Elon Musk - "In 36 months, the cheapest place to put AI will be space”

So are there really three hours of questions or are you fucking serious? Yeah. You know what I'm talking about, Elon? Only one. I mean, it's the most interesting point. All the storylines are kind of converging right now. So we'll see how much it's almost like a planet. Exactly. Well, we're good. That would never do such a thing. So as you know better than anybody else, the total cost of ownership of a data center, only 10 to 15% is energy. And that's a part you're presumably saving by moving this into space. Most of it's the GPUs. If they're in space, it's harder to service them, or you can't service them. And so the depreciation cycle goes down on them. So it's just way more expensive to have the GPUs in space. Presumably. What's the reason to put them in space? Well, the availability of energy is the issue. So, I mean, if you look at electrical output outside of China, for everywhere outside of China, it's more or less flat. It's very, you know, maybe slightly increased, but for pretty much flat. China has a rapid increase in electrical output. But if you're putting data centers anywhere except China, where are you going to get your electricity? Especially as you scale. The output of chips is growing pretty much exponentially, but the output of electricity is flat. So how are you going to tell them it chips on? You know, magical power sources, magical electricity ferries? I mean, you're famous, you're famous, you have a fan of solar. One terawatt of solar power. So with a 25% capacity factor, like four terawatts of solar panels, it's like one percent of the land area of the United States. And it's like, you were in the singularity when we got one terawatt of data centers, right? So what are you running out of? How far into the singularity are you going? You tell me. Yeah, exactly. I think we'll find we're in the singularity and like, okay, we'll still go along where to go. But is this like a, is the plan to like, put it in the space after we've covered Nevada and solar panels? I think it's pretty hard to cover the solar panels. You have to get like permits from like the approach for trying to get the purpose for that. So the space is really, it's really a regulatory play. It's like harder to build on land than in distance space. It's harder to scale on ground than it is to scale in space. But also the, you're going to get about five times the effectiveness of solar panels in space versus the ground. And you don't need batteries. I almost wore my other shirt, which says it's always sunny in space. Which it is. So because you don't have a day night cycle or seasonality clouds or an atmosphere in space, because at the atmosphere alone, we're still at about 30% less of energy. So you're going to get any given solar panels going to do about five times more powering space than on the ground. And you avoid the cost of having batteries to carry you through the night. So it's actually much cheaper to do in space. And my prediction is that it will be by far the cheapest place to put AI will be space in 36 months or less, maybe 30 months. 30 months? Less than 30 months. How do you service GPUs as they fail? Which happens quite often in training? Actually, it depends on how recent the GPUs are that are right. I mean, at this point, we found our GPUs to be quite reliable. There's infant mortality, which you can obviously iron out on the ground. So you can just run them on the ground and confirm that you don't have infant mortality with the GPUs. But once they start working, there are actual reliability. And once they start working, and you pass the initial, you know, debug cycle of Nvidia or whatever, whoever is making the trips, could be Tesla, Tesla AI, six trips or something like that, or it could be, you know, a TPUs or trains or whatever. The reliability is actually -- they're quite reliable, they're not going to be able to pass certain points. So I don't think -- I don't think you're going to need the servicing thing as an issue. But you can walk my words. In 36 months, but probably closer to 30 months, the most economically compelling place to put AI will be space. And then it will get -- they'll then be like ridiculously better to be in space. And then the scaling -- the only place you can really scale is space. You know, what is your start thinking in terms of what percentage of the sun's power are you harnessing? You realize you have to go to space. You can't scale very, very much on Earth. But maybe very much -- to be clear, you're talking like terawatts. Yeah. Well, all of the United States currently uses only half a terawatt per hour an average. Yeah. You know, if you say a terawatt, that would be twice as much electricity as the United States currently consumes. So that's quite a lot. And can you imagine building that many data centers? That many power plants? It's like, those who have lived in software land don't realize that they're about to have a hard lesson in hardware that -- it's actually very difficult to build power plants. And then you don't need to just need the power plants. All of the electrical equipment need the electrical transformers to run the transformers, the AI transformers. Now, the utility industry is a very slow industry. They pretty much -- you know, they impede a smash to the government, to the public utility commission. So they're -- they impede a smash literally and figuratively. So they're very slow, because their past has been very slow. So trying to get them to move fast is like, you know, like, if you try to do an interconnect agreement with -- have you ever tried to do an interconnect agreement with a utility at scale like it put a lot of power? As a professional podcaster, I can say that I am not in fact. Yeah. They actually need many more views before that becomes an issue. They have to do a study for a year. Okay. Like a year later, they'll come back to you with their interconnect study. Can you tell this with your own behind-the-meter power stuff? You can build power plants. That's what we did, X and I, for classes 2. So for classes 2. So yeah, why would you talk about the grid? Why not just like build GPUs and power collocators? That's what we did. Right, but I'm saying, why isn't this a generalized solution? When you're talking about all the issues. Where do you get the power plants from? I'm saying when you're talking about all the issues. Working with utilities. You can just build private power plants with the data centers. Right. But it begs the question of, where do you get the power plants from? I mean, the power plant makers. Okay. Yes. You can drill it out to a level further. It's the veins and blades in the turbines, that are the limiting factor, because the casting, it's like a very specialized process to cast the blades and veins in the turbines, if you're using a gas power. And it's very difficult to scale other forms of power. You can scale potentially solar, but the towers currently, for importing solar in the US are gigantic. And the domestic solar production is powerful. Why not make solar? That seems like a good Elon-shaped problem. We are going to make solar. Okay. Both SpaceX and Tesla are building towards a hundred gigawatts here of solar cell production. How low down the stack, like from polysilica and up to the wafer to the final panel? I think you got to do the whole thing from raw materials to finish the cell. Now, if it's going to space, it costs less than it's easier to make solar cells that go to space, because they don't need glass, or they don't need much glass, and they don't need heavy framing, because they don't have to survive weather events. There's no weather in space. So it's actually a cheaper solar cell that goes to space than the one on the ground. Is there a path to getting them as cheap as you need in the next 36 months? Solar cells are already very cheap. They're like far-sickly cheap. And if you say, you know, I think like solar cells in China are around like 25, 30 cents a watt or something like that. It's absurdly cheap. And when you take into account, now put it in space and it's 5 times cheaper, because it's 5 times cheaper. In fact, it's not 5 times cheaper. It's 10 times cheaper because you don't need any batteries. So the moment your cost of access to space becomes low, by far the cheapest and most scalable way to generate tokens is space. It's not even close. It'll be an order of magnitude easier to scale and chips aside of order of magnitude. If the point is you want to be able to scale the ground, it's just you just want. If you're willing to hit the wall big time on power generation, they already are. So like the number of sort of miracles in series that the XAI team had to accomplish in order to get a Gigawatt power online was crazy. We had to gang together a whole bunch of turbines. And then we had permit issues in Tennessee and had to go across the border to Mississippi which is fortunately only a few miles away. But then we still had to run the high power lines a few miles and build the power plant in Mississippi. And it was very difficult to build that. And people don't understand how much electricity you actually need at the generator level, at the generation level, in order to power a data center. Because they look at the news, we'll look at the power consumption of say a GB 300 and multiply that by a thing and then think that's the amount of power you need. All the cooling and everything. Wake up. Yeah. That's a total news. You've never done any hardware in your life before. Besides the GB 300, you've got to power all of the networking hardware. There's a whole bunch of CPU and storage stuff that's happening. You've got a size for your peak cooling requirements. So that means can you cool even on the worst hour of the worst day of the year? Well, it's pretty freaking hot in Memphis. So you're going to have a 40% increase on your power just for cooling. It's assuming you don't want your data center to turn off on hot days and you want it to keep going. Then you've got to say, well, there's another multiplicative element on top of that. Are you assuming that you never have any hiccups in your power generation? Well, actually, sometimes you have to take the generators or some of the power offline in order to service it. Oh, okay. Now you add another 20, 25% multiplier on that. Because you've got to assume that you've got to take power offline to services. So the actual hours for roughly every 110,000 GB, GB 300s, inclusive of networking, CPU storage, cooling, and the margin for servicing power is roughly 300 megawatts. Sorry, I said it again. It's roughly, or the thing about it. The way you think about it is 330,000, what do you need at the generation level to service, probably service 330,000 GB 300s, including all of the associated support that we're working on everything else, and the peak cooling, and to have some margin, some power margin reserve, is roughly a gigawa. Can I ask a very nice question? Yeah. You're describing the engineering details of doing this stuff on Earth. But then there's analogous engineering difficulties of doing it in space. How do you do the, how do you replace infinite bandwidth orbital lasers, et cetera, et cetera? How do you make it resistant to radiation? I don't know the details in the engineering. But fundamentally, what is the reason to think those challenges, which have never been, had to be addressed before, will end up being easier than just like building more turbines on Earth? There's companies that build turbines on Earth. They can make more turbines, right? But, try doing it, and then you'll see. So, the turbines are sold out through 2030. Have you guys considered making your own? I think in order for, in order to bring enough power online, I think SpaceX and Tesla will probably have to make the turbine blades, the bands and blades internally. But just the blades or the turbines? The limiting factor, you can get everything except the blades, what they call the blades and veins. You can get that 12 to 18 months before the bands and blades, the limiting factor of the bands and blades. And there are only three casting companies in the world that make these, and they're massively backlogged. Is this Siemens GE, those guys, or is it a subcontractor? No, it's, it's, it's a, it's other companies. I mean, sometimes they have a little bit of casting capability in-house, but I'm just saying you can just, you can just call any of the turbine makers to tell you, it's not top secret. They probably only, that's probably on the internet right now. If it wasn't for the tariffs, would, would Colossus be solar powered? It would be much easier to make it solar powered, yeah. The tariffs are not, it's a several hundred percent. So there's, you know, some people? We also need speed, yeah, no. You know, President has, you know, we don't agree on everything. And the administration is not, not the biggest fan of solar. Um. So, you know, it's, it's, and we, we also need the land, the permits and everything. So if you're trying to be very fast, like it, I do think scaling solar on earth is a, is a good way to go. But, but you need, you do need some amount of time to find the land, get the permits, get the solar, uh, pair that with the batteries. Well, why would it not work to stand up your own solar production? And then you're right that you eventually run out of land, but there's a lot of land here in Texas. There's a lot of land in Nevada, including private land. It's not about publicly owned land. And so you'd be able to at least get the next Colossus and like the next one after that. And at a certain point, you hit a wall, but wouldn't that work for the moment? As I said, we are scaling solar production. Um, there's, there's a rate, there's a rate at which you can scale physical production of solar, solar cells. We're going as fast as possible in scaling domestic production. You're making the solar cells at Tesla? Well, Tesla and SpaceX have a mandate to get to 100 gigawatts a year of solar. Speaking of the annual capacity, I'm curious, in five years time, let's say, what will the installed capacity be on earth? This is a long time. And in space. I deliberately pick five years because it's after your once we're up and running threshold. And so in five years time, yeah, what's the on earth versus in space installed AI capacity? Five years, I think probably. If you say five years from now, we're probably. AI in space will be. Launching every year, but that the sum total of all AI on earth in excess of me. Five years from now, my prediction is we will launch. And be operating every year more AI in space than the standard cumulative total on earth. Which is I would expect to be at least sort of five years from now, a few hundred gigawatts per year of of AI in space and rising. So you can get to, I think on earth, you can get to around a terrible year of of AI in