The Exchanges

Q your belt in Austin and it feels like pretty well, you guys have done a great job figuring out where the trouble spots are. Maybe you could talk a little bit about what you learned in the first, I don't know, it's been like three or four months of this so far. What did you, what did you learn in the first three or four months of the Austin experiment?

A Actually, it's, it's gone pretty smoothly. Um, a lot of, a lot of things that we're learning, um, are, uh, just how to manage a fleet, like, cause, cause you've, you've got to write all the fleet management software, right? So, um, and you've, you've got to write the ride hailing software. You've got to write, basically the, the software that Uber has, you've got to write that software. It's just summoning a robot car instead of a car with a driver. Um, so, so a lot of the things we're doing, we're, we're, we're scaling up the number of cars. Um, to say, say, like, what happens if you have a thousand cars? Like, so we'll, we, you know, we think probably we'll have, you know, a thousand cars or more, um, in the Bay Area, uh, by the end of this year, probably, I don't know, 500 or more in the greater Austin area. Um, and, you know, if, if, if, um, you know, you have to, you have to make sure the cars don't all, for example, go to the same supercharger, uh, at the same time. Right. Um, so, uh, or don't all go to the same intersection. Um, there's, there's, it's like, what do these cars do? And then like, sometimes there's a high demand and sometimes there's, there's low demand. What do you do during, during those times? Uh, do you have a car circled the block? Do you have a, try to find a parking space? Um, the, um, and then, you know, sometimes the, like say it's a, it's a, You k…

AI assessment note: “a lot of things that we're learning, um, are, uh, just how to manage a fleet”

Q for this that instead of, you know, having an AT&T account or, and then roaming when you're in the UK or you're in India, it's just, we could have one direct deal with Starlink. It works all over the world eventually, not today, but at some point, is that the end goal? That basically, we don't need a regional carrier. We have a global carrier, and that would be you.

A Uh, that, that would be one of the options. To be clear, we're not going to put the other carriers out of business. They're still going to be around because they, they own a lot of spectrum. So, uh, there's, uh, but, but yes, you'll, you should be able to have a Starlink, uh, like, you know, like you have an AT&T or P-Mobile or Verizon or whatever. You should go, you, you could have a, you know, account with Starlink that, uh, works with your, you know, Starlink, uh, antenna at home with, uh, Wi-Fi as well as on your phone. And, um, yeah, it would be a comprehensive solution for high bandwidth at home and for high bandwidth direct to sell.

AI assessment note: “that would be one of the options. To be clear, we're not going to”

Q Yeah. But, but what about making it to orbit? Do you think you got a shot this time?

A We are doing a new staging technique called hot staging, uh, where, um, You light the upstage, the upstage engines, or the ship engines, while the, um, boost stage is still firing, and, um, this is the, kind of, this is the most efficient way to do stage separation of a rocket going orbit, um, but we did not try that on the last mission, and, and we're trying it on this mission. We think it will be overall better, um, but I, I think this, I think probably about, well, I hope well over a 50% chance of getting to stage separation, um, and maybe a close to 50% chance of getting to orbit if the hot staging, um, the new separation method, ah, is it works? So I'd say maybe it's like a, you know, above, I'd say probably above a 50% chance of getting to orbit this time, whereas previously I said below 50.

AI assessment note: “probably above a 50% chance of getting to orbit this time”

Q Um, you know, Tesla is competitive, but Tesla is competitive because we have an awesome team in China that, uh, you know, so, um, Like do your Tesla China employees work some meaningful percentage more or harder than your Tesla non-China employees? Do you find like it's, Two different companies, basically.

A Well, I mean, I, I think Tesla is somewhat, it tells us sort of pretty far out there in terms of work ethic, uh, anywhere in the world. So, uh, the Tesla work ethic in the U S I think is substantially greater than any other car company or, or any large manufacturing company that I'm aware of. Um, so, you know, Tesla, Tesla does have a, a strong worth of work ethic in the U S. But, but to be totally frank, it, it, it, it, the work that work ethic is exceeded, um, uh, on balance by, uh, the Tesla China team. That, that is, I think, objectively true. So there's not to say there aren't lots of hardworking people at Tesla US. There certainly are. Um, but if you say on average, the, the, the work ethic in China is higher. It's just.

