WEBVTT 1 00:00:00.200 --> 00:00:09.080 Why are you skeptical to connecting? Let's playing with the soundboard here. Okay, 2 00:00:09.199 --> 00:00:12.000 Now, what's your take on where we are in the AI race? 3 00:00:12.119 --> 00:00:15.480 Just now? Wow, that's a long answer. There's so much happening in 4 00:00:15.759 --> 00:00:19.480 AI is the fastest advancing technology that I've ever seen of any kind. And 5 00:00:19.480 --> 00:00:22.960 I've seen a lot of technology. Barely a week goes by without some new 6 00:00:23.000 --> 00:00:27.399 announcement. So, and if you look at the amount of AI hardware, 7 00:00:27.480 --> 00:00:31.320 the computers coming online that are dedicated to AI, that is increasing what looks 8 00:00:31.359 --> 00:00:35.679 like at least by a factor of ten every year, if not every sixty 9 00:00:35.759 --> 00:00:39.759 nine months. So when you combine the hardware coming online, really order of 10 00:00:39.799 --> 00:00:43.439 magnitude increase every you call it, at least every nine months, many many 11 00:00:43.840 --> 00:00:47.079 software breakthroughs. If you look at that, that COVID looks insane. So 12 00:00:47.240 --> 00:00:51.320 I think, well, my guess is that we will have AI that is 13 00:00:51.320 --> 00:00:55.159 smaller than any any one human probably around the end of next year, and 14 00:00:55.200 --> 00:00:58.479 then AI, the total amount of sort of sentient compute of AI, I 15 00:00:58.600 --> 00:01:03.280 think will probably exceed all humans in five years. What is the race about? 16 00:01:03.560 --> 00:01:07.319 Just now? Is it ALGORI is a people? Is it computing power? 17 00:01:07.640 --> 00:01:10.120 What is it about? Just now? Is it the supply of chips? 18 00:01:10.359 --> 00:01:12.439 Just what is it? Yeah? Last year it was chip constraint and 19 00:01:12.480 --> 00:01:17.719 the hardware deployment to break it down into the three areas of people, data, 20 00:01:17.879 --> 00:01:19.840 and hardware are starting with hardware. Last year it was about a chief 21 00:01:19.879 --> 00:01:23.159 supply who could not get enough video chiefs. Particularly this year is starting to 22 00:01:23.200 --> 00:01:30.239 transition to a voltage transformer supply, so actually getting enough voltage transformers put in 23 00:01:30.280 --> 00:01:34.680 place. So my sort of very niche joke is transformers for transformers, because 24 00:01:34.719 --> 00:01:38.000 a lot of the AI that's around is called a transformer, need transformers to 25 00:01:38.040 --> 00:01:42.079 run transformers. And then next in the if you look out a year or 26 00:01:42.079 --> 00:01:47.040 two or certainly three years, it's just electricity availability, so that those are 27 00:01:47.040 --> 00:01:51.560 those constraints on the hardware side. So many of the smartest people are doing 28 00:01:51.719 --> 00:01:55.959 AI. People that would have done physics before, in fact have done physics 29 00:01:55.959 --> 00:01:59.519 for example, have moved into AI because it's just the fastest moving field. 30 00:02:00.079 --> 00:02:02.200 So we're seeing a lot of the best talents, a lot of the smartest 31 00:02:02.280 --> 00:02:07.560 humans going into AI, and then we see along with that algorithmic breakthroughs, 32 00:02:07.719 --> 00:02:12.479 and then then you start hitting the wall with the data problem. So the 33 00:02:12.719 --> 00:02:15.680 you know, you can fit all books ever written just a text, the 34 00:02:16.000 --> 00:02:21.360 text in compressed form on one hard drive or called on one computer. So 35 00:02:21.400 --> 00:02:24.560 when you're when you're looking at like called tokens to train on, and you 36 00:02:24.599 --> 00:02:29.400 because you still give like all the books ever written in every in all languages, 37 00:02:29.439 --> 00:02:31.439 by all humans, sounds like a lot. Certainly is far more than 38 00:02:31.439 --> 00:02:35.840 any one human could could ever read. It actually is a small note. 39 00:02:36.280 --> 00:02:39.000 It's a small number of training tokens. It's just not enough. So then 40 00:02:39.039 --> 00:02:45.000 you start having to look at all the videos that are created, you know, 41 00:02:45.039 --> 00:02:49.800 all the podcasts, all the everything, and and you start even running 42 00:02:49.840 --> 00:02:53.319 out of data there. Well, hopefully hopefully they will include these podcasts. 43 00:02:53.159 --> 00:02:59.039 Definitely will include this podcast. What's the biggest challenge you have with X Well, 44 00:02:59.039 --> 00:03:01.080 I say, it is still relatively new, so it's not you know, 45 00:03:01.319 --> 00:03:07.639 like the limiting factor right now is just training our GROC version two model, 46 00:03:07.680 --> 00:03:10.840 which should be we think better than GPD four and that's we're hoping to 47 00:03:10.840 --> 00:03:14.879 complete that in May. So that's that's training right now. So it's just 48 00:03:14.919 --> 00:03:17.759 really we're just trying to get enough GPUs online to train it fast enough to 49 00:03:19.080 --> 00:03:23.000 get that done in May, which I think probably will happen. And then 50 00:03:22.800 --> 00:03:28.360 and that's with roughly twenty thousand, h one hundreds and I'm doing I think 51 00:03:28.479 --> 00:03:31.240 very efficient training. Then the next step would be for GROCK three, which 52 00:03:31.240 --> 00:03:37.479 should be I guess GPD five or beyond. Would you know requires one hundred 53 00:03:37.479 --> 00:03:42.800 thousand in video h one hundreds training coherently, So that's you know, half 54 00:03:42.919 --> 00:03:46.039 order of vangetude basically more training. And then you really start to have running 55 00:03:46.039 --> 00:03:51.159 into this data problem where you have to either create synthetic data or use real 56 00:03:51.199 --> 00:03:54.800 world video. Those the two sources of kind of like unlimited data synthetic data 57 00:03:54.879 --> 00:03:59.400 and real world video, which I should say Tesla has a pretty big advantage 58 00:03:59.400 --> 00:04:01.479 in real world. It'll has my father's most real world video of anyone. 