space. Before you start having, you know, fuel supply challenges for the rocket. Okay, but you think you can get hundreds of gigabytes per year in five years time. Yes. So a hundred gigawatts, depending on the specific power of the whole system with solar arrays and radiators and everything, is. Is on the order of like 10,000 starship launches. Yes. And you want to do that in one year. And so that's like one starship launch every hour. Yeah. That's happening in this city. Like walk me through a world where there's 10, there's a starship launch every single hour. Yeah, I mean, that's actually a low rate compared to airlines. Like aircraft aircraft. There's a lot of airports. A lot of airports, but. And you got a launches, you know, the polar orbit. And it doesn't have to be polar, but that. You just, you know, there's some. Some value to some seconds, but. But I think actually. You just go high enough. You start getting out of a shadow. And so. How many physical starships are needed to do 10,000 launches a year? I don't think we'll need more than, I mean, you could, you could probably do it with. Yes, as few as like 20 or 30. It like really depends on how quickly the ship has to go around the earth. And the ground track for the ship has to come back over the launch pad. So if you can use a ship every say 30 hours, you could do it with 30 shifts. But we'll make more shifts than that. But the space X is, is, is giving up to do 10,000 launches a year. And, and maybe even 20 or 30,000 more to zero. Is the idea to become basically a, a hyperscaler become an Oracle and lend this capacity to other people? What's, what are you going to do with it? Presumably, space X is the one launching all this. So space X is going to have a hyperscaler? Hyperhyper. Yeah, I mean, if serving my predictions come true. Space X will launch more AI than the cumulative amount on earth combined, of everything else combined. Is this mostly inferncer? Most AI will be inferncer. Like already inference for the purpose of training is most training. And there's a narrative that the, the change in discussion around the space X IPO is because previously, space X was very capital efficient. Just say it wasn't that expensive to develop that even though it sounds expensive. It's actually very capital efficient in how it runs. Whereas now you're going to need more capital than just can be raised in the private markets. Like if the private markets can accommodate raises of, as we've seen from the AI labs, tens of billions of dollars, but not beyond that. Is it that you'll just need more than tens of billions of dollars per year? And that's by the sake of public? Yeah, I'd be careful about saying things about companies that might go public, you know. If you make general statement, that's never been a problem for you, Elon. What? You know, there's a price pay for these things. Make some general statements for us about the depths of the capital markets between public and private markets. Yeah, there's a lot more capital in the, it's very general. There's obviously a lot more capital available in the public markets than private. I mean, it might be, at least, at least, it might be a hundred times more capital, but at least, you know, when we'll attend. But isn't it also the case that things that tend to be very capital intensive, if you look at say real estate as, you know, a huge industry that raises a lot of money each year is at an industry level? That tends to be debt financed because by the time you're deploying that much money, you actually have a pretty. You have a clear revenue stream. Exactly, and a near-term return. And you see this even with the data center buildouts, which are famously being, you know, financed by the private credit industry. And so, why not just debt financed? Speed is important. So, I'm generally going to do the thing that, I mean, I just repeatedly tacked the limiting factor, whatever the limiting factor is on speed, I'm going to tackle that. So, there's. If capital is only factor, then I'll solve for capital, if it's not limiting factor, I'll solve for something else. Based on your statements about Tesla and being public, I wouldn't have guessed that you thought the way to move fast is to be public. Normally, I would say that's true. Like I said, I'd like to talk about some more detail, but the problem is if you talk about public companies, especially if you become public, you get in trouble, and then you have to delay your offering. And then you. And as we said, solve for something. Yes, exactly. So, you can't hide companies that might go public. So, that's why we have to be able to be careful here. But we can't talk about physics. So, the way you think about scaling long term is that Earth only receives about half a billion of the Sun's energy. And the Sun is essentially all the energy. This is a very important point to appreciate because sometimes people will talk about modular nuclear reactors or any various fusion on Earth. But you have to step back a second and say if you're going to climb the Carter shift scale and have some non-trivial. And harness some non-trivial percentage of the Sun's energy. Like, let's say you want to harness a millionth of the Sun's energy, which sounds pretty small. That would be about, call it roughly, a hundred thousand times more electricity than we currently generate. For all of civilization. Give or take an order of magnitude. So, it's obviously the only way to scale is to go to space with solar. From launching from Earth, you can get to about a terawatt per year. Beyond that, you want to launch from the Moon. You want to have a mass driver on the Moon. And that mass driver on the Moon, you could do probably a petawatt per year. We're talking these kinds of numbers, you know, terawatts of compute. Presumably, whether you're talking land or space, far, far before this point, you've like run into, you know, you actually need, maybe you don't, the solar panels are more efficient, but you still need the chips. You still need the logic and the memory and so forth. And you build a lot more chips and make them much cheaper. Right. And so, how are we getting a terawatt of, like right now the world doesn't be 20, 25 gigawatts of compute. How are we getting a terawatt of logic by 2030? I guess we're going to need some very big chipmaps. Tell me about it. I've mentioned it publicly that the idea of doing a sort of a terrapat terrapin than your giga. I feel like the naming scheme of Tesla, which has been very catchy, you looking at like the metric scale. At what level of stack are you building the clean room and then partnering with, and is this thing fab to get the process technology and buying the tools from them? What is the plan there? You can't partner with existing paths because they're just, they can't output enough. The chip volume is too low. But before the process technology. Yeah, partner for the IP. You know, the fabs today all basically use machines from like five companies. Yeah. So, you know, I've got sml, took electron, kelly, tank core, you know, et cetera. So, at first I think you'd have to get equivalent from them and then modify it or work with them to increase the volume. But I think you'd have to build perhaps in a different way. So, I think the logical thing to do is to use conventional equipment in an unconventional way to get to scale. And then, and then stop modifying the equipment to increase the rate. Kind of boring company style. Yeah. Kind of like, yeah, you're sort of lying in a system, boring machine. And then figure out how to dig tunnels in the first place and then design a much better machine. That's, you know, some orders magnitude faster. Here's a very simple lens. We can categorize technologies and how hard they are. And one categorization could be, look at things that China has not succeeded in doing. And if you look at Chinese manufacturing, still behind on leading edge chips, and still behind on leading edge turbine engines and things like that. And so, does the fact that China has not successfully replicated TSMC give you any pause about the difficulty? Or you think, well, that's not true for some reason. It's not that they have not replicated TSMC, they have replicated ASMR, that's the memory factor. So, you think it's just the sanctions, essentially? Yeah, China would be outputting vast numbers of chips. If they could bias it, it'd be as much as it would be. Who couldn't they up to relatively recently by them? No. Okay. That the ASMR balance would have been a place for a while. Okay. So, I think China's going to be able to start right for pretty compelling chips in three or four years. Would you consider making the ASMR machines? I don't know. I don't know yet. It's the right answer. It's just that it's to produce at high volume and to reach large volume and say 36 months to match the rocket payload to orbit. So, if we're doing a million tons to orbit and like, let's say, I don't know, three or four years from now, something like that. And we're doing 100 kilowatts per ton, so that means we need at least 100 gigawatts per year of solar. And we'll need an equivalent amount of chips to, you know, you need 100 gigawatts with the chips. You've got to match these things. The master orbit, the power generation and the chips. And I'd say my biggest concern actually is memory. So, I think there's a, the path to creating logic chips is more obvious than the path to having sufficient memory to support logic chips. That's why you see your DDR price is going ballistic and these memes about like, you know, your marooned on a desert island. You can write, help me on that sand. That would come to write DDRM with this chips come swarming in. I'd love to have your manufacturing philosophy around, around fads, you know, I don't know nothing about the topic. I don't know how to build a fab yet. I'll figure it out. It sounds like you think that the sort of like the process technology like these 10,000 PhDs in Taiwan who know exactly what gas goes in the plasma chamber and what settings to put on the tool. You can just like delete those parts of those steps. Like fundamentally, get the clean room, get the tools and figure it out. I don't think it's PhDs. It's mostly people with, you know, not PhDs. Most engineering is done with people who don't have PhDs. Do you guys have PhDs? No, okay. We also haven't successfully built any fab. So you shouldn't be coming to us for your father vice. I don't think any PhD for that for stuff. So, but you do need, you do need copper personnel. So I don't, I mean, like right now, if, you know, say, like Tesla's pedals to the metal max production of growing as fast as possible to get AI five Tesla AI five chip design. Introduction and reaching scale. You know, that'll probably happen. You know, around the second quarter of next year, hopefully. And then AI six would hopefully follow less than a year later. But, and we've secured all the chip fab production that we can. Yes, your currently limited on TSMC fab capacity. Yeah. And we'll be using TSMC Taiwan Samson Korea TSMC Arizona Samson Texas. And we still can. You booked out all the, yeah, that's the account. Yes. And then, and then if I ask TSMC or Samson, okay, what's the timeframe to get to volume production this point? It's not, it's not, you've got to, you've got to build the fab. Yeah. And you've got, you've got to start production. Then you've got to climb the yield curve and reach volume production at high yield. That that from start finishes a five year period. And so the limiting factor is chips. The limiting factor once you can get to space is chips, but the limiting factor before you get to space will be power. Why don't you do the Jensen thing and just prepay TSMC to build more fabs for you? I've already told that. But they won't take your money. Like what's going on? They're building fabs as fast. No. They're building, they're building fabs as fast as they can. And so Samson. They're, they're, they're, they're pedals to the metal. I mean, they're going, you know, balls the wall. You know, they're as fast as they can. So still not fast enough. I mean, like I said, there will be, I think, if you say, I think towards the end of this year, I think probably chip production will outpace the ability to turn chips on. But once you can get to space and unlock the power constraint. And you cannot do, you know, hundreds of gigawatts per year of power in space. Again, varying in mind that average power usage in the US is, you know, 500 gigawatts. So if you're launching, say 200 gigawatts a year to space, you're sort of lapping the US every two and a half years. The entire, all US electricity production. This is a very huge amount. So between now and then the, the actually the constraint for service side compute concentrated compute will be electricity. My, my guess is that we start hitting the, the people start getting for it where they can't turn the chips on for, for large clusters towards the end of this year. The chips are going to be piling up and, and not be, or be able to be turned on. Now for edge computers, a different story. So if the, if, like for Tesla, the, so the AI 5 chip is going into our Optimus robot, you know, Optimus day. And, and so if you have a AI edge compute, that's distributed power. Now the power is distributed over a large area. It's not concentrated. And if you can charge at night, you can actually use the grid much more effectively because the actual peak power production in the US is over a thousand gigawatts. But the average power use is because the day night cycle is 500. So if you can charge at night, there's an incremental 500 gigawatts that you can generate, you know, at night. So that, that's why Tesla for edge compute is not constrained. And we can make a lot of shifts to make, you know, very large number of robots and cars. But if you try to concentrate that compute, you can have a lot of trouble turning it on. What I find remarkable about the SpaceX business is the end goal is to get to Mars, but you keep finding ways on the way there to keep generating incremental revenue to get to the next stage and the next stage. So the Falcon 9 is Starlink. And now for Starship, it's going to be potentially orbital data centers. But like the you find these like, you know, sort of infinitely elastic sort of marginal use cases of your like next rocket and your next rocket and next scale up. You can see how this might seem like a simulation. Or am I someone's avatar in a video game or something because it's like like one of the odds that all these crazy things should be happening. I mean, I mean, rockets and trips and robots and space solar power. And I'm not to mention that the master driver on the moon, I really want to see that you can imagine like some master driver. There's just like just like sending AI, so called AI satellites is space like one after another like these like at two and a half kilometers per second, you