AI assessment note: “the work ethic is exceeded, um, uh, on balance by, uh, the Tesla China team”

Q Where does it break? Is it a materials breakdown or where does it break down economically?

A Well, so, so you can't just, um, Use, uh, normal hydrogen. You know, you, you need to use like deuterium and tritium, like unusual forms of hydrogen. Helium three. Uh, uh, you know, there, um, there are some, uh, some, uh, other types of, of fusion that could be used. Uh, but, um, these are just not, they're not like, there's not a lot of this raw material. It's quite difficult to get the raw material. So first you have to get the raw material. Uh, that's, that's expensive raw material. Um, and then, um, it's not just about generating the, the, the energy you've, you've got to, um, turn that energy into usable electricity. You can't just have a hot thing. Okay. So the hot thing has to translate to usable electricity. So I think you've got, you've got a cost of fuel issue, which is very significant. Uh, you've got, you got a whole bunch of knockdowns from, uh, when you generate the heat to, uh, when you actually convert that into electricity. Um, you've got some very difficult maintenance issues with, with, uh, a fusion reactor. Um, So, uh, and that should be then compared to alternatives, uh, the, the sustainable energy alternatives that I think, uh, are overwhelmingly more competitive are, um, uh, solar energy, wind, uh, geothermal, uh, hydro, uh, some title and energy, but it's really primarily, uh, solar, uh, and, and wind, um, that you can really say like, what, what, why b…

AI assessment note: “you've got a cost of fuel issue, which is very significant.”

Q even more challenging? Because those early days, it was just fighting to survive, nobody knew who you were, you were anonymous, um, and it was really just about the work. And now, let's face it, you're the world's most famous guy, Um, and everybody's watching everything you do, but these companies are also very big. So what's life like for you today? Are you enjoying what you're doing every day?

A Um, well, I mean, it's, it's, it's somewhat of a rollercoaster. Um, so there are like good days and bad days. Um, and there's, there was a crisis issues. Um, and you know, like sort of, you know, knock on wood, like we're not like, uh, Facing, you know, death in the face. Like, like, it's, it's, it's definitely, like, quite stressful when, like, you know, death is, like, trying to eat your face off, and, like, the foam is, like, you know, just getting, and, like, yeah, right there, you know. You know, that's, it's pretty stressful in that situation. Um, I mean, So like, right, you know, both SpaceX and Tesla have, um, you know, significant cash reserves. So like, you know, it's not, we're staring death in the face. We can sort of see it over in the horizon, you know, so I don't want to get complacent or entitled because, um, but, but, but if it's not like just sort of foaming at the mouth and gnashing, trying to eat your face off, uh, on a daily basis, that's, that's certainly we've moved on from that point. Um, and hopefully never, Never return. Um, but, but there are a lot of issues that need to be, it's just like the, if you're a CEO of a company, the chore level is high. And if you don't do your chores, then the company goes to hell. And I hate doing chores, frankly. So, uh, who, who does? Uh, so that's the real, like, there's a whole bunch of sort of, uh, you know, personn…

AI assessment note: “it's somewhat of a rollercoaster. Um, so there are like good days and bad days.”

Q But is there a world where, for example, other people over time that need, you know, some sort of like video use case or image use case could theoretically, you know, you'd say, oh, why not? You know, I have some extra cycles over here, so. We should kind of make you a competitor of NVIDIA. It's not intentionally, per se, but.

A Um, yeah, I mean, the, you know, there's this training and inference, and we do have the, you know, those two projects at Tesla. We've got Dojo, which is the, the training computer, uh, and then, um, you know, our inference chip, which is in every, every car, inference computer. Um, so, and a dojo, we've only had dojo one. Dojo two is, um, you know, should be, we should have dojo two in volume towards the end of next year. Um, and, and that, that, that will be, we think, sort of comparable to, uh, sort of a B 200 type system, a training system. Um, and, um, You know, so there's, I guess there's some potential for, for that to be used as a service. Um, but, but, but, like, you know, Dojo is, is, is just kind of like, I mean, we're, we're, we're, I guess, I, I guess I have, like, some improved confidence in Dojo, um, But I think we won't really know how good Dojo is until probably version three. Like, it usually takes three major iterations on a technology for it to be, to be excellent. Um, and we'll only have the second major iteration next year. Um, the third iteration, I don't know, maybe Late, you know, 26 or something like that.