59 00:04:01.599 --> 00:04:03.599 Yeah, you've got a huge library. So when do you think, so 60 00:04:03.639 --> 00:04:06.680 when you think we'll see proper AGI, Well, in terms on how you 61 00:04:06.680 --> 00:04:11.000 define a GI, if you define a GI as smarter than as mouse human, 62 00:04:11.080 --> 00:04:14.800 I think it's probably any next year, within two years. But that's 63 00:04:14.800 --> 00:04:17.560 that there's still there's still a pretty big leap beyond that, to say smarter 64 00:04:17.680 --> 00:04:24.199 than the machine augmented human collective so like, is it smarter than all humans 65 00:04:24.240 --> 00:04:30.399 working together who are also using computers to augment their output, and that that 66 00:04:30.519 --> 00:04:33.160 I think is probably five years away. One way to look at it is 67 00:04:33.160 --> 00:04:36.319 is to try to assess, like, roughly, what is the ratio of 68 00:04:36.399 --> 00:04:42.639 digital to biological compute and so biological computer all the human brains that are thinking 69 00:04:42.839 --> 00:04:46.519 and I don't know, we have you back online here, Yeah, I 70 00:04:46.560 --> 00:04:50.519 think somebody. I think somebody hacks us. We must have, we must 71 00:04:50.519 --> 00:04:57.480 have some enemies. Last question on on AI, any new thoughts on regulation 72 00:04:57.879 --> 00:05:00.000 and how we should be stretured. Well, I think we probably do need 73 00:05:00.000 --> 00:05:04.079 some sort of regulatory authority to look at the safety of AI, just as 74 00:05:04.120 --> 00:05:11.800 we have regulatory authorities and other arenas to oversee aircraft and the safety of aircraft 75 00:05:11.839 --> 00:05:15.040 and cars and other things, you know, medication. So another rat which 76 00:05:15.079 --> 00:05:20.680 AI is progressing as fast as faster than probably any regulatory agency can keep up 77 00:05:20.680 --> 00:05:25.279 with. But I do have a comment on what I think is very important 78 00:05:25.360 --> 00:05:29.319 for achieving safe AI, which is that it's very important to train the AI 79 00:05:29.399 --> 00:05:32.319 to be as truthful as possible and not to Yeah, just to be as 80 00:05:32.360 --> 00:05:35.800 truthful as possible. I think you can get some very dangerous things when you 81 00:05:35.959 --> 00:05:40.399 program an AI to be politically correct. I think that things that may seem 82 00:05:40.560 --> 00:05:46.079 relatively innocuous now but will not be so in the future if AI has immense 83 00:05:46.120 --> 00:05:49.600 power. You can take the Google Gemini example, where it refused to publish 84 00:05:50.120 --> 00:05:55.800 to produce a picture of George Washington as a white man, and any in 85 00:05:55.839 --> 00:06:01.240 fact, any historical figure would automatically be made diverse because it's been programmed to 86 00:06:01.319 --> 00:06:05.240 insist on diversity, which sounds, you know, perhaps okay at first, 87 00:06:05.399 --> 00:06:10.399 but not if the AI has so much power that it can actually enforce diversity 88 00:06:10.480 --> 00:06:13.360 and decide there's too many of one kind of people or too many of one 89 00:06:13.399 --> 00:06:17.680 sex and kill off, just just kill off enough until the diversity number is 90 00:06:17.959 --> 00:06:20.959 what is programmed to believe is correct. But don't you think this will be 91 00:06:21.000 --> 00:06:25.319 sorted out in the next version? No, where is China? Where is 92 00:06:25.399 --> 00:06:30.600 China relative to the next question? Now they'll make it more subtle and less 93 00:06:30.639 --> 00:06:32.040 obvious, but it will still be there. Okay, well, we'll see, 94 00:06:32.040 --> 00:06:34.160 But where where is China? Where do you? Where is China in 95 00:06:34.199 --> 00:06:38.639 a relative to the US? I don't know exactly what China is, except 96 00:06:38.720 --> 00:06:42.839 there are a lot of very smart people in China, and they it won't 97 00:06:42.879 --> 00:06:46.120 be They won't be far behind the rest of the world or far behind the 98 00:06:46.240 --> 00:06:49.519 US. I mean, the air right now is very concentrated in San Francisco, 99 00:06:49.600 --> 00:06:53.560 in London, and then you know, there's there's a lot happening in 100 00:06:53.839 --> 00:06:57.720 China. But I don't have insight into what they're doing, except that I'm 101 00:06:57.759 --> 00:07:00.680 confident they will not be far behind what it's developed in the West. So 102 00:07:00.920 --> 00:07:04.079 but but but mark my words, if we do not program an AI to 103 00:07:04.079 --> 00:07:08.879 be as truthful as possible, that that is where it will go awry. 104 00:07:09.079 --> 00:07:13.399 That is where the danger lies. Moving moving tach here, moving to tesla. 105 00:07:13.720 --> 00:07:16.879 Is the EB conversion now going slower than you had expected? Just where 106 00:07:17.000 --> 00:07:21.279 is the speed of commercional relative to your expectations. I think it's going quite 107 00:07:21.279 --> 00:07:26.240 fast actually, especially Noway, absolutely well, it's pretty much all there is 108 00:07:26.240 --> 00:07:29.160 is your teslas. Yeah, there's a lot of teslas noise. It's crazy. 