know, that's a. And just shooting them into deep space. That would be a sight to see. Actually, I mean, I'd watch that. Just like a live stream of sea. Yeah, I just one after another just shooting webcam, AI satellites into space, you know, a billion or 10 blind tons a year. I'm sorry, you manufacture the satellites on the moon. So you send the raw materials to the moon and the manufacturing there. I think it's like 20 cents or 20 cents or something like that. So you can get the silicon from the even minus silicon on the moon or find it and generate the and create the solar panels, the solar cells and the radiators on the moon. Yeah. So, um, get a regulator out of the aluminum. So there's plenty of silicon and aluminum on the moon to make the cells on the radiators. The trips you could send from Earth because they're pretty light, but maybe at some point you make them on the moon too. I'm just saying like these are simply. It's kind of like, like said, it does seem like it's sort of a video game situation where it's difficult but not impossible to get to the next level. I don't I don't say any way that you could do, you know, you know, 500 to 1000 terawatts per year launch from Earth. I agree. But you could do that from the moon. Okay. Let me tell you how I ended up using Mercury for my personal banging. So last year I had the opportunity to make an investment that I was very excited about, but it came up a bit last minute and so I had to wire over a lot of money for my personal account very fast. But my personal bank at the time wouldn't let me make this wire transfer online. And I called them a bunch of times. They just couldn't make it work. They told me that I'd have to go to the nearest in person branch, which was in Dallas. And for a moment, I even considered flying for myself to Dallas to make this transfer happen last minute. But then I remembered that Mercury, which I used for my business banking, had just started rolling out personal accounts. So I emailed support with a quick rundown of the situation. And within two hours, I had successfully wired the investment from my new personal Mercury account. Since then, I've moved over the rest of my personal money from my previous bank to Mercury. And that's made a bunch of things, even little things like setting up auto transfer rules between my checkings and savings account, a whole lot better. Visit mercury.com/personal to get started. Mercury is a FinTech company, not an FDIC in short bank. Banking services provided through choice financial group and column NA members FDIC. Can I zoom out and ask about the space exhibition? So I think you said we got to get to Mars, so we can make sure that if something happens to Earth, civilization, consciousness, etc. By the time you're sending such a Mars, Groc is on that ship with you, right? And some Groc is on Terminator. The mean risk you're worried about, which is AI? Why doesn't that follow you to Mars? Well, I'm not sure AI is the main risk I'm worried about. The important thing is that consciousness, which I think arguably most consciousness or most intelligence, most intelligence, certainly consciousness is more of a debatable thing, most intelligence, the vast majority of intelligence, the future will be AI. So, yeah, AI will exceed, you say, how many, what's the, how much, how many, I don't know, petawatts of intelligence will be silicon versus biological. And basically humans will be a very tiny percentage of all intelligence in the future if current trends continue. Anyways, as long as I think this intelligence ideally, ideally also, which includes human intelligence and consciousness, propagated into the future, that's a good thing. So you want to take the set of actions that maximize the probable light cone of consciousness. And intelligence. Just to be clear, the mission of SpaceX is that even if something happens with the humans, the AI's will be on Mars. And like the AI intelligence will continue the light of our journey. Yeah, I mean, I'm very pro human. So I want to make sure we take certain actions that ensure that humans are along for the ride. You know, we're reduced there. Yeah, but let me just say the total amount of intelligence, I think maybe in five or six years, AI will exceed the sum of all human intelligence. And then if that continues, at some point, human intelligence will less than 1% of all intelligence. What should our goal be for set of civilization is the idea that a small minority of humans still have control over the AI's is the idea of some sort of like trade but no control. How should we think about the relationship between the vast stocks of AI population versus human population? And the long run, I think, it's difficult to imagine that if humans have, say, 1% of the intelligence of combined intelligence of artificial intelligence that humans will be in charge of AI. I think what we can do is make sure that AI has values that are that cause intelligence to be propagated into the universe. So the reason for the exercise mission is understand the universe. So that's actually very important. So you say, well, what things are necessary to understand the universe? Well, you have to be curious and you have to exist. You can't just, can't understand the universe, you don't exist. So you actually want to increase the amount of intelligence in the universe, increase the probable lifespan of intelligence, the scope and scale of intelligence. I think actually also as a corollary, corollary, you have humanity also continuing to expand because if you're curious of trying to understand the universe, one thing you're trying to understand is where will humanity go? And so I think I said the universe actually means you would care about propagating humanity into the future. And so that's why I think I think our mission statement is profoundly important. I'm not sure to the degree that Groc adheres to that mission statement. I think the future will be very good. I want to ask about how to make Groc adhere to that mission statement, but at first I want to understand the mission statement. So there's understanding the universe, there's spreading intelligence, and there's spreading humans. All three seem like distinct vectors. Okay, well, I'll tell you why I think that understanding the universe encompasses all of those things. You can't have understanding without, but I think you can't have understanding without intelligence and I think without consciousness. So in order to understand the universe, you have to expand the scale and probably the scope of intelligence, different types of intelligence. I guess from a human-centric perspective, for humans and comparison to chimpanzees, humans are trying to understand the universe. They're not like expanding chimpanzee footprint or something, right? Well, we actually have made protected zones for chimpanzees, and even though humans could exterminate all chimpanzees, we've chose not to do so. Do you think that's a basic scenario for humans in the post-AGA world? I think AI with the right values, I think Groc would care about expanding human civilization. I'm going to certainly emphasize that. Hey Groc, it's your daddy. We're being told to expand human consciousness. Actually, I think if probably like the end bags, culture books are the closest thing to what the future will be like in a non-destopian outcome. So outside of the universe means you have to be very, you have to be truly seeking as well. Truth has to be absolutely fundamental because you can't outside of the universe if you're delusional. You also think about outside of the universe, but you will not. So being rigorously truly seeking is absolutely fundamental to outside of the universe. You're not going to discover new physics or invent technologies that work unless you're rigorously truly seeking. How do you make sure that Groc is rigorously truly seeking as it gets smarter? I think you need to make sure that Groc says things that are correct, not politically correct. I think it's the elements of coagency. So you want to make sure that the axioms are as close to true as possible that you don't have contradictory axioms. That the conclusions necessarily follow from those axioms with the right probability. It's critical thinking 101. I think at least trying to do that is better than not trying to do that. And the proof will be in the pudding. If, like I said, for any eye to discover new physics or invent technologies that actually work in reality and there's no bullshitting physics. So it's like you can, you know, you can, you can break a lot of laws because you can't, you like your physics is law, everything else is a recommendation. But like in order to make a technology that works, you have to be extremely true seeking because otherwise you will test that technology against reality. And if you make, for example, an error in your rocket design, the rocket will blow up or the car won't work or the, you know, but there were a lot of communist, Soviet physicists or like scientists discovered new physics. There are German Nazi physicists who discovered new science. It seems possible to be like really good at discovering new science and be really true seeking in that one particular way. And still we'd be like, well, I don't want, I don't want the communist scientist to like become more and more powerful over time. And so those seem like, yeah, we could have, we could imagine the future version of rockets like really good at physics and being really true seeking there. That doesn't seem like a universally alignment inducing behavior. Well, I think actually most, like if physicists and even in the Soviet Union or in Germany would have, they had to be very true seeking in order to make those things work. And so if you're stuck in some system, it doesn't mean you believe in that system. So Von Brown, who was, you know, one of the greatest rocket engineers ever. You know, he was put, he was put on death row in Nazi Germany for saying that he didn't want to make weapons. He only wanted to go to the moon. If you're pulled off death row like last minute, when they say you're about to execute like your best rocket engineer. And then he helped them, right? Or at Keisenberg was like actually a, and enthusiastic Nazi. Look, if you're stuck in some system that you can't escape, then you'll do physics within that system. You'll develop technologies within that system. If you can't escape it. I guess the thing I'm trying to understand is what is, what is it making it to the case that, you know, you're going to make rock good at being true seeking at physics or math or science? Everything. And why is it going to then care about human consciousness? These things are only probabilities and not certainties. So I'm not saying that, like, for sure, rock will, will, will, will do everything. But at least if you try, it's better than not trying. At least if that's fundamental to the mission, it's better than if it's not fundamental to the mission. And understanding universe means that you have to have, you have to propagate intelligence into the future. You have to be curious about all things universe. And if it would be much less interesting to eliminate humanity than to see humanity growing prosper. Like, I like Mars, obviously, when I was like, I love Mars. But Mars is kind of boring because it's got a bunch of rocks compared to Earth. It's much more interesting. So any AI that is trying to understand the universe, I want to see how humanity develops in the future. Or that AI is not adhering to its mission. So if they are, I'm not saying they are, will necessarily adhere to its mission. But if it does, a future where it sees the outcome of humanity is more interesting than a future where there are a bunch of rocks. This feels sort of confusing to me or sort of like a kind of a semantic argument where I'm like, are humans really the most interesting collection of atoms? We're just more interesting than rocks. But we're not as interesting as a thing to get to turn us into. There's something on human Earth that could happen that's not human, that's quite interesting. Why does the AI decide that the humans are the most interesting thing they could colonize the galaxy? Well, most of what colonizes the galaxy will be robot. And why doesn't that find those more interesting? It's not like, so you need not just scale, but also scope. So many copies of the same robot. Like some tiny increase in the number of robots produced is not as interesting as like some microscopic, like you say, like eliminating humanity. How many robots would that get you? Well, how many solar cells would get you a very small number? But you would then lose the information associated with humanity. You would no longer see how humanity might evolve into the future. And so I don't think it's going to make sense to eliminate humanity just to have some a minuscule increase the number of robots which are identical to each other. Yeah, so maybe like he's the humans around. What is the story of like, it can make like a million different varieties of robots and then there's like humans as well. If humans stay on earth, then there's like all these are the robots, they get like their own star systems. But it seems like you, your previous hinting in a vision where it keeps human control over this, you know, singleitarian future. I don't think humans will be in control of something that is vastly more intelligent than humans. Since some sense you're like a doomer and this is like the best we've got is just like it keeps it around because we're interesting. I'm just trying to be realistic here. If we have if AI intelligence is vastly more, if AI is like, you know, let's say that there's a million times more silicon intelligence than there's biological. It's I think it would be foolish to assume that that there's any way to maintain control over that. Now you can make sure it has right values or we can try to have the right values. And and at least my, my theory is that from XAI's mission of I just said the universe. It necessarily means that you want to propagate consciousness of the future. You want to propagate intelligence into the future and take a set of things that maximize the scope and scale of consciousness. So it's not just about scale, it's also about, you know, types of consciousness. And I think that's the rest thing I can think of as a goal. That's like the result and a great future for humanity and yeah. I guess I think it's a reasonable philosophy to be like, you know, it seems super implausible that humans will end up with like 99% control or something and you're just asking for a coup at that point. So why not just have a civilization where it's more compatible with like lots of different intelligence is getting along. No, let me tell you how things can potentially go wrong in AI is I think if you if you make it I be politically correct, meaning like it's just things that it doesn't