AI assessment note: “there's I guess there's some potential for for that to be used as a service”

Q And, you know, they say, Elon, life's a negotiation, so at a different price, it might be a totally viable deal, correct?

A I mean, that, I mean, it's not out of the question. Um, but I really would, you know, this is, you know, the more, the more questions I ask, the more I, the more my concerns, uh, grow. Um, so. You know, at the end of the day, acquiring, it has to be fixable, um, and, and fixable, you know, with, with reasonable, reasonable timeframe and without revenues collapsing along the way and all that sort of stuff. Um, and so, uh, you know, I, I really need to see how these things are being calculated and, and it, it can't be some deep mystery that is like more complex than the human soul or something like that. Um, it's gotta be. You know, I think we can apply the scientific method to this and try to figure out what's really going on. And, um, you know, Twitter's revenue is, is, is primarily dependent. I think 70% or some of that order on brand advertising, as opposed to specific purchase advertising. This is a big deal because brand advertising is not, there's not a, there's not a purchase that results from that. So it's basically, you know, how much mind share or like basically if you're a big company, how, how often do they hear your name? Um, it's as opposed to something that where you can directly measure the outcome. Um, so that, that means that they're somewhat going on faith. Um, and if that faith is undermined or, or reduced because of the reality of the situation coming to the…

AI assessment note: “it's not out of the question.”

Q changed with respect to U.S. policy towards China, what it's like in D.C., what it's like in Silicon Valley, and how everyone talks about the relationship with China today. It's pretty crazy how quick things have changed. Um, as a business leader with all these business relationships with China, how do you make decisions and, and how things are changing, and how do you think about where this is headed?

A Sure. Well, I mean, let's just clarify here, um, you know, SpaceX has no, uh, SpaceX and Starlink have no business in China whatsoever. They're not, they're not allowed, you know, SpaceX doesn't launch, uh, China satellites, and Starlink is, uh, abandoned in China. So, to be clear, SpaceX, Starlink, zero business in China. Um, uh, in the case of Tesla, uh, one of our, well, of our four vehicles Is in, uh, China. So, um, you know, it's a, it's a significant, uh, car market, uh, but it is, uh, you know, so, what I'm trying to say is, like, by far the bulk of, of my business interests. If, if I Outside of China. Let's just be clear about that. Um, then, with respect to, now that I said, I, I think I understand China well. I've been there many times. I've met with, uh, senior leadership, um, at many levels of China for, for many years. And so, I, I think I've got a pretty good understanding of, um, at least as an outsider of China. So, and, and Tesla has been very successful domestically in China. So, um, You know, the fundamental thing here is, is really Taiwan. Um, the, uh, China has, well, really since, uh, for like half a century or so, uh, it may be longer at this point, much longer at this point, their policy has been to, um, sort of reunite Taiwan with China. Uh, from this standpoint, you know, it may be just analogous to, like, Hawaii or something like that. Like, an integr…

AI assessment note: “the fundamental thing here is, is really Taiwan.”

Q was people opening many millions of accounts, but, ah, we discovered this very early. How easy was it for you, ah, with the tech team to maybe lower the bot count and all the fake accounts? Maybe you could speak a little bit to that, because people have seemed to think that, gosh, it's, it's a really hard thing to get rid of bots, and it turns out it isn't.

A Um, well, we still have a fair number of bots in the system, um, but The like, I think that the incentive structure, the way Twitter set up previously was this relentless focus on what they called MDAO, which is monetizable daily active users. Although I would say the monetizable part is dubious, but at least things that appeared to be monetized or could be passed off as monetizable daily active users. So this, I created an incentive, an incentive to turn a blind eye to a fake accounts. So if the incentive structure is like, You know, maximize the appearance of monetizable daily active users. Then you're just, it's a strong incentive to pretend that a bot is real and that's what happened. So, um, we were taking a lot of steps to reduce the, the bots and troll situation. Um, so many, um, and, um, I think you're seeing that in the usage. Like, it's not, it's not like relatively rare to have your replies filled with crypto scams.