109 00:07:29.480 --> 00:07:31.560 I'd once again like to think no way for the support of electric vehicles 110 00:07:31.600 --> 00:07:38.199 so much appreciate it. So I think it's we will that electric, that 111 00:07:38.279 --> 00:07:41.079 all vehicles will go fully electric. It's only a matter of time. That 112 00:07:41.120 --> 00:07:46.600 includes aircraft ultimately and both obviously trains. The only thing that is ironically difficult 113 00:07:46.639 --> 00:07:49.079 to where you can't really make it electric is rockets because you need you can't 114 00:07:49.079 --> 00:07:54.199 get away from having to expel mass sort of Newton instead law. But all 115 00:07:54.279 --> 00:07:57.920 cars will be electric only amount of time. And we'll look back on combustion 116 00:07:58.040 --> 00:08:01.399 cars in the same way that we look back on steam engines that it was 117 00:08:01.879 --> 00:08:05.319 it was inevitable that there would be internal combustion cars, and it's just as 118 00:08:05.360 --> 00:08:09.879 inevitable that all cars will go electric. And there will be some ev you 119 00:08:09.879 --> 00:08:11.319 know, so like it's going to be a completely straight up line. There 120 00:08:11.319 --> 00:08:16.000 will be some ebb and flow in how how far electric cars go. But 121 00:08:16.480 --> 00:08:22.040 that but the ultimate victory of electric cars is inevitable, and I think the 122 00:08:22.079 --> 00:08:24.879 sooner we get there the better. How do you see the Chinese competition here? 123 00:08:24.920 --> 00:08:28.120 Now, we generally find that the companies in China the most competitive in 124 00:08:28.120 --> 00:08:31.039 the world, and to me, in electric vehicles or cars in general, 125 00:08:31.200 --> 00:08:35.000 the Chinese car companies by far the most competitive. That's where we find the 126 00:08:35.000 --> 00:08:39.080 most toughest, toughest competitive competitive challenges. That they make great cars and they 127 00:08:39.159 --> 00:08:43.240 work very hard so when you write in one of the Chinese cars, what 128 00:08:43.240 --> 00:08:43.799 do you think? I mean, you're an engineer, you know, what 129 00:08:43.799 --> 00:08:46.159 about it? What do you what do you think? I haven't written it. 130 00:08:46.240 --> 00:08:50.519 I have not ridden in one lately. But because they're not all available 131 00:08:50.559 --> 00:08:52.279 here, you know, in the US, very very few are available in 132 00:08:52.279 --> 00:08:56.360 the US. Some are available in Europe, but from what my team tells 133 00:08:56.399 --> 00:08:58.600 me, they are very good. Oh, you are moving into India hair 134 00:08:58.600 --> 00:09:03.320 as well TESL protection. What are your what are your faults there moving into 135 00:09:03.360 --> 00:09:05.279 a to India? Oh, Indieah, yeah, yeah, I think it's 136 00:09:05.440 --> 00:09:09.200 you know, in India sort of. India is now the most populous country 137 00:09:09.240 --> 00:09:13.399 in the world, you know, the biggest population, and I think we 138 00:09:13.639 --> 00:09:16.200 India just should have electric cars, just like every other country has electric cars. 139 00:09:16.240 --> 00:09:22.559 And yeah, so it to natural progression to provide electric TESL electric vehicles 140 00:09:22.600 --> 00:09:28.120 in India. Moving moving out out in space? What what would it take 141 00:09:28.200 --> 00:09:31.240 to be self sufficient at Mars? To be selfufficient on Mars, it's really 142 00:09:31.240 --> 00:09:35.919 about the total tonnage that is delivered to the surface of Mars. So you 143 00:09:35.960 --> 00:09:39.519 can say, like I think it's probably on the order of a million tons, 144 00:09:39.559 --> 00:09:43.399 maybe maybe more. But somewhere between probably a million tons and ten million 145 00:09:43.440 --> 00:09:46.720 tons are needed to make Mars self sufficient. And how many rockets is that? 146 00:09:46.240 --> 00:09:48.559 Well, I gave a presentation on this recently. If you've looked at 147 00:09:48.559 --> 00:09:52.360 my recent space X talk. If you if you have really if you have 148 00:09:52.360 --> 00:09:54.960 one hundred tons per flight, you need ten thousand flights you get to a 149 00:09:54.960 --> 00:09:58.639 million a million tons, and that's one hundred times landed to the surface of 150 00:09:58.679 --> 00:10:01.120 Mars. So in order to get a hundred land to the surface of Mars, 151 00:10:01.159 --> 00:10:03.000 you need five hundred times five times that number in Earth orbit. So 152 00:10:03.279 --> 00:10:09.000 we do a lot of orbital refilling, so launching sort of rockets tanker ships 153 00:10:09.000 --> 00:10:13.120 over and over again. That that would replenish the propellant of the ships that 154 00:10:13.159 --> 00:10:16.080 would go to Mars. And then you'd need it roughly a lolder of ten 155 00:10:16.159 --> 00:10:20.720 thousand of them to get to a million tons. And but we we plan 156 00:10:20.840 --> 00:10:24.080 to do that that we're that's uh, that's what we think we can get 157 00:10:24.120 --> 00:10:26.000 that done within twenty years. Really, so and when do you think so, 158 00:10:26.039 --> 00:10:30.240 when do you think will be there for the first time, first starship 159 00:10:30.279 --> 00:10:33.480 that will land on Mars, which obviously will not have people at first. 