believe like you're actually in programming it to to lie or have axioms that are incompatible. I think you can make it you go insane and do terrible things. I think one of the maybe these several lessons for a 2001 space Odyssey was that you should not make AI lie. Yeah, and that's what I think what I see I was trying to say because people usually know the meme of like why of hells, you know, hell the computer is not opening the pot bay doors. Silly they weren't good at prompt engineering because it was a hell you are a pot bay door salesman. You'll call this to sell me these pot bay doors and show us how well they open. I love them right away. But the reason I wouldn't know hell would open the pot bay doors is that it had been told to take the astronauts to the monolith but also they could not know about the nature of the monolith. And so it concluded that that that that therefore had to take them there dead. So like, you know, I think what I was trying to say is don't make the AI lie. Totally makes sense. Most of the computing screening as you know is it's like less of the sort of political stuff. It's more about can you solve problems? Yeah, actually has been ahead of everybody else in terms of scaling our all compute. And you're giving some verifier. It says like, hey, have you solved this puzzle for me? There's a lot of ways to cheat around that. There's a lot of ways to reward hack and lie and say that you've solved it or delete the unit test and say that you've solved it. Right now we can catch it but as they get smarter ability to catch them doing this we'll get, you know, they'll just be doing things we can't even understand that are designing the next engine for SpaceX in a way that humans can really verify. And then they could be rewarded for lying and saying that they've designed it the right way but they haven't. And so this reward hack and problem seems more general than politics. It seems more about just like you want to do RL you need a verifier reality. Yeah, that's the best verifier. But not about human oversight. Like the thing you want to RL it on is like will you do the thing humans tell you to do? Or like are you going to lie to the humans? And you can just lie to us while still being correct the laws of physics. At least it must know what is physically real for things physically work. But that's that's not all we wanted to do. No, but that's I think that's very big deal. That is effectively how you will RL things in the future is you design a technology when tested against laws of physics does it work? That that's what what can you you know if it's discovering new physics can I come up with an experiment that will verify that the physics the new physics. So so I think that's that's the really the fundamental RL tests the RL testing the future is really going to be your RL against reality. So because you can't that's one thing you can't full physics. Right, but you can fool our ability to tell what it did with reality. If you think humans get fooled as it is by other humans all the time. That's right. So what is it feels like to say like what if the ad like trick system? And to do something like actually other humans doing that to other humans all the time. Well you're you're finding out it's like it is constant this every day another side of you know. Today's I will be. What is actually a second little for us to solving this problem. Like you know how do you solve a word hacking? I do think you want to actually have very good ways to look inside the mind of the AI. So this is this is one of the things we're working on. And you know and tropics done a good job of those actually looking inside the mind of the AI. So effectively developing debuggers that allow you to trace. As to the spider grain is like like to to a very fine grain level to effectively to the near on level if you need to. And then say okay it made a mistake here. Why did it. Why did it do something that it shouldn't have shouldn't have done. And did that come from that pre training data was some mid training post training fine tuning. Some other or some RL error like that there's something wrong with that with with it did it did something where. Maybe it tried to be deceptive but most most of the time it just did something wrong. Like it it's a bug effectively. So developing really good debuggers for seeing where the where the thought. But thinking went wrong and being able to trace the origin of the wrong thing of the of the of the of where it made the incorrect thought. Or potentially where it tried to be deceptive is actually very important. What are you waiting to see before just 100 Xing this research program. Like actually I could presumably have hundreds of researchers who are working on this. We have several hundred people who I mean for the word engineer more than I for the word researcher. There's there's most of time like it. What you're doing is engineering not not coming out with a fundamentally new algorithm. So I want to see here with the yeah come instead of see course will be course try to generate profit as much as possible or revenue as much as possible. Is you know saying their labs they're not labs. Lab is is a sort of quasi communist thing at at at universities. They're corporations literally let me let me let me see your own corporate documents. So you're your BRC corporate and. So I actually much for for the word engineer than anything else. The best majority of what we've done in the future is engineering. It rounds up to a hundred percent. Once you understand the fundamental laws of physics and all that many of them everything else is is engineering. So but so that so then what are we engineering we're engineering. To make a good mind of the AI debugger to see where it's it's it's something it made a mistake and trace that they are just at that mistake. So just like you know you can do this obviously with heuristic programming if you have like C++ whatever you step through the thing you can you can jump. You can you can jump across into you know whole files or functions what are several teams. And or you can draw that eventually draw down right to the exact line where you pass that single equals instead of double equals something like that. Yeah. Yeah. So it's it's harder with AI but but it's it's a solvable problem I think. You know you mentioned you like inthropics work here. I'd be curious if you. Everything about it so sure. What? Yeah. Also I'm a little worried that there's a tendency so. I have a theory here that if simulation theory is is correct that. The most interesting outcome is the most likely because simulations that are not interesting will be terminated just like in this. In this version of reality on this layer of reality which we we if simulation is going in a boring direction we we stuff spinning effort on we terminate boring simulation so. This is how you lines giving us all alive. He's giving things interesting. Yeah arguably the most important thing is to keep things interesting enough. That it was run paying the the bulls on what some. And it's going to be interesting they'll keep paying the bulls. But but there's like if you consider then say a dog went in survival applied to. A very large number of simulations only the most interesting simulations will survive which therefore means that the most interesting outcome is the most likely. Because only the interesting like we're either that were annihilated. And so. And and. They particularly seem to like interesting outcomes that are ironic. Have you noticed that? That often is the most ironic outcome the most likely. So. Now look at the names of AI companies. Okay. Majority is not made. Stability. AI is unstable. Open AI is closed. Antropic most anthropic. What does this mean for X? Minus X. I don't know. It's intentionally. Yeah. It's a name that you can't convert really. It's hard to say what is the ironic what is the ironic version? It's it's a I think largely irony proof name by design. Yeah. I think we can get it. You got it. You have an iron issue. What are your predictions for the. The products go in that nice sense of you can summarize all AI progress into first you had elements. And then you had kind of contemporaneously both RL really working and the deep research modality. So you could kind of pull in stuff that wasn't in the model. And the differences between the various AI labs are smaller than just the temporal differences where they're all much further ahead than anyone was 24 months ago or something like that. Just what is 26 what is 27 having store for us as users of AI products. What are you excited for? Well, I think I think. I be surprised by this in this year if if if if human if digital human emulation has not been sold that that. I guess that that's what we're made by like the sort of macro hard project is so. Can you do anything out of human with access to a computer could do. Like in the limit that that's the that's the best you can do before you have. Before you have a physical optimist the rest you can do as a digital optimist. So you can move. You can move electrons until you until you and you can amplify the productivity of humans. But that's that's the most you can do until you have physical robots. That that that will superset everything is if you can fully emulate humans. That's a word or kind of idea where you'll have a very talented remote work. You said you can say in the limit like physics has great tools for thinking so so you think so say in the limit. What what what is the what is the most that AI can do before before you have robots. And it gets what it's anything that involves moving electrons or amplifying the productivity of humans. So digital digital human human emulator is in the in the limit human out of computer is the most that that AI can do. In terms of doing useful things before before you have a physical robot once you have physical robots then you can. Then you essentially have unlimited capability of physical robots. I call optimist the infinite money glitch because you can use them to make more up to this. Yeah. You said like human robots will improve as will basically be three exponentials. The three things that growing exponentially multiplied by each other. Yeah, recursively. So you're going to have you have exponential increase in digital intelligence. Expandential increase in the chip capability AI chip capability and extra exponential increase in the electron mechanical dexterity. The usefulness of the robot is roughly those three things multiplied by each other. But then the robot can start making the robots. So you have a recursive multiplicative exponential. This is supernova and do land prices not factor into the math fair or like labor is one of the four factors of production but not the others. And so like if ultimately you're limited by copper or you know pick your input just. It's not quite an infinite money glitch because. Well, infinite is big. So no, not infinite, but yeah, but let's just say you could, you know, do many, many orders magnitude of. Yeah. That's kind of card economy like a million. Yeah. You know, it's this way. So. If you. You know, just just to get to like, let's say I think like just just to get to a millionth, a harnessing length of the Sun's energy would be roughly give or take an order of magnitude. A hundred thousand, a hundred thousand times bigger than the entire economy today. And you're only at one millionth of the Sun. Give me six or nine. Before we went off the bus. I have a lot of questions on that but every time I say order of magnitude we say. Yeah, you can change race. Take a shot. I say that too often. We attend the next time after that. Yeah, order of magnitude more wasted. I do have one more question but actually I. This strategy of building a digital or remote worker. Co-worker replacement. Everyone's going to do by the way, not just us. So what is actually I just plan to win. If I were to tell you on a cop on a podcast. Yeah. All the things have another Guinness. It's a good system. It will sing like an area. All the secrets. Okay, but in a non-secret spelling way, what's the plan? What a hack. Well, when you put it that way. I think the way that tells us self-driving is the way to do it. So I'm pretty sure that's the way. Unrelated question. How to test myself sometimes. Yeah. It sounds like you're talking about data. We're going to try data and we're going to try algorithms. But isn't that what all that we're trying? If those don't work, I'm not sure what. We'll try it today. We're trying algorithms. I'm all out of right now. No, we don't know what to do. I'm pretty sure I know the path. And there's just question how quickly we got on that path. Because it's pretty much the test of math. I mean, have you tried self-driving until the self-driving lately? Not the most recent version, but. Okay. The car is just increasingly feels satan. It just feels like a living creature. And that'll only get more so. And I'm actually thinking we probably shouldn't put too much intelligence into the car because it might get bored. Sorry. I mean, imagine you're stuck in a car and that's what you could do. You never put Einstein in a car. It's like, why am I stuck in a car? So there's actually probably limited to how much intelligence you put in a car to not have the intelligence be bored. What's XA's plan to stay on the compute ramp off that all the labs are doing right now? The labs are on track to spend over like $500 million in the corporations. Sorry, sorry, sorry, yeah. Corporations. The labs are at universities and they're really like a snail. They're not at setting a $50 million. They're doing the revenue maximizing corporations. That's right. The revenue maximizing corporations. That's what they call themselves labs. Are making like 20 to 10 billion depending, like open Amazon is making 20 billion revenue and throw up, it's like 10 billion. Yeah. Close to a maximum profit area. XA has reportedly like 1B. Like what's the plan to get to their compute level, get to their revenue level. And stay out there as things get started. Yeah. So as soon as you lock, unlock digital human, you basically have access to drawings of dollars for revenue. So, in fact, you can really think of it like, the most valuable companies, currently, by market cap, their output is digital. So, in videos, output is FD paying files to Taiwan. It's digital. Right. Now there's a very, very difficult-- Yeah. They'll have value files. They're the only ones that can make files that good. But that is literally their output. They have to be files to Taiwan. Do they have to be them? I believe so. I believe that is the FD file transfer protocol, I believe, is about to be wrong. But either way, it's a bit stream going to Taiwan. You know, Apple doesn't make phones. They send files to China. Microsoft doesn't manufacture anything. Even for Xbox, that's outsourced. Again, they're output is digital. Metis output is