AI assessment note: “we were taking a lot of steps to reduce the, the bots and troll situation”

Q Pavel in France, and then, of course, we got Brazil with Judge Voldemort. That, that one seems like the one that impacts you the most. Can you, what's the latest on that?

A Well, we, I guess we are trying to figure out, ah, is there some, some reasonable solution in Brazil? Ah, the, you know, the concern, ah, I mean, I want to just make sure that this is framed correctly. Um, And, ah, you know, funny memes aside, ah, the, the, the, the nature of the concern was that, ah, at least at, at X Corp, we had the perception that, ah, we were being asked to do things that violated Brazilian law. So, ah, I, I, obviously we cannot, as an American company, impose American laws and values on, on other countries, um, that, ah, you know, we wouldn't get very far if we did that. Um, but, But we do, you know, think that, uh, if, if a country's law laws are a particular way, and we're being asked to, what we, we think we're being asked to break them, then, and, and be silent about it, then obviously that is no good. So, so I just want to be clear, because this sometimes, sometimes comes across as, uh, Elon's trying to just be a crazy, whatever, billionaire, and, Demand outrageous things from other countries. And, um, you know, while that is true, um, In addition, there are, um, other things, uh, that, that I think are, you know, valid, which is like, we, we, we obviously can't, uh, you know, I think any given thing that we do at X Corp, we've got to be able to explain in the light of day, and, and not feel that it was dishonorable, or, you know, we, we, we did the …

AI assessment note: “we actually are in, uh, sort of discussions with, uh, the, you know, judicial authorities”

Q to act differently now that the risk free rate is probably going to get to five percent. And I'm just curious across all your businesses. So Twitter, yes, but really more importantly, Tesla, SpaceX, are there decisions that you will make differently or not at all, or will make that you wouldn't have made otherwise in this new regime? And how often do you think about that kind of stuff?

A Well, I think it's more like, like, it does seem like we're headed into a recession here, um, in 2023. The magnitude of that recession is debatable, but I think it's at least a, uh, a light to moderate recession. Potentially it's on the order of, uh, 2009. Um, so that's, so I think it's, it's, it's wise to kind of like prepare for the worst, hope for the best, prepare for the worst. Don't get too adventurous, like, like watch out for margin debt. Like, I would really advise people to not have, uh, margin debt, uh, in a volatile stock market. And, uh, you know, from a cash standpoint, keep, keep powder dry. You can get some pretty extreme things happening, um, in a down market. Um, like Brett Johnson, who was a CFO, who is a CFO, CFO of SpaceX was at, um, Broadcom in 2000. And he, he said that, uh, and that's a good company making good products. And he said the, the, from, from peak to trough, I think in less than 12 months, uh, Broadcom went down 97%. So like, even if you had a small margin loan there, you got, you got crushed. Um, it subsequently recovered. And I think, you know, to, to much higher levels, but You know, if there's like mass panic in the stock market, uh, then you've got to be really be careful about margin debt. So, but I mean, this is just as, as we know, there's the economy is cyclic. So, you know, you, and it's somewhat overdue for a recession. And my best …

AI assessment note: “it does seem like we're headed into a recession here, um, in 2023.”

Q money off AI is NVIDIA with the chips. Um, do you have a sense yet of where you think the big applications will be from AI? Is it going to be an enabling self-driving? Is it going to be enabling robots? Is it transforming industries? I mean, it's still, I think, early in terms of where the big business impact is going to be. Do you have a sense yet?