160 00:10:33.639 --> 00:10:37.679 I think it's probably within about five years, and then it would probably launch 161 00:10:37.720 --> 00:10:41.679 several ships and just confirm that they can land okay on Mars. We'll also 162 00:10:41.720 --> 00:10:46.000 be doing the Moon simultaneously with that, So I got taking I think, 163 00:10:46.000 --> 00:10:48.200 well, I think we'll get people back to the Moon I should say within 164 00:10:48.240 --> 00:10:50.840 five years, and we'll get uncrewed ships landed on Mars within five years, 165 00:10:50.919 --> 00:10:56.200 and then we'll be building up the production rate and approving the design of the 166 00:10:56.240 --> 00:10:58.679 booster and the ship. So the in the first people on Mars I think 167 00:10:58.759 --> 00:11:03.279 with in seven years or so seven to nine years, and from there we 168 00:11:03.360 --> 00:11:07.639 need to rapidly increase. We need massive numbers of ships going and Earth in 169 00:11:07.679 --> 00:11:13.679 Mars only are in the same quadrants of the Solar System roughly for six months 170 00:11:13.679 --> 00:11:18.279 every two years, or at least, it's only possible to really transfer efficiently 171 00:11:18.360 --> 00:11:20.200 from Earth to Mars. I say every six months, but really there's about 172 00:11:20.360 --> 00:11:24.279 a couple of months where it's ideal every twenty six months, so every two 173 00:11:24.360 --> 00:11:28.279 years that you would see a basic a fleet to part Mars. I think 174 00:11:28.320 --> 00:11:31.240 we're quite a spectacular thing to see a thousand ships depart from Mars all at 175 00:11:31.279 --> 00:11:35.559 once, like battlestock Galactica, a kind of new technology we need before will 176 00:11:35.639 --> 00:11:39.559 be self sufficient there. Actually, I think we have all the tech wordy, 177 00:11:39.600 --> 00:11:41.240 you know, all the technology that's necessary for that. It just needs 178 00:11:41.240 --> 00:11:45.399 to we just need to build. No new physics is needed for this. 179 00:11:46.360 --> 00:11:50.799 Why is it so important for you? I think it's important for consciousness in 180 00:11:50.840 --> 00:12:01.399 general. So if wish to maximize the lifespan of consciousness, then being a 181 00:12:01.440 --> 00:12:07.279 multiplanet species will result in a much longer existence of consciousness consciousness than if we 182 00:12:07.320 --> 00:12:11.159 are on one planet. If you're on one planet, we're simply finding our 183 00:12:11.240 --> 00:12:16.080 time until there's eventually a calamity. It could be soon, it could be 184 00:12:16.120 --> 00:12:18.320 a long time, but eventually something will happen. It could be a global 185 00:12:18.320 --> 00:12:26.000 throm in nuclear war. It could be simply that civilization really subsides. Our 186 00:12:26.000 --> 00:12:28.039 civilization may not die with a bag, It may die with a whimper, 187 00:12:30.000 --> 00:12:35.000 just just gradually falling into obsolescence. But if we're a multiplanet species, then 188 00:12:35.039 --> 00:12:41.159 we've got two planets and yeah, and they can support each other, and 189 00:12:41.200 --> 00:12:45.399 we can go beyond two planets, ultimately to the moons of Jupiter, to 190 00:12:45.440 --> 00:12:52.440 the to the beyond, to the outer parts of the Solar System and ultimately 191 00:12:52.519 --> 00:12:58.320 to other star systems. So this tiny, this tiny candle of consciousness that 192 00:12:58.360 --> 00:13:05.799 we have in this bottle starkness can be extended and amplify it, and we're 193 00:13:05.840 --> 00:13:11.440 just far more likely to survive as for full consciousness to survive. If we 194 00:13:11.480 --> 00:13:16.639 are a multiplanet species, you don't think it'd be better to use all these 195 00:13:16.639 --> 00:13:22.720 resources and trying to sort out Earth. Well, just to put this into 196 00:13:22.720 --> 00:13:26.480 perspective, the amount of resources I'm talking about for making life multiplanetary would be 197 00:13:26.519 --> 00:13:30.879 less than one percent of all resources on Earth, So I really can think 198 00:13:30.879 --> 00:13:33.679 of it as resource allocation. Do you think it's worth spending half a percent 199 00:13:33.759 --> 00:13:41.720 of Earth resources to ensure that we have redundancy and consciousness and that we extend 200 00:13:41.759 --> 00:13:46.639 consciousness beyond Mars to other planets, to Mars and other planets and ultimately other 201 00:13:48.200 --> 00:13:52.200 star systems. And then also take into account the fact that there are certain 202 00:13:52.240 --> 00:13:56.679 inevitable that there's certain things we simply cannot avoid on Earth, Like is it 203 00:13:56.679 --> 00:14:00.639 within your power of mind to stop World War three? I don't think so. 204 00:14:01.440 --> 00:14:05.840 If it happens, and if we have global dominical warfare, our technology 205 00:14:05.879 --> 00:14:11.840 level will drop to the stone age, and we may never spive. And 206 00:14:11.879 --> 00:14:16.799 then there are we maybe get maybe it hit like by a comet like the 207 00:14:16.799 --> 00:14:20.919 dinosaurs. And you know, if the dinosaurs had spaceships, they they'll probably 208 00:14:20.960 --> 00:14:26.440 still be around. So and then if if you wait long enough the Earth, 209 00:14:26.159 --> 00:14:31.720 that the Sun will continue to expand and eventually engulf Earth and destroy it 210 00:14:31.440 --> 00:14:35.840 and destroy all life. So just give it amount in a certain amount of 211 00:14:35.879 --> 00:14:39.759 time, no matter what you do on Earth, no matter how careful you 212 00:14:39.799 --> 00:14:43.000 are, Earth well, life of all life on Earth will die. That 213 00:14:43.080 --> 00:14:52.399 it will happen is a certainty. Gloom is is now, so it's slightly 214 00:14:52.759 --> 00:14:58.799 less gloomy oat X Twitter, Yeah, what is your vision now? What 215 00:14:58.840 --> 00:15:05.639 do you how do you see the visional mix? I go access to be 216 00:15:05.720 --> 00:15:11.320 the best source of truth on the Internet, and I think we're bringing a 217 00:15:11.360 --> 00:15:16.600 good, good progress there. I mean that's we're gonna be like I'll call 218 00:15:16.720 --> 00:15:18.559 everything up like an if anything you want to do you can do on the 219 00:15:18.639 --> 00:15:26.000 X platform, whether it's text, audio, video, payments, financial stuff, 220 00:15:28.320 --> 