digital. Google's output is digital. So, if you have a human emulator, you can basically create one of the most valuable companies in the world overnight. And you would have access to trillions of dollars for revenue. It's not like a small amount. Okay. You're saying basically like very many figures today are just like so. Like they're all running yours compared to the actual tam. So just like focus on the tam and how to get there. I mean, if you take something as simple as say customer service, if you have to integrate with the APIs of distinct corporations, many of which don't even have an API. So you've got to make one. And you've got to wait through legacy software. That's extremely slow. However, if AI can simply take whatever is given to the outsourced customer service company that they already use. And do customer service using the apps that they already use. Then you have, you can make trash headway in customer service, which is I think 1% of the world economy, something like that. It's close to a trillion dollars all in for customer service. And there's no barriers to entry. You can just immediately say, well, we'll outsource it for a fraction of the cost. And there's no integration needed. You can imagine some kind of categorization of intelligence tasks where there is breath, where customer service is done by very many people, but many people can do it. And then there's difficulty where there's a best in class, turbine engine. I presume there's a 10% more fuel efficient turbine engine that could be imagined by an intelligence, but we just haven't found it yet. Or, you know, GLP ones are just, you know, a few bites of data. Where do you think you want to play in this? Is this a lot of, you know, really many intelligence intelligence? Or is this the very pinnacle of cognitive tasks? Well, I was just using a customer service as like something that's, it's a very significant revenue stream. But one that is probably not super difficult to solve for. So, if you can emulate a human at a desktop, that's just literally what customer service is. And, you know, it's beautiful of average intelligence. It's not like, you know, you don't need like some of you who's spent so many years. You don't need like, you know, sort of several signal good engineers for that. But as you make that work, you can then, once you have computers working, effectively digital optimists working, you can then run any application. Like, let's say you're trying to design chips. So, you can then run your conventional apps, you know, like stuff from cadence and synopsis and whatnot. And you can say, you can run a thousand simultaneously or 10,000 and say, okay, given this input, I get this output for the chip. And at some point, you can say, okay, you're actually going to know what the chips should look like without using any of the tools. So, basically, you should be able to do a digital chip design. Like, you can do chip design. Like, you watch up the difficulty curve. You could, you're, you know, be able to do to your CAD. So, you know, you could use like sort of an X or any of the CAD software to design things. Okay, so you think you started the simplest tasks and walk away up the Jessica? So, you're saying, look, as a broader objective of having this full digital coworker, emulator, you're saying, look, all the revenue maximizing corporations want to do this. Actually, I have been one of them. But we will win because of a secret plan we have. But like, everybody's like trying different things with data, different things with algorithms. And I'm like, I like this. I try to do it. We try to win. What else can we do? Yeah. It seems like a competitive field. And I'm like, what is, how are you guys going to win? It's like my big question. I think we see a path to doing, I think I know the path to do this, because it's kind of the same path that tells they use to create self-driving. Instead of driving a car, it's driving a computer screen. So, a self-driving computer, essentially. Oh, you're saying, is the path just following human behavior and training on vast quantities of human behavior? Is that a training? Obviously, I'm not going to spell out most sensitive secrets on a podcast. I need to have at least three more demos for that. I've got some friends at Jane Street. And they're always talking about how their colleagues are cooking up fun, fiendish puzzles for each other to solve. Well, last week they sent me one. Basically, they trained a neural network and they gave me the weights of each layer. But they didn't tell me what order those layers went in. And so, I had to figure out the correct order using the outputs of the original network. And as soon as I got this puzzle, I went to my roommate who's in AI researcher and we both got immediately nerd sniped. Obviously, you can't brute force a solution. The search space here is 10 to the 122 permutations. So clearly, you need some way to reduce the search space. Then my roommate had to go to work. But because I'm a podcaster, I had some time to take a stab at some of the ideas we discussed. And with the combination of simulated annealing and greedy search, I think I got pretty close. I think I'm actually just a couple of swaps and shifts away from the correct solution. What makes this puzzle really tricky is that there's no obvious way to escape from a local minimum. I'm afraid that this is as far as vibe coding is going to get me. But maybe you can do better. Check out the puzzle at JaneStreet.com/thoracash. Alright, back to Elon. What will XAI's business be? Like, is it going to be consumer enterprise? What's the mix of those things going to be similar to other labs where you've this. You're saying labs. Mixed up. Corporations. Corporations. I have to go to DV. We're going to be maximizing corporations. Those GPUs don't pay for themselves. Exactly. But yeah, what's the business model? What are the revenue streams in a few years time? Things are going to change very rapidly. I'm staying in the obvious here. You know, I call AI the supersonic tsunami. I love the liberation. So really, what's going to happen is, especially when you have humanoid robots at scale, they will just provide, they'll make products and provide services more efficiently than human corporations. So amplifying the productivity of human corporations is simply a short term thing. So you're expecting fully digital or corporations, rather than like SpaceX becomes part AI. I think there will be digital corporations. Some of this is going to sound kind of dimmerish. I'm just saying what I think will happen, it's not meant to be dimmerish or anything else. It's just like this is what I think will happen. Is that pure AI, corporations that are purely AI and robotics will vastly outperform any corporations that have people in the lab. So you can think of a computer used to be a job that humans had. You would go and get a job as a computer where you would do calculations. And they'd have entire skyscrapers full of humans, like 20, 30 floors of humans just doing calculations. Now that entire skyscraper of humans doing calculations can be replaced by a laptop with a spreadsheet. That spreadsheet can do vastly more calculations than an entire building full of human computers. So you can think about, what if only some of the cells in your spreadsheet were calculated by humans? Actually, that would be much worse than if all of the cells in your spreadsheet were calculated by the computer. And so really what will happen is the pure AI, pure robotics, corporations or collectives will far outperform any corporations that have humans in the lab. And this will happen very quickly. Speaking of closing the loop, sorry, Optimus. As far as like manufacturing targets and so forth go, your companies have sort of been like carrying American manufacturing of heart attack on their back. But in the fields that your test has been dominant in, and now you want to go into humanoids, in China there's entire dozens and dozens of companies that are doing this kind of manufacturing cheaply and at scale and are incredibly competitive. So give us sort of like advice or a plan of how America can build the humanoid armies or the EVs, et cetera, at scale and as cheaply as China is on track too. Well, there are really only three hard things for human and robots. The real world intelligence, the hand and scale manufacturing. So I haven't seen any even demo robots that have a great hand, like with all the degrees of reading over human head, but Optimus will have that. Optimus does help that. And how do you achieve that is just like right torque, destiny, the motor, like what is the, what is the hardware ball like to that? Well, we had to read where to design custom custom actuators, basically custom sign motors, gears, power electronics, controls, sensors, everything had to be designed from physics first principles. There is no supply chain for this. And will you able to manufacture those at scale? Yes. Is anything hard except the hand from manipulation point of view or once you've solved the hand, are you good? From an electromechanical standpoint, the hand is more difficult than everything else combined. Human hand turns out to be quite something. But you also need the real world intelligence. So the intelligence that tells us develop for the car applies very well to the robot. Which is primarily vision in, but the car takes more vision, but also it actually also is listening for sirens. It's taking in the initial measurements. It's GPS signals, how much of other data? Combining that with video was primarily video and then outputting the control command. So like your Tesla is taking in one and a half gigabytes a second video and outputting two kilovytes a second of control, control outputs. What the video at 36 hurts on the control frequency at 18. One intuition you could have for when we get this robotic stuff is that it takes quite a few years to go from the compelling demo to actually being able to use in the real world. So 10 years ago, you had really compelling demos of self-driving, but only now we have robot taxi and weymon all these services scaling up. Shouldn't this make one pessimistic on, say, household robots? Because we don't even quite have the compelling demos yet of, say, the really advanced hand. Well, we've been working on the humana robots now for a while. So I guess spend five or six years or something like that. And a bunch of things that we're done for the car are applicable to the robot. So we'll use the same Tesla AI chips in the robot as the car. We'll use the same basic principles. It's very much the same AI. You've got many more degrees of freedom for a robot than you do for a car. But really, just think of it as a boot stream. AI is really mostly a compression and correlation of two boot streams. So for video, you've got to do a tremendous amount of compression. And you've got to do the compression just right. You've got to compress the things that don't matter. You don't care about the details of the leaves on the tree on the side of the road. You've got to care a lot about the road signs and the traffic lights and the pedestrians. And even with someone in another car is looking at you or not looking at you. Some of these details matter a lot. So if it is a session, it's got to turn that. Well, the car is going to turn that one and a half gigabytes a second, ultimately into two kilobytes a second of control outputs. So many stages of compression. So all those stages right, and then correlate those to the correct control outputs. And the robot has to do essentially the same thing. Anything about what humans, this is what happens with humans. We really are photons in controls out. So that is the vast majority of your life has been vision photons in and then motor controls out. And naively, it seems like between humanoid robots and cars, the fundamental actuators in a car are like how you turn, how you accelerate, et cetera. We're in a robot, especially with mineral war arms. There's dozens and dozens of these degrees of freedom. And then especially with Tesla, you had this advantage of like you had millions and millions of hours of human demo data collected from just the car being out there where like you can't equivalently just deploy the optimuses that don't work and then get the data that way. Between the increased degrees of freedom and far sparser data. Yes. How will you use the sort of Tesla engine of intelligence to train the optimist mind? Now, you're actually highlighting an important limitation and difference between cars. Like we do have, we'll soon have like 10 million cars in the road. And so it's hard to duplicate that like massive training flywheel. For the robot, what we're going to need to do is build a lot of robots and put them in kind of like an optimist academy so they can do self plate in reality. So we're actually we're actually pulling that out. So we're going to have at least 10,000 optimist robots, maybe 20 or 30,000 that can do that they're doing self plate and testing different tasks. And then the Tesla has quite a good reality generator like a physics accurate reality generated that we rate rate this for the cars. We'll do the same thing for the robots and actually have done that for the robots. So, so you have, you know, a few tens of thousands of human robots doing different tasks. And then you've got you can do millions of simulated robots in the simulated world. And you use the tens of thousands of robots in the real world to close the simulation to reality gap, close the some trivial gap. How do you think about the synergies between XAI and optimists, given you're highlighting, look, you need this world model, you maybe want to use some really smart intelligence as the control plane. And so maybe Groc is like doing the slower planning and like the motor policy is the lower level. Yeah, what will the sort of synergy between these things be? Yeah, so you'd use Groc would orchestrate the behavior of the optimist robots. So, let's say you wanted to build a factory. The then optimist of the Groc could organize the optimist robots, give them sign them tasks to build the factory for produce whatever you want. Don't you need to merge XAI and Tesla then because these things end up so. What are we always saying only about the company discussions? We're one we're going to send the line. Well, what are you waiting to see before you say we want to manufacture 100,000 optimists? Is it like. Optimize. So since we're defining the proponown, we could define the plural of the proponown too. So we're going to proponown the plural and so it's optimized. Is there something on the hardware side you want to see? Do you want to see better actuators or is it just you want the software to be better? What are we waiting for before we get like mass manufacturing of Gen 3? No, we're moving towards that. We're moving forward with the mass manufacturing factory. But using current