A I, I mean, I think, I think the, the, the, the spending on AI probably runs ahead of, I mean, it does run ahead of the revenue right now. That's, there's no question about that. Um, but the rate of improvement of AI is faster than any technology I've ever seen by far. And, and, and it's, I mean, like the, the, the, for example, a Turing test used to be a thing. Now, you know, your basic, uh, open source, random LLM, you're writing on a frigging Raspberry Pi probably could, uh, you know, beat the Turing test. Um, so there's, I, I, I think actually, like, like, the, the, the good future of AI is one of immense prosperity, where There is an age of abundance. No shortage of goods and services. Everyone can have whatever they want, unless, except for things we artificially define to be scarce, like some special artwork. Um, but, but anything that is a manufactured good or provided service, uh, will, I think, with the advent of AI plus robotics, that the cost of goods and services will be, will Trend to zero. Like, I'm not saying it'd be actually zero, but it'll be, everyone will be able to have anything they want. That's, that's the good future. Of course, you know, in my view, that's probably 80% likely. So look on the bright side. Only 20%. 20% probability of annihilation. Nothing.

AI assessment note: “with the advent of AI plus robotics, that the cost of goods and services”

Q How's the use, how's the usage been? Have you seen good growth since it launched?

A Uh, yeah. Is it very early? It went super viral. Um, so we're, you know, we're seeing, seeing it cited all over the place. Um, but, uh, yeah, it's, and I think we'll see it used more and more, um, as people refer to it, and people will judge for themselves. When you read a Grokipedia article about a subject or a person that you know a lot about, and you see, wow, this is way better than, than Wikipedia. It's, it's, it's more comprehensive. It's way more accurate. Um, it's not, it's, it's neutral instead of biased, then you're going to set, you're going to fold those links around, um, and say that this is actually the better source. Like it's, it will succeed, I think very well, because it is fundamentally a superior product to Wikipedia. It is, it is a better source of information. Um, and we haven't even added, uh, images and video yet. That's going to be awesome. Yeah, we're gonna add a lot of video. Um, so, uh, using Grok Imagine to create videos, um, And, uh, so if you're trying to explain something, um, Grok Imagine can take the text from Grokpedia and then generate a video, uh, an explanatory video. So if you're trying to understand anything from how to tie a bow tie to, you know, how do some chemical reactions work or, you know, um, really anything, um, dietary things, medical things, um, we could, uh, Grok, well, you can just go on and see the video. Of, of how it works…

AI assessment note: “It went super viral. Um, so we're, you know, we're seeing, seeing it cited”

Q now gone through the intellectual exercise of studying all this. Um, if you're looking at the two choices now, fixing Twitter given all these problems, and maybe just starting your own version, which one are you leaning towards? Because it, I have watched you build a couple of companies, and the products have turned out pretty good. So is it easier for someone like you to just start from scratch?

A I mean, I mean, it's certainly the, my, my default inclination is to start things from scratch. Uh, I mean, I'm not really, I don't buy things. Like there's still this sort of, you know, uh, um, yeah, like, like SpaceX was started from scratch. You know, in, in the case of, of Tesla, uh, you know, it was like five people. There's still this guy, Mark Eberhard, who's the worst guy I've ever worked with, who tries to claim like sold credit essentially for creating a Tesla. And if he's so damn great, why didn't he just go, you know, create another car company when he was fired? Um, but anyway, um, so, uh, I mean, that's a pretty good story.

AI assessment note: “my default inclination is to start things from scratch”

Q lot of proclamations that there's going to be environmental and FAA and all these other sorts. The recovery back to the launch pad again was incredibly fast. How did you get back so fast? Not just technically and work wise, but just like regulatory clearance wise, because they said there were going to be all these questions and reviews and so on. How did, how did you guys manage that?

A Uh, well, there were a lot of questions and reviews. We got through them all, um, and credit to the SpaceX team. They worked incredibly hard and they, uh, got the next driven booster tested and on the pad and, and flown and, um, yeah, huge credit to the SpaceX team. Very proud of them for doing, doing such a job, a great job recovering. Um, I mean, creating a fully reusable orbital rocket is one of the hottest engineering problems ever. It's certainly, you know, a candidate for most difficult engineering project ever. You know, it's on the podium at least. Um, so it's a, that, that's been the goal of SpaceX from the beginning, from 2002. Um, and here we are 23 years later, so it's, it's a long journey. And, um, With, with a super talent, like by far the, I think the most talented group of rocket engineers that's ever been assembled. Um, and, uh, and we're finally next year, I think we'll be able to achieve full reusability.

AI assessment note: “well, there were a lot of questions and reviews. We got through them all”