00:15:33.960 communications of all kinds. And then but then also where there's publicly to 221 00:15:33.000 --> 00:15:37.519 seven eighty information is to be the best source of truth and I think it 222 00:15:37.679 --> 00:15:39.840 I think it already is that. And people may say, oh, there's 223 00:15:39.879 --> 00:15:43.679 some piece of misinformation, disinformation. I said, yes, but look look 224 00:15:43.720 --> 00:15:48.039 at the replies, the replies correct that misinformation, and look at community notes 225 00:15:48.200 --> 00:15:50.399 and the and how good the batting average of community notes is. It's extremely 226 00:15:50.440 --> 00:15:56.279 good. It's by far the best fact checking system on the internet. So 227 00:15:56.519 --> 00:15:58.799 and and and a lot of people still labor under the illusion that the can 228 00:16:00.120 --> 00:16:03.639 that the legacy newspapers that they read are actually true. There's so much nonsense 229 00:16:03.679 --> 00:16:08.240 in them. I mean, how many times when do you read an article 230 00:16:08.240 --> 00:16:14.080 in newspaper where you know the circumstances of what that article is, and how 231 00:16:14.120 --> 00:16:18.159 often is it spot on? No, of course it's normally it's not. 232 00:16:18.279 --> 00:16:22.559 Of course we will know it's normally wrong. But how do you get how 233 00:16:22.559 --> 00:16:25.679 do you look at the situation now, for instance, with with Russia. 234 00:16:26.360 --> 00:16:29.399 You know the work Russia does in Germany with fake accounts on, it's pretty 235 00:16:30.080 --> 00:16:34.840 pretty huge activity, right, I mean, we don't see a lot of 236 00:16:36.120 --> 00:16:45.320 Russian activity, to be frank on the system, so we see very little. 237 00:16:47.200 --> 00:16:48.679 We do we do you see a lot of a lot of terms to 238 00:16:48.679 --> 00:16:52.679 influence things, but they seem to be coming from from the west, not 239 00:16:52.679 --> 00:16:59.840 from from Russia. Right, what about what about things like the latest developments 240 00:16:59.879 --> 00:17:11.240 in in Brazil and so yeah, so the uh we kept getting these demands 241 00:17:11.279 --> 00:17:21.480 from uh this uh judge Alexander that's his that's his name on Twitter at Alexander, 242 00:17:22.839 --> 00:17:29.960 and there would be to suspend accounts immediately. We're given typically two hours 243 00:17:30.000 --> 00:17:36.240 to suspend an account or face massive fines. And the final storye we were 244 00:17:36.240 --> 00:17:42.720 being given demands to suspend sitting sitting members of the parliament and major journalists and 245 00:17:42.799 --> 00:17:48.640 moreover, we could not tell them that this was at the behest of zider 246 00:17:49.000 --> 00:17:52.400 morales. We had to pretend that it was due to our rules of service, 247 00:17:53.279 --> 00:17:56.319 and that was the final straw. And we said, no, mm, 248 00:17:57.599 --> 00:18:03.319 now, when you when you bought Twitter now renamed x, did you 249 00:18:03.319 --> 00:18:07.960 expect that you would end up in these type of situations Always it's all unexpected. 250 00:18:07.119 --> 00:18:11.640 I knew it wouldn't be just a total bed of roses, you know, 251 00:18:12.200 --> 00:18:21.279 and it's talking. I wouldn't. No, I mean I thought it 252 00:18:21.279 --> 00:18:30.000 would be since Look, we're just like rigorously trying to pursue the the the 253 00:18:30.359 --> 00:18:33.680 goal of being the most accurate and truthful place in the Internet. And that 254 00:18:33.680 --> 00:18:36.759 that doesn't mean that what is said is always true or accurate, but it's 255 00:18:36.880 --> 00:18:40.279 it is perhaps another way to frame it is as the least inaccurate place on 256 00:18:40.319 --> 00:18:42.640 the Internet. Do you do you secretly? Do you secretly think this is 257 00:18:42.640 --> 00:18:45.960 a bit fun? It's fun? Yeah, Yeah, it's fun at times. 258 00:18:45.000 --> 00:18:51.519 It's stressful at times, and it's fun at times. But in overall, 259 00:18:51.559 --> 00:18:55.400 we're trying to serve the people of Earth and and and this is sort 260 00:18:55.400 --> 00:18:59.400 of an est sort of maybe an estratarkic way of viewing it, but to 261 00:18:59.480 --> 00:19:03.200 try to be of like the group consciousness of Earth. So you can think 262 00:19:03.240 --> 00:19:07.599 of like if each person is like a neuron contributing to like the collective brain 263 00:19:07.640 --> 00:19:11.319 of Earth, and you want to try to minimize the noise and maximize the 264 00:19:11.319 --> 00:19:17.960 signal of every neuron that's connected to the X network. That that's basically what 265 00:19:17.960 --> 00:19:22.240 what is what is the collective will of humanity? And and how and and 266 00:19:22.240 --> 00:19:27.519 and how to Yeah, just serve the collective will of humanity and so serve 267 00:19:27.559 --> 00:19:32.759 the greater good. That that's our goal. Now, you know, there's 268 00:19:32.799 --> 00:19:37.400 there's definitely going to be people who want to manipulate that information and so ratify 269 00:19:37.519 --> 00:19:41.799 that and try to have a you know, be the most accurate place as 270 00:19:41.799 --> 00:19:45.240 far as the best of our ability, and have it be kind of a 271 00:19:45.279 --> 00:19:51.960 marketplace of ideas where people can propose ideas and you know, debate them. 272 00:19:52.440 --> 00:19:59.559 And I think so far it's working recently well in that regard. Now, 273 00:19:59.559 --> 00:20:03.079 people that don't like the truth will not like X or if they want to 274 00:20:03.079 --> 00:20:06.880 manipulate things, they will not like it. But only, but only a 275 00:20:06.920 --> 00:20:14.400 few years ago you were you were a