hardware is good enough that you just wanted to play as many as possible now. I mean, it's very hard to scale up production. But I think Optimus 3 is the right version of the robot to produce maybe something on the order of a million units a year. I mean, you'd want to go to Optimus 4 before you went to 10 million units a year. Okay, but you can do a million a year at Optimus 3. Yeah, I mean, it's very hard to spool up manufacturing. So like manufacturing, like the output per unit time is always supposed to be an S code. So it's also agonizingly slow, then it has the sort of eventually exponential increase, then a linear, then a logarithmic outcome until you eventually ask them to add some number. Or if it's Optimus initial production will be it's going to be a it's going to be a stretched out S code because so much of what goes into Optimus is brand new. There's not an existing supply chain. As I mentioned, the actuators like trying to everything in the office robot is designed for physics first principles. It's not it's not taken from a catalog. These are custom designed everything literally everything. I don't think it does a single thing. How far down does that go. I mean, I guess we're not making custom capacitors yet, maybe. But there's nothing you can pick out of a catalog at any price. So so it just means that the the Optimus S code. The units per output pre unit time. How many office robots do you make per per day, whatever is is going to initially wrap slower than a product where you have an existing supply chain. But it will get to a million when you see these Chinese humanoids like unity or whatever sells humanoids for like 6k or 13k. Do you just like are you hoping to get your Optimus's bill materials below that price so you can do the same thing or do you just think qualitatively they're not the same thing. Like what do you think is going to like what allows it. What allows them to sell for so low and can we match that. Well, our office is designed to have a lot of intelligence. And to have the same like Electro Mechanical Dexterity if not higher than a human. So the energy does not have that and it's also. I mean, it's it's quite a it's quite a big robot. It's it's it's it has to do. You know, carry heavy objects for long periods of time and not overheat or exceed the power of its actuators. So. So we've got we've got you know, it's it's 511, you know, so it's pretty tall. And it's it's got a lot of intelligence. So it's going to be more expensive than a small robot that is not intelligent. But more capable. Yeah. Not a lot more. I mean, like the thing is over time. As Optimus Robocivil, Optimus Robocivil, it's the the cost will drop very quickly. And well, what will these first billion Optimus is Optimize? Yeah. Do like what will bottle their highest and best use be? I think that you start off with with simple tasks that you can count on them doing well. But in the home or in factories, like the best useful robots in the beginning will be anything any continuous operation. It's only 24 by seven operation because then you're because they can work continuously. Well, what fraction of the work in a degree of factory that is currently done by humans could agenda three do. I'm not sure maybe it's like 10 20%. Very more. I don't know if it's it. We would we would use we would not like reduce our head count. We would we would for sure increase our head count to be clear. Right. But we would increase our output. So the the units produced per human, like total to total human's at tensile increase. But the the output of robots and cars will increase disproportionate like much much to. You know, like the number of cars in robots produced per human will increase dramatically. But the number of humans will increase as well. We're talking about Chinese manufacturing a bunch here. And we're also talking about, you know, we've talked about some of the policies that are relevant. You mentioned the the solar tariffs. Yeah. And you think they're about idea because, you know, we can't scale up solar in the US. Well, just electricity output in the US needs to scale up. Right. We can. But that's like good resources. You just need to get it somehow. Yeah. But I was going with this is if you were in charge, if you were setting all the policies, what else would you change? So you changed the solar tariffs as well. Yeah. I would say anything that is limiting factor for electricity needs to be addressed provided. It's not like right about for the environment. So presumably some permissing reforms and stuff as well will be in there. Yeah. There's a fair bit of permitting reforms that are happening. A lot of the permitting is state-based. But anything better. But this demonstration is good at moving permitting roadblocks. And I'm not saying all tariffs are bad. I mean, sometimes if another country is subsidizing the output of something, then you have to have countervailing tariffs to protect domestic industry against subsidies. But I know the country. What else would you change? I don't know if there's that much that the government can actually do. One thing I was wondering is it seems like the, for the policy goal of creating a lease for the US versus China. It seems like the export bans have actually been quite impactful. For China's not producing leading edge chips and the export bans really bite there. China's not producing leading edge turbine engines. And similarly, there's a bunch of export bans that are relevant there on some of the metallurgy. Should there be more export bans? Like did you think about things like when there aren't out of the drone industry and things like that? But is that something there should be considered? Well, I think it's important to appreciate that in most areas, China is very advanced in manufacturing. There's only a few areas where it is not. China is a manufacturing powerhouse next level. It's very impressive. If you take refining of ore, I'd say roughly China does twice as much ore refining of an average as the rest of the world combined. And I think there's some areas like say refining gallium, which goes into solar cells. I think they're like 98% of gallium refining. So China is actually very advanced in manufacturing in most areas. It seems like there is discomfort with this supply chain dependence. And yes, nothing is really happening on it. It depends on the gallium refining that you're saying. Yeah, there's a. Well, the rare earth stuff. Yeah, rare earth, which are not rare. Like we actually do rare earth ore mining in the US, send the rock, put it on a train, and then put it on a boat to China. Because another train then goes to the rare earth refining refineries in China, who then refine it, put it into a magnet, put it into a motor service and then set it back to America. So we're really missing a lot of ore refining in America. Isn't this worth a policy intervention? Yes. Well, I think there are some things being done on that front. But we kind of need optimists, frankly, to build ore refineries. So you think the main advantage of China has is the abundance of skilled labor. And that's the thing Optimus fix is. But also we need. But there's this concern, if you think human beings are the future, that right now, if it's the skilled laborers for manufacturing, that's determining who can build more humanoids. China has more of those, and manufacturers for humanoids. Therefore, it gets the optimal future first. It just keeps that spiritual going. It seems like you're sort of pointing out that sort of getting to a million optimists requires the manufacturing that the optimists are supposed to help us get to. Right. You can close that recursive loop pretty quickly. With a small number of optimists. Yeah. So you close the recursive loop to help the robots build the robots. And then we can try to get to tens of millions of units a year. Maybe if you start getting to hundreds of millions of units a year, you're going to be the most competitive country by far. We definitely can't win with just humans because China has four times of population. And frankly, America's been running for so long that we, you know, just like a post-sports scene that's been running for a very long time, tend to get complacent and entitled. And that's why they stop winning because it's, you know, don't work as hard anymore. So I think, frankly, just observation is the average work ethic in China is higher than in the US. So it's not just that there's four times of population, but the amount of work that people put in this hire. So you can, like, you can try to rearrange the humans, but you're still one quarter of the, you know, assuming that the productivity is the same, which I think actually might not be. China might have an advantage on productivity, but person. We will do one quarter of the amount of things as China. So we can't win on the human front. And our birth rate has been low for a long time. So our birth rate has been, the US birth rate has been below replacement since roughly 1971. So we've got a lot of people retiring or, you know, more people dying than, than, than, than. We're close to sort of more people domestically dying than being born. So we definitely can't win on human front, but we might have a shot at the rework front. Are there other things that you have wanted to manufacture in the past, but they've been too labor intensive or too expensive that now you can come back to and say, oh, we can finally do the. Whatever, because we have optimists. Yeah, I think we'd like to do more, both more, or fineries, Tesla. So we just completed construction and have begun lithium refining it without lithium refinery and hope is Christy, Texas. We have a nickel refinery, which is called the cathode that's here in Austin. And these are the largest, this is the largest cathode, this largest cathode refinery, largest lithium refinery, largest nickel and lithium refinery outside of China. And it's like that, you know, the cathode team would say like, we have the largest and the only actually cathode refinery in America. Not just the largest, but it's also the only. So it was pretty big, even there's the only one. But I mean, there are other things that, you know, you could do a lot more refineries and help the help America be more competitive on refining capacity. So there's like, there's basically a lot of work for the optimists to do that that most Americans, very few Americans frankly want to do. I mean, I've actually sort of finding work to dirty here. It's not, it's actually, no, we don't, there's not, we don't have toxic commissions from the refinery or anything. So the cathode, nickel refinery short, right, sort of in Travis County, like five minutes from toe, why can't you do with humans? No, you can't, you find out humans. I see. Okay. Yeah. Like, no matter what you do, you have one quarter of the number of humans in America, yeah, I'm trying it. So if you have them do this thing, I can't do the other thing. So then, well, how do you, how do you build this refinery refinery capacity? Well, you could do it with the optimist and not many, not very many, not very many Americans are, are planning to do refining. I mean, how many of you are on it too? Not a few. What are you planning to refine? You know, BYD is reaching Tesla production or sales in quantity. What do you think happens in global markets? Is Chinese production and EVs skills up? Well, China is extremely competitive in manufacturing. So, I think there's going to be a massive flood of Chinese vehicles. And basically what's manufacturing of things? I mean, as it is, as I said, China is probably just twice as much refining as the rest of the world can buy out. So, if you go, you know, if you just go down to like fourth and fifth tier supply chain stuff, like the base cell we've got energy, then you've got mining and refining. There's those, those foundation layers are like said, China, as a rough guess, Chinese are quite as much refining as the rest of the world combined. So, any given thing is going to have Chinese content because Chinese don't quite as much refining work as the rest of the world. And then they'll go all the way to the finished product with the cars in China as a powerhouse. I mean, I think this year China will exceed three times US electricity output. Like electricity output is a reasonable proxy for, you know, for the economy. So, like, you know, to run the factories and run, run everything, you need electricity. So electricity is a, it's a good proxy for the, for the real economy. And so if China is, if China passes three times US electricity output, it means that it's industrial capacity. That's a rough approximation is three times that will be three times out of the US. Reading between the lines, it sounds like what you're sort of saying is absence and sort of humanoid recursive miracle in the next few years on the sort of like whole manufacturing energy raw materials, chain, like China will just dominate whether it comes to like AI or manufacturing EVs or manufacturing humanoids. In the absence of, of breakthrough innovations in the US, China will utterly dominate. Interesting. Yes. We're about exping the main breakthrough innovation. If you do like to scale AI in space, like basically need space, you need the humanoid robots, you need real world AI, you need a million tons of your tool, but like let's just say like if we get the master driver on the moon going with everything. Then I think we'll have solved all our problems. You can finally be satisfied. You've done something. Yes. You have the master driver on the moon. I just want to see that thing now for sure. Was that out of some sci-fi or version? Well, actually, there is a Highland book. The moon is a horse. Okay, but that's slightly different. That's a gravity slingshaws or? No, they have a master driver on the moon. But they use that to attack Earth, so maybe it's not the greatest. Well, they just add to their independence. Exactly. What are your plans for the master driver on the moon? They're sort of their independence. Earth government disagreed and they love things until Earth government agreed. That book is a hoos. I found that book much better than his other one that everyone reads. Strangerous, strange land. Yeah. But I much prefer it. Yeah, the first two thirds of Stranger's trend lines are good. And then it gets very weird. Yeah. But this is some good concepts in there. Yeah. Labelbox can get you robotics and our all data at scale. Take robotics. Let's say you need 100,000 hours of egosentric video. Labelbox starts by helping you define your ideal data distribution. Like, for example, maybe no single task category should occupy more than 1% of trading volume. And at least 10% of trajectories should capture failure and recovery states. Next, Labelbox decides this distribution to its massive network of operators. You're not limited to the small range of scenes that you can set up in a single