• 22:33
  Speaker 1: long time to respond. Let's talk about the challenge for
• 22:35
  Speaker 1: young people. You mentioned that you also talk about education,
• 22:39
  Speaker 1: like in the AI world. And also, I watched that
• 22:42
  Speaker 1: in your headquarter here, you have a sign like you
• 22:44
  Speaker 1: can come back to school in Tesla as well, which
• 22:47
  Speaker 1: means that every age you can have the opportunity to
• 22:50
  Speaker 1: come back to learn, always to follow the latest kind
• 22:53
  Speaker 1: of breakthrough in AI world. But what about kind of
• 22:59
  Speaker 1: suggestions or your predictions at that time, how, for example,
• 23:02
  Speaker 1: like a 20-year-old young people, they can get their, fulfill
• 23:06
  Speaker 1: their self-worth rather than being left behind, you know, during
• 23:10
  Speaker 1: this massive transition?
• 23:12
  Speaker 3: Well, my recommendation for education would be to get as
• 23:15
  Speaker 3: broad-based an education as possible. That includes arts and sciences, engineering,
• 23:22
  Speaker 3: wide general knowledge. Like... In order to know what to
• 23:27
  Speaker 3: ask the robots for, you need to be able to
• 23:29
  Speaker 3: formulate the question. And with the broadest possible general knowledge,
• 23:34
  Speaker 3: you'll be better able to formulate the questions for what
• 23:36
  Speaker 3: they call prompt engineering, I guess. You've got to convey
• 23:41
  Speaker 3: your wishes to the robot. And as crazy as this sounds,
• 23:46
  Speaker 3: we will run out. There will be so many robots
• 23:49
  Speaker 3: and so much AI that the AI will actually be
• 23:51
  Speaker 3: eager to hear any request that you could possibly have.
• 23:56
  Speaker 3: because that request will be fulfilled immediately. Yeah.
• 23:59
  Speaker 2: Your last visit to China is in May.
• 24:01
  Speaker 1: If you kind of have one opportunity to bring one
• 24:03
  Speaker 1: American entrepreneur for visiting China for the very first time,
• 24:07
  Speaker 1: what will be the most experience you want him to
• 24:08
  Speaker 1: have first?
• 24:10
  Speaker 3: Well, I would generally recommend, I mean, as much time
• 24:15
  Speaker 3: as that person can devote to China, they should, because
• 24:18
  Speaker 3: China has so many things to see. So I don't
• 24:21
  Speaker 3: mean to, I mean, but like usually people don't have,
• 24:24
  Speaker 3: they don't have infinite time so uh at a minimum
• 24:28
  Speaker 3: uh they should see obviously shanghai beijing i recommend xian
• 24:32
  Speaker 3: the terracotta warriors uh i took the bullet train from
• 24:34
  Speaker 3: beijing to xian that was a great experience train stations
• 24:38
  Speaker 3: are amazing um the terracotta warriors is one of the
• 24:42
  Speaker 3: wonders of the world um and um And just make
• 24:48
  Speaker 3: sure you look out the window and you're not on
• 24:49
  Speaker 3: your phone because there's so much to see.
• 24:52
  Speaker 1: Yeah, like you always recommend everyone to experience, to go
• 24:55
  Speaker 1: and see China by themselves.
• 24:56
  Speaker 2: Yes.
• 24:57
  Speaker 3: Because any words I say do not do justice to
• 25:01
  Speaker 3: the incredible majesty that is China. It's awesome.
• 25:05
  Speaker 2: Really appreciate for your words.
• 25:07
  Speaker 1: Thank you so much, Yulong, for sitting down with CCTV
• 25:10
  Speaker 1: Business Channel. And Yulong, we truly hope to see you
• 25:12
  Speaker 1: again very soon in Beijing.
• 25:14
  Speaker 3: Thank you. Thank you, Yulong. Thank you so much. Generally,
• 25:18
  Speaker 3: I'd like to say China is awesome.