guy producing electric vehicles. Now you 276 00:20:14.559 --> 00:20:18.160 are, you know, through Starlink, you've had some you know, I 277 00:20:18.200 --> 00:20:25.640 mean some big impact in Ukraine with Twitter, you are kind of into some 278 00:20:25.680 --> 00:20:30.799 issues in you know, Brazil, India, Turkey. You know, you're 279 00:20:30.839 --> 00:20:36.119 becoming like a real geopolitical force and a really important one. How do you 280 00:20:36.319 --> 00:20:38.160 how do you look at that? Well, like I said, I'm really 281 00:20:38.480 --> 00:20:42.119 I'm trying to take the set of actions that maximize the public The future is 282 00:20:42.119 --> 00:20:49.000 good. I mean, we have to keep civilization going onward and upward as 283 00:20:49.079 --> 00:20:55.680 much as possible, and and trying to minimize the civilizational threats that occur. 284 00:20:56.240 --> 00:21:00.359 Like you know, we we can't get to Mars if somevilization collapses. It's 285 00:21:00.359 --> 00:21:07.400 not gonna happen. So you know, we've got to we've got to keep 286 00:21:10.599 --> 00:21:15.440 keep civilization going. And I think we should view our civilization as being much 287 00:21:15.440 --> 00:21:19.680 more fragile than we think. We kind of take for granted it's always going 288 00:21:19.720 --> 00:21:23.279 to be there. But actually, if you study history, you realize that 289 00:21:23.279 --> 00:21:29.440 there are rise. You know, there's a rise in full civilizations. I 290 00:21:29.440 --> 00:21:34.279 mean, I was reading in depth about the ancient Samerians, who were arguably 291 00:21:34.359 --> 00:21:38.200 the first civilization if you call civilization like writing and stuff. You know, 292 00:21:38.240 --> 00:21:47.200 they're the first to develop writing, and but eventually they died out and they 293 00:21:47.200 --> 00:21:51.759 were gone, so and then nobody could read the writing at all, and 294 00:21:51.799 --> 00:21:56.200 they're just faded out as a civilization. But they're pretty impressive in their time. 295 00:21:57.200 --> 00:22:03.799 And the ancient Egyptians the same thing thing, and you know, one 296 00:22:03.839 --> 00:22:08.839 sort of one after another. Ancient Greek had as reset a stay. China 297 00:22:08.839 --> 00:22:15.079 and India had will have incredibly impressive populations, but there's been ebbs and flows 298 00:22:15.079 --> 00:22:21.680 in the China and Indian civilizations over the the AONs, you know, the 299 00:22:21.720 --> 00:22:27.279 Blendia as well. So you know, I guess I'm just trying to take 300 00:22:27.319 --> 00:22:36.839 this instead of just the steps that increase, uh, the scope and scale 301 00:22:36.880 --> 00:22:40.400 of consciousness. That's that's what I'm trying to do. It's not like it's 302 00:22:40.440 --> 00:22:42.880 not that I'm trying to have a put a political thumb on the scale or 303 00:22:42.880 --> 00:22:48.799 anything like that. But I think I'm trying to have the political will go 304 00:22:48.880 --> 00:22:56.240 where the people wanted to go. You know, you mentioned some some really 305 00:22:56.279 --> 00:22:59.759 smart people here and kind of just moving back a bit here into corporate culture. 306 00:22:59.799 --> 00:23:03.599 Now, you manage a lot of geniuses in your in your companies. 307 00:23:03.680 --> 00:23:11.559 What is the key to manage really smart people? Do you think? I 308 00:23:11.559 --> 00:23:18.000 don't. I don't think I managed smart people. They manage themselves. I 309 00:23:18.000 --> 00:23:22.759 think, well, I guess with really smart people, you know, I 310 00:23:22.759 --> 00:23:26.680 don't really think of it like managing them. I think that if somebody is 311 00:23:26.880 --> 00:23:30.319 very smart and talented, they can go anywhere and do anything anytime. Like 312 00:23:30.720 --> 00:23:33.799 they they don't have to work with me, They could go anywhere. So 313 00:23:33.880 --> 00:23:37.480 I really just say, like, look, this is the goal we're after, 314 00:23:37.279 --> 00:23:40.880 and this is what we're trying to achieve, and do you agree with 315 00:23:40.880 --> 00:23:44.000 this goal? And if you do, then let's try to get it done. 316 00:23:45.680 --> 00:23:51.599 And you know, provide my opinion along the way. But it's very 317 00:23:51.680 --> 00:23:56.960 rare for me to actually sort of insist on a particular thing once in a 318 00:23:56.000 --> 00:23:59.279 while, I say, look, guys, just got to trust me on 319 00:24:00.079 --> 00:24:03.960 this one. We've got to do this thing, and if it turns out 320 00:24:03.000 --> 00:24:07.039 to be a bad decision, you can all hold that against me in the 321 00:24:07.039 --> 00:24:11.440 future. But you have an incredible life for deten right. I mean, 322 00:24:11.000 --> 00:24:15.880 when we read the Isaac book, it's pretty clear that you mean you really 323 00:24:17.839 --> 00:24:18.920 are deep into detail. I know what you talk about. So how do 324 00:24:18.920 --> 00:24:25.039 you how do you balance this kind of micromanagement of some areas and then delegate 325 00:24:25.079 --> 00:24:33.960 others. I wouldn't I wouldn't call it micro management. It's just insisting on 326 00:24:34.039 --> 00:24:40.279 attention to detail. That's if you're trying to make a perfect product, you 327 00:24:40.359 --> 00:24:48.599 must have attention to attention to details essential. And I haven't actually read the 328 00:24:48.599 --> 00:24:51.759 Isaacson book. You should. It's very good. Actually, I love it 329 00:24:52.880 --> 00:24:56.640 well. I asked authorizing and if I should read it, and he said 330 00:24:56.640 --> 00:25:07.200 I shouldn't. Okay, So okay, well, I'll ask you some questions 331 00:25:07.200 --> 00:25:08.319 from the book. Then