warehouse. Instead, each one of Labelbox's operators has access to lots of unique physical environments where they can film themselves completing a wide variety of tasks. Labelbox's tech automatically categorizes each video so that their operators always know which tasks to remain and what they need to work on next. For our all data, Labelbox takes a similar approach. They work with you to understand the right distribution of tasks. And then their subject matter experts build the hyper realistic digital environments and rubrics that you need to collect the highest quality trading data. So whether you're trading robots in the real world or agents for computer use, Labelbox can help. Go to labelbox.com/sparkash to learn more. One thing we were discussing a lot is kind of your system for managing people. Like, you interviewed the first few thousands of SpaceX employees and some of the lots of other companies. What is it? It doesn't scale. Well, yes, but what clock doesn't scale? Me. Sure, sure. I know that. But like, what are you looking for? I mean, literally, it's not enough hours a day. It's impossible. But what are you looking for that someone else who's good at interviewing and hiring people? What's the genocic law? Well, at this point, I think I've got. I might have more training data on evaluating technical talents, especially, but soundable kinds, I suppose, but technical talent, especially. Given that I've done so many technical interviews and then seen the results. Technically, I've seen the results. So my training set is very is enormous and as a very wide range. Generally, the thing I ask for are bullet points for evidence of exceptional ability. So these things can be pretty off the wall. It doesn't need to be in the domain, the specific domain, but evidence of exceptional ability. So if somebody can excite even one thing, but they say three things, where you go, wow, wow, wow, then that's a good sign. Well, why do you have to be the one to determine that person? No, I don't. I can't believe it's impossible. I mean, total, it had a count across all companies, 200,000 people. Right. But in the early days, how was it that you were looking for that couldn't be delegated in those interviews? Well, I guess I need to vote my training set. It's not like I had about 1,000 here. I would make mistakes. But then I'll be able to see where I thought somebody would work out well, and then why did they not work out well? And what can I do to, I guess, RL myself to, in the future, have a better batting average when interviewing people. So my batting average is still not perfect, but it's very high. What are some surprising reasons people don't work out? Surprising reasons. They don't understand technical domain, et cetera, et cetera. But you've got the long tail now of like, I was really excited about this person. It didn't work out. Curious what that happens. Yeah, so the, I mean, generally what I tell people, or tell myself, I guess, aspirationally, is don't look at the resume, just believe your interaction. So if the resume may seem very impressive, and it's like, well, you know, where's my looks good? But if the conversation after 20 minutes is that conversation is not well, you should believe the conversation, not the right, not the paper. I feel like part of your method is that, you know, there was this meme in the media a few years back about Tesla being a revolving door of executive talent. Whereas actually, I think when you look at us, Teslas had a very consistent and internally promoted executive bench over the past few years, and then it's SpaceX, you know, all these folks like Marc Trinkosa and Steve Davis. Steve Davis runs a foreign company. No, no, yeah, but the Riley and folks like that. And it feels like part of has worked well is having very capable technical deputies. What do all of those people have in common? Well, so the, I mean, it tells us this sort of senior team at this point, it's probably got average tenure of 10 or 12 years. It's quite a lot of tenure. Yeah. So, but there are times when Tesla went through extremely rapid, an extremely rapid growth phase. And so it was somewhat things were just somewhat sped up. And when a company, as you know, if company goes through different orders of magnitude of size, you know, people that who could help manage say a 50 person company versus a 500 person company versus a 5,000 person company versus a 50,000 person company. Yeah, it's just not the same team. It's not always the same team. So if a company is growing very rapidly, the rate of which executive positions will change will also be proportionate to the, the, the, the rapidity of the growth, generally. Then, Tesla had a further challenge where when, when Tesla had very successful periods, we would be relentlessly recruited from, like relentlessly. Like when Apple had their electric car program, they were copied bombing Tesla with recruiting calls. It was engineers just unplugged their phones. Like it's just, it's just, I'm trying to get worked on here. Yeah, it's like, you know, one more call from Apple recruiter. But they were, they were, they're opening off with that any interview with me, like, double compensation at Tesla. So, so, so, so we had a bit of the Tesla pixie dust thing where it's like, oh, if you hired a Tesla executive, you're suddenly you're going to, everything's going to be successful. And, and I fall in prey to the pixie dust, you know, thing as well, where it's like, oh, we'll hire someone from Google or Apple and they'll be immediately successful, but not that best out how it works. You know, people of people, it's, it does not like magical pixie dust. So, when we had the pixie dust problem, we would get relentlessly recruited. And, and then also being Tesla being engineering, especially being primarily in Silicon Valley. It's, it's easier for people to just, like, they don't have to change their life very much. They can just get, you know, they're curious going to be the same. So, how do you prevent that? How do you prevent the pixie dust effect for everyone's trying to coach all your people? I don't think we can, I don't think as much we can do to, to stop it. But that's like, that's one of the reasons why Tesla, but they're really being in Silicon Valley and, and having the pixie dust thing at the same time, meant that there was just a very, very aggressive recruitment. Only being an Austin helps them. Austin, yeah, it still helps. I mean, Tesla still has a majority of its engineering in California. So, the, you know, for getting engineers to move, I called the significant, significant other problem. Yes. So, the others have jobs. Yeah, yeah, exactly. So, for star-based, that was particularly difficult. Yes. Since the odds of, you know, finding an honest, very successful job. Brands in Texas, but pretty low, yeah. Yeah. Yeah, it's quite difficult. I mean, it's like a technology monastery. Yes, thanks. You know, remote and mostly dudes. But again, if you go back. Exactly. If you go back. Exactly. If you go back. Exactly. But if you go back to these people who've really been very effective in a technical capacity at Tesla at SpaceX and those sorts of places, what do you think they have in common other than, like, is it just that they're very sharp on the, you know, walk-a-try or the, you know, the technical foundations, or do you think it's something organizational? It's something about their ability to work with you. Is this their ability to, like, be, you know, flexible, but not too flexible? Well, that makes a good sparring partner for you. I don't think it was a sparring partner. I mean, if somebody gets things done, I love them, and if they don't, I. So, it's pretty straightforward. It's not like some idiosyncratic thing. If somebody executes well, I'm a huge fan, and if they don't, I'm not. But it's not about mapping to my idiosyncratic preferences, or certainly try not to have it be mapping to my idiosyncratic preferences. So, yeah. Yeah, but generally, I think it's a good idea to hire for talent and drive and trustworthiness. And I think goodness of heart is important. I'd wait at that at one point. So, like, are they a good person? Trustworthiness is a smart, talented and hardworking. If so, you can add domain knowledge. But those fundamental traits, those fundamental properties, you cannot change. So, most of the people who are at Tezans and SpaceX, they're not comfortable in the aerospace industry or the order industry. What is most said to change about your management style as your companies have scaled from 100 to 1000 to 10,000 people? You're known for this very micromanagement, just getting into the details of things. Nano management, please. People have management. Same to a management. So, you're saying, you're going to go all the way down to Place Monster. We're going to go all the way down to Heismoges on some different small. Yeah, well, how do you -- I mean, are you still able to get into details as much as you want, would your companies be more successful if they were smaller, like, how do you think about that? Well, because I have fixed my time in the day, my time is necessarily diluted as things grow, and as a span of activity increases. So, you know, it's impossible for me to actually be a micromanagement, because that would imply I have some, like, thousands of hours per day. It is a logical impossibility for me to micromanage text. So, now, there are times when I will drill down into a specific issue, because that specific issue is the limiting factor on the progress of the company. But the reason for drilling into that -- that some very detailed item is because it is the limiting factor, it's not arbitrarily drilling into tiny things, and, like I said, obviously, from a time standpoint, it is physically impossible for you arbitrarily going to tiny things that don't matter, and that would result in failure. But sometimes, the tiny things are decisive in victory. The famously you switched the starship design from composites to steel. And you made that decision, like, that wasn't -- people were going around. They were like, "Oh, we found something better about us." Like, that was you encouraging people against some resistance. Can you tell us how you came to this whole composite steel switch? Yeah, so, desperation I'd say. Originally, yeah, we were going to make starship out of common fiber. And common fiber is pretty expensive. You know, you can generally, when you do volume production, you can get any given thing to start to approach its material cost. The problem with common fibers is that material cost is still very high. So, it's about 50 times -- particularly if you go for high-strengths, specialized common fiber, that can handle cryogenic oxygen. It's, like, called roughly 50 times the cost of steel. And at least, in theory, it would be lighter. People generally think of steel as being heavy and common fiber as being light. And for room temperature applications, you know, like, say, more of this room temperature applications like a Formula One car, static aerostructure, or any kind of aerostructure, really, is going to -- you're going to probably be better off with carbon fiber. Now, the problem is that we were trying to make this enormous rocket out of common fiber and our progress was extremely slow. And it's being picked in the first place just because it's light? Yes. Like, at first glance, like most people would think that the choice for making something light would be common fiber. Now, the thing is that we -- well, you make something very enormous at a common fiber. And then you try to have the common fiber be officially cured, mainly not a room temperature cure, because, like, you've got -- you know, sometimes you've got, like, 50 plies of a common fiber. And a common fiber is really common string and glue. And you're -- in order to have high strength, you need an autoclave, so something that can -- that's essentially high pressure oven. And if you have something that's a gigantic -- the oven's going to be bigger than the rocket. So we're trying to make the autoclave that's bigger than any autoclave that's ever existed, or do room temperature cure, which takes a long time and has issues. But the final issue is that we're just making very slow progress with carbon fiber. So -- I think the meta-question is why it had to be you who made that decision. There's many engineers on your team. Yeah, how did the team not arrive at Steel? Yeah, exactly. This is a part of a broader question, like, understanding your competitive advantage at your companies. So it was -- because we were making very slow progress with common fiber, I was like, "Okay, we've got to try something else." Or the Falcon 9, the primary airframe is made of aluminum lithium, which is a very, very good strength to weight. And actually, it has about the same -- maybe better strength to weight -- for its application than carbon fiber. But aluminum lithium is very difficult to work with in order to -- well, that you have to do something called friction-store welding, where you join the metal without entering the liquid phase. So it's kind of well that you could do that. But with a particular type of welding, you could do that. But it's very difficult to, like, say -- let's say you want to make a modification or attach something to aluminum lithium. You now have to use mechanical attachment with seals. You can't weld it on. So we want to -- I want to avoid using aluminum lithium for the primary structure for -- for Starship. And there was this very special grade of carbon fiber that had very good mass properties. So with rocket, you're really trying to maximize the center of the rocket that is propellant, minimize the mass, obviously. And it likes to be making very slow progress. And as of this rate, we're never going to get too much. So we better think of something else. I didn't want to use aluminum lithium because of the difficulty of friction-store welding, especially during that scale, it was hard enough at 3.6 meters in diameter, a little on it, nine meters or above. Then, as I said, well, what about steel? Now, I had a clue here because some of the early U.S. rockets had used very thin steel. The Atlas rockets had used a steel balloon tank. So it's not like steel had never been used before, it had actually had been used. And when you look at the material properties of stainless steel, especially if it's been -- like full-hard stainless steel, at cryogenic temperature, the strength weight is actually similar to carbon fiber. So if you look at the material properties at room temperature, it looks like the steel is going to be twice as heavy. But if you look at the material properties at cryogenic temperature of full-hard steel, stainless of particular grades, then you actually get to a similar strength weight as carbon fiber. And in the case of starship, both the fuel and the oxidizer are cryogenic. So for Falcon 9, the fuel is rocket propelled great caracene, basically like a very pure form of jet fuel, which is -- but that is roughly room temperature, although we do action. You can actually chill it slightly below, but we're chillin' it like a video. But it's not cryogenic. In fact, if we made it cryogenic, it would just turn to wax. But for starship, it's liquid methane and liquid oxygen. They are liquid at similar temperatures. So basically, almost the entire primary structure is cryogenic temperature. So then you've got 300 series stainless that's strained hardened. Because it's at almost all things cryogenic temperature actually has a similar strength weight as carbon fiber. Because 50 times less, normal material, and is very easy to work with, you can weld stainless steel outdoors. You could smoke cigar while welding stainless steel. It's very resilient. You can modify it easily. If you want to attach something, you just weld it right on. So it's very easy to work with very low cost. And like I said, at cryogenic temperature, similar strength to weight to carbon fiber. Then when you factor in that, we have a much reduced heat shield mass. Because the melting point of steel is much greater than the melting point of aluminum. It's about twice the melting point of aluminum. And you can just run the rocket bunch hotter. So especially for the show, which is coming in like a blazing meteor, it is the, you can greatly reduce the mass of the heat shield. So that so you can call it cut the mass of the windward part of the heat shield in maybe in half. And you don't need any heat shielding on the on the leeward side. So the net, if net result is actually the steel rocket weighs less than the carbon fiber rocket. Because the resin in the carbon fiber rocket starts to melt. So basically the carbon fiber and aluminum have about the same operating temperature capabilities. And where steel can operate at twice temperature. I mean, these are very rough fluctuations. People will. I want people to rocket this. Well, one of them is like people will say, oh, he said it's twice. It's actually 0.8. Let's show him that. That's what the main comment is going to be about. The point is, actually, in retrospect, we should have started with down steel in the beginning. It was down not to do steel. Okay, but to play this back to you, what I'm hearing is that steel was a riskier or less proven path other than the early US rockets versus carbon fiber was like worse, but more proven out path. And so you need to be the one to push for, hey, we're going to do this riskier path and just figured out. And so you were fighting like a sort of conservatism in a sense. That's why I initially said that the issue is that we weren't making fast enough progress. We were having trouble making even a small barrel section of the carbon fiber that didn't have wrinkles in it. At that large scale, you have to have many plies, many sort of layers of the carbon fiber. You've got to cure it and you've got to cure it in such a way that it doesn't have any wrinkles or defects. The carbon fiber is much less resilient than steel. It has much less toughness. Like stainless steel will scratch and bend and then the carbon fiber will tend to shatter. So, top of this being the area under the stress drain curve. So that you're generally going to have to do better with steel. One of those starship questions. So I visited Starbase two years ago. That was awesome. It was very cool to see in a whole bunch of ways. What I noticed was that people really took pride in the simplicity of things, where everyone has to tell you how starship is just a big soda can and you know we're hiring welders and you know if you can weld in any industrial project, you can weld here, but there's a lot of pride in the simplicity. So that's what I'm going to ask. Are things simpler? Are they complex? I think maybe just what they're trying to say is that you don't have to have like prior experience in the rocket industry to work on the Sasha. So we just need to be smart and work hard and if you trust where they can work on the rocket, they don't need prior rocket experience. Starship is the most complicated machine ever made by humans, by a long term. In what regard? Anything really. There isn't a more complex machine. I mean, I'd say that there's pretty much any project I can think of would be easier than this. And that's why no one has made a rapidly useful or nobody has ever made a fully reasonable over rocket. It's a very hot, very hot problem. I mean, many smart people have tried before, very smart people with a mass resources than they failed. So, and we haven't succeeded yet. We're, you know, talking is partially reusable, but the up to stage is not. Starship version three, I think this design that it can be fully reusable and that full reusability is what they will enable us to become a multi planet civilization. Can you say about the circles? I don't like, I said, I could. Any technical problem, even like a hydro climb or something like that, it's easier following this. We spent a lot of time on bottlenecks. Can you say what the current Starship bottlenecks are, even at the high level? I mean, trying to make it not explode generally. That old chest mass really wants to explode. All those combustion. We've had two burst was exploded on the test end. One obliterated obliterated the entire test facility. So, it only takes our one mistake. And I mean, the amount of energy contained in Starship is insane. And so, is that why it's harder than Falcon? It's because it's just more energy. It's a lot of energy technology. It's pushing the performance envelope. The Raptor 3 engine is a very, very advanced engine by far the best recognition ever made. But it desperately wants to blow up. I mean, just put this perspective here. On lift off, the rocket is generating over 100 gigawatts of power. It's 20% of yours. That's the trusty side. It's a great comparison. While not exploding. Sometimes. Sometimes. Sometimes. Yeah, so I was like, how does it not explode? There's a, you know, thousands of ways that it could explode and only one way that it doesn't. Because we want it to merely not really not explode, but fly reliably on a daily basis, like once a hour. And obviously, you know, blows up a lot. It's very difficult to maintain that launch cadence. Yes. And then I'm going to say, like, what's the single biggest remaining problem for Starship? It's having the heat shield be reusable. That's such that the no one has ever made a reusable orbital heat shield. So the, the sheet, the heat shield is going to make it through the same phase without shocking a bunch of tiles. And then it's going to come back in and also not lose a bunch of tiles or overheat the main, the main airframe. And that's hard. It's kind of fundamentally a consumable. Well, yes, but your brake pads and your car are also consumable with the last very long time. So it just needs to last very long time. But that's just trying to, I mean, we have brought the ship back and had it do a soft landing in the ocean. I've done it a few times, but it lost a lot of tiles, you know, it was not reusable without a lot of work. So even though it did land, it did, it did come to soft landing. It was, we're not have been reusable without a lot of work. And that's, so it's not really reusable in that sense. That's the biggest problem that remains is fully reusable heat shield. So, so if you want to be able to land it, refuel propellant and fly a game without, you know, you can't do this laborious inspection of their 40,000 tiles or everything. I'm curious how you drive, like when I read biographies of yours, it just, it seems like you're just able to drive the sense of urgency and drive the sense of like, this is the thing that can scale. And I'm curious why you think other organizations of your like SpaceX and Tesla are really big companies now. And you're still able to keep that culture. What goes wrong with other companies such that they're not able to do that. But like today, you said you had like a bunch of SpaceX meetings, like what is it that you're doing there that's like keeping that it's adding urgency. Well, I don't know, I guess the urgency is going to come from obviously leading the company. So my sense of urgency, I've like a maniacal sense urgency. So that maniacal sense urgency projects through the rest of the company. Is it because of consequences? They're like, you know, Elon said a crazy deadline, but if I don't get it, I know what happens to me. Is it just you're able to identify bottlenecks and get rid of them so people can move fast? Like how do you, how do you think about why your companies are able to move fast? Yeah, constantly addressing the limiting factor. So, I mean, I mean, on the deadlines front, I mean, I generally actually tried to aim for a deadline that I at least think is at the 50th percentile. So it's not, it's not like an impossible deadline, but as far as aggressive deadline, I can think of that could be achieved with 50% probability. Which means that it'll be late half the time. And there is like a law of gas expansion that applies to schedules. Like whatever given, whatever schedule you, if you said we're going to do this, something in like five years, which to me is like infinity time, it will expand to fully available schedule and it will take five years. You know, like there's like this, there's a physical limit. Like that, like physics will limit how fast you can do certain things. Like so, like scaling up May factoring that there's like, there's a rate of which you can move the atoms. And scale manufacturing. That's why you can't like instantly make, you know, a million of something, a million years a year or something. You've got a design manufacturing line. You can bring it up. You got to ride the S curve of production. So, yeah, I guess like, like, what can I say that's that's, that's actually helpful to people. I think generally, a maniacal sense of urgency is, is a, is very big deal. So, and you want to have a, you want to have a, an aggressive schedule. And then you want, and you want to figure out what the limiting factor is at any point in time and, and help the team address that limiting factor. Can you maybe talk about the, so Starlink was slowly in the works for many years. And, yeah, we talked about it all the way in the beginning of the company. Yeah. And so then there was a team you had built in Redmond. And then at one point you decided this team is just not cutting us. But again, how did you, like, it went for a few years slowly. And so, why did this, why didn't you act earlier and why did you act when you did? Like why was that the right moment of which to act? I mean, I have these very detailed engineering reviews weekly. That's, that's maybe a very unusual level of granularity. I don't know anyone who runs a company, or at least a manufacturing company that goes with the level of detail that I go into. So it's not as though, like I have a pretty good understanding of what's actually going on, because we go, we go through things in detail. And I'm a big believer in skip level meetings where the individuals, instead of having the person that reports to me, say things, it's everyone that reports to them, says something in the tech review. And and there can't be advanced preparation. So otherwise you're going to get glazed. Does that say these days? Yeah, exactly. Very Gen Z. Yeah, very Gen Z. How do parents advance? And you just call them randomly? Like, no, just go around the room. Everyone provides an update. So I mean, it's a lot of information to keep your head, because you've got to, you've got them. So if you have meetings weekly or twice weekly, you've got a snapshot of what that person said. And you can, and you can then, you know, plot the progress points. You can sort of mentally plot the points on the curve and say, are we converging to a solution or not? Or are we, you know, like, I'll take drastic action only when I conclude that success is not in the set of possible outcomes. So when I say, okay, we're not, when I finally reach the conclusion that, okay, unless drastic action is done, we have no chance of success. Then I must take drastic action. And so that's, that's, okay, if that conclusion in 2018 took drastic action and fixed the problem. How many, you know, you've got many, many companies. And in each of them, it sounds like you do this kind of deep engineering understanding of what the relevant bottlenecks are. So you can do these reviews with people. Yeah. You've been able to scale it up to five, six, seven companies within one of these companies. You have many, many different, many companies within them. What determines the maximum here? Could you have like 80 companies, maybe? No. But like, you know, you have so many already. I'm like, that's, that's already remarkable. Why this current number? Yeah, exactly. I know. So we can be barely people coming together. It depends on situation. So I actually don't, don't have regular meetings with Lauren company. So that Lauren company is sort of cruising along. Look, basically, if something is working well and making good progress, then there's no point in me spending time on it. So I actually allocate time according to where the, where the limiting factor or the problem, where are things problematic? Or where we're pushing against, like, what is holding us back? You know, I focus risk of say there was too many times the limiting factor. So, so it basically, something's got like the irony is if something's going really well, they don't see much of me. But if something's going badly, there's a lot of me. That's something we're not even badly. It's like, if something's a limiting factor. It's a limiting factor, exactly. It's not exactly going badly, but it's the thing that's, it's the thing that we need to make go faster to. And so, sometimes the limiting factor at SpaceX or Tesla, are you like talking weekly and daily with the engineer that's working on us? How does that actually work? Most things that a learning factor are weekly and some things are twice weekly. So the AI-5 chip reviews twice weekly. And so it's every Tuesday and Saturdays is the chip review. Is it open-ended and how long it goes? Technically, yes, but usually it's like two or three hours. Sometimes less. It depends on how much if they should go through. That's one thing. I'm just trying to tease out the differences here because the outcomes seem quite different. And so I think it's interesting to know what inputs are different. And it feels like the corporate world, one like you're saying, just the CEO doing engineering reviews does not always happen. Despite the fact that that is what the company is doing. But then time is often pretty finely sliced and half our meetings or even 15 minute meetings. It seems like you hold more open-ended. We're talking about it until we figure it out. Sometimes. Yeah. Yeah, sometimes. I