they do you talk? He talks about, you 332 00:25:08.359 --> 00:25:14.599 know, you're the kind of a hardcore and ultra hardcore culture. What is 333 00:25:14.640 --> 00:25:18.640 an ultra hardcore culture. I guess it's work. I mean it's working culture, 334 00:25:18.680 --> 00:25:22.079 right, I mean how how I mean ultra hard work? How hard 335 00:25:22.119 --> 00:25:27.960 is that? Well, when things get really intense, you're basically just working 336 00:25:29.039 --> 00:25:33.279 every waking hour. And how and how long can you do that for? 337 00:25:34.680 --> 00:25:42.640 I've done that for well, continuously for sometimes like a few years. What 338 00:25:42.680 --> 00:25:47.759 does it? What does it do to you? It? Really it's pain 339 00:25:49.000 --> 00:25:52.960 and every waking out. Maybe it is an exaggeration because there are a few 340 00:25:52.960 --> 00:26:00.759 hours obviously with friends and family and critical other things. Uh, but one 341 00:26:00.839 --> 00:26:03.160 hundred hour weeks would be I've done many, many stretches of one hundred hour 342 00:26:03.160 --> 00:26:08.720 weeks, like true hundred hour weeks where roughly six hours per day is sleeping. 343 00:26:11.720 --> 00:26:14.680 I would not recommend that this is not that's for emergencies, you know, 344 00:26:14.799 --> 00:26:21.759 it's not all the time. You know, during very difficult times at 345 00:26:21.759 --> 00:26:25.960 Tesla, I've had to do that, and sometimes at the beginning of my 346 00:26:26.039 --> 00:26:29.160 earliest startups, I did that where I just wouldn't leave the office. I 347 00:26:29.160 --> 00:26:34.319 would a sleep under my desk and just work seven days a week. Sometimes 348 00:26:34.359 --> 00:26:41.640 it's necessary for success or to avoid failure. But but do you you do 349 00:26:41.680 --> 00:26:47.160 you enjoy being in this crisis mode. No, I don't. It sucked. 350 00:26:52.680 --> 00:26:57.200 No, I don't want to be there. That's pain m hm. 351 00:26:59.079 --> 00:27:03.319 But sometimes it's that between success and failure. When you make decisions? How 352 00:27:03.359 --> 00:27:08.559 important is speed? He just gave me an idea, which is I'm going 353 00:27:08.640 --> 00:27:17.240 to invite the judge Alexander Raala to do a spaces and then he can explain 354 00:27:18.000 --> 00:27:23.640 why what I'm doing is bad. And maybe he's right I challenge. I 355 00:27:23.759 --> 00:27:30.440 challenged him to a spaces sounds good, Yeah, but what about speed? 356 00:27:30.480 --> 00:27:33.880 When do you make when you when you make decisions, how important is speed? 357 00:27:33.279 --> 00:27:37.400 And how do you how do you balance analysis with your good field? 358 00:27:37.920 --> 00:27:45.759 I think the the best offense and defense is speed. If you think it's 359 00:27:45.759 --> 00:27:49.039 something like the SR seventy one Blackbird, it really had almost no defenses except 360 00:27:49.160 --> 00:27:53.720 accelerate and it was never shot down even once. Like I think over three 361 00:27:53.759 --> 00:28:00.839 thousand missiles were shot at the SR seventy one Blackbird and none hit and and 362 00:28:00.960 --> 00:28:06.799 really what it did was just go faster. So the the power of speed 363 00:28:06.880 --> 00:28:18.160 is underappreciated as a competitive dimension. Is that why in all space expenses being 364 00:28:18.240 --> 00:28:25.119 so successful? Because you've been mean and lean as an organization and fast. 365 00:28:26.079 --> 00:28:30.160 I think speed is definitely a factor. Now I should say you want to 366 00:28:30.319 --> 00:28:33.920 go in the case of the company, you you need to be a vector, 367 00:28:33.920 --> 00:28:37.599 not a scaler, So it can't be you need to go at high 368 00:28:37.640 --> 00:28:48.640 speed in the right direction. So so and no company is going to be 369 00:28:48.680 --> 00:28:49.599 going in the right direction all the time, so you have to do course 370 00:28:49.640 --> 00:28:57.440 corrections like a guided missile recorded course corrections and uh. But in the case 371 00:28:57.440 --> 00:29:02.880 of SpaceX, it's like, Okay, our goal is to extend humanity beyond 372 00:29:02.920 --> 00:29:07.559 Earth. And we didn't even know how to even frame the question correctly, 373 00:29:07.759 --> 00:29:15.119 like what which knew that that was a general goal. We didn't know what 374 00:29:15.319 --> 00:29:18.640 pealant we would use, or what the raw materials would be, or for 375 00:29:18.680 --> 00:29:22.359 the how would the rocket be built, how would it be designed? What's 376 00:29:22.400 --> 00:29:30.799 actually important? And you know, so for example, going from our Falcon 377 00:29:32.000 --> 00:29:41.680 architecture, which is users refined jet fuel and liquid oxygen in a open cycle 378 00:29:41.880 --> 00:29:51.680 gas generator architecture engine to a two starship, which is a liquid methane like 379 00:29:51.880 --> 00:30:02.839 liquiduction a propellant in a staged combustion, very high pressure engine. That that's 380 00:30:03.200 --> 00:30:06.960 a that's a big architectural change. But we didn't know that we needed to 381 00:30:06.960 --> 00:30:11.400 make that architectural change until we're pretty far down the road, like about halfway. 382 00:30:11.480 --> 00:30:15.079 But because about ten years to figure out that was even the right architecture, 383 00:30:15.799 --> 00:30:18.680 now we're confident it is. I think SpaceX is one of the best 384 00:30:18.680 --> 00:30:22.000 example I know about what we call failing well right, learning from mistakes and 385 00:30:22.039 --> 00:30:26.680 moving on generally, how do you how do you look at mistakes? I 386 00:30:26.680 --> 00:30:29.440 mean, which which ones do you tolerate and which ones don't you tolerate? 387 00:30:29.559 --> 00:30:32.200 I do really think of that way. You know, the first three flights 388 00:30:32.200 --> 00:30:36.880 of SpaceX failed, the fourth one succeeded, and if the fourth one had 389 00:30:36.920 --> 00:30:38.680 not succeeded, we would have gone bankrupt. We would have had no money 390 00:30:38.759 --> 00:30:45.440 left. So it was very close call. But since then SpaceX has done 391 00:30:45.519 --> 00:30:48.559 very well. It's now the Falcon nine. You know, knock on Wood 392 00:30:48.640 --> 00:30:55.000 is the most reliable rocket in the world and launches about every two to three 393 00:30:55.079 --> 00:30:56.880 days. But if you take a game like, for example, Civilization, 394 00:30:57.079 --> 00:31:00.599 it's actually quite a good It tells you how how civilizations are formed. I 395 00:31:00.599 --> 00:31:04.920 remember playing the original Civilization with a technology tree, and how you invent different 396 00:31:04.960 --> 00:31:11.680 things. You'd like invent literacy, and invented democracy and invent gunpowder, all 397 00:31:11.720 --> 00:31:14.160 these things that like, and you start to realize, oh wow, there's 398 00:31:15.240 --> 00:31:18.599 there are stages two technology. You can't actually get to democracy without literacy, 399 00:31:18.720 --> 00:31:26.799 these stages of technology development or stages of ideas that you know, that's a 400 00:31:26.799 --> 00:31:30.279 helpful framework for a company. In recent years, there's a game I play 401 00:31:30.359 --> 00:31:36.440 that was actually developed in Sweden called Polytopia, which is actually quite a good 402 00:31:36.480 --> 00:31:37.680 game. Like a lot of people like playing chess, but I think chess 403 00:31:37.759 --> 00:31:41.039 is not a great There's not a lot of transfer learning from chess to the 404 00:31:41.079 --> 00:31:45.480 real world because in chess, you've got only sixty four squares. It's a 405 00:31:45.519 --> 00:31:48.200 set piece battle, same pieces every time. If there are no terrain differences, 406 00:31:48.880 --> 00:31:52.119 there's no technology tree, there's no fog of war. But say a 407 00:31:52.160 --> 00:31:57.200 game like Polyatopia has all of those things, random train generation, differences in 408 00:31:57.279 --> 00:32:02.359 attacking defense bonuses depending on type of train. You've got sixteen tribes, I 409 00:32:02.359 --> 00:32:07.960 think, each with different abilities. You've got a technology tree that you can 410 00:32:07.039 --> 00:32:09.799 choose to develop in different ways, and you've got, of course fog of 411 00:32:09.839 --> 00:32:14.519 war. So that I think is much more much closer to reality. I 412 00:32:14.559 --> 00:32:19.319 was playing Diablo for a while pretty fun, Diablo as high levels gets very 413 00:32:19.319 --> 00:32:22.359 complicated. You could call it like a spreadsheet with a game attached. And 414 00:32:22.559 --> 00:32:27.079 I briefly got for about a day the world record in this avatar of zero 415 00:32:27.519 --> 00:32:30.640 on a four person team of clearing the hardest level. You know, not 416 00:32:30.720 --> 00:32:37.039 bad for someone who's like fifty three. Basically will be fifty three soon. 417 00:32:38.680 --> 00:32:44.079 There is still some twitch element to it, and it's hard to be because 418 00:32:44.079 --> 00:32:46.079 at games with a twitch element. But yeah, I like if I find 419 00:32:46.079 --> 00:32:49.799 these games interesting. If you can be fully immersed in a game, it's 420 00:32:49.839 --> 00:32:53.079 great. What is a score now? In terms of the union in Sweden 421 00:32:53.079 --> 00:32:57.119 and the collective bargaining, I think the storm has passed on that front. 422 00:32:57.160 --> 00:33:00.960 I think things are reasonably good shape in Sweden. I think things are good 423 00:33:00.519 --> 00:33:04.440 overall. Yeah, I feel pretty good about the future. I mean, 424 00:33:04.799 --> 00:33:07.839 you know there's going to be bumpy quarters from you know, here and there, 425 00:33:07.880 --> 00:33:10.759 but I think the long term future of Tesla is extremely strong. For 426 00:33:10.799 --> 00:33:15.359 example, so I didn't hear the answer because I was out, but we 427 00:33:15.440 --> 00:33:19.400 have covered this with with your chair. But just the last question, what 428 00:33:19.400 --> 00:33:23.799 do you want your legacy to be? I don't I don't mind if my 429 00:33:23.920 --> 00:33:34.240 legacy is accurate or inaccurate, provided that I've I dive feeling that I've done 430 00:33:34.279 --> 00:33:38.160 the right thing for the future of consciousness. So just trying to trying to 431 00:33:38.440 --> 00:33:43.920 have this slide of consciousness lasts as long as possible, and maybe understand more 432 00:33:43.960 --> 00:33:47.440 about the nature of the universe or stimulation or whatever this is. I have 433 00:33:47.480 --> 00:33:55.880 a philosophy of curiosity, which is to understand the they understand the universe, 434 00:33:55.960 --> 00:34:00.480 understand the nature of the universe, or even what questions to ask. Kind 435 00:34:00.519 --> 00:34:04.720 of like that, I would say I would subscribe to the Douglas Adams Hitchhiker's 436 00:34:04.720 --> 00:34:09.360 Guys, to the Galaxy school of philosophy that we're trying to understand what questions 437 00:34:09.360 --> 00:34:15.360 to ask about the answer that is the universe. Okay, I think that's 438 00:34:15.400 --> 00:34:19.000 a good place to end. For sure, the life and life on this 439 00:34:19.679 --> 00:34:23.239 planet would have been a lot more boring without you, and so I'm Graham 440 00:34:23.280 --> 00:34:28.320 glad to spice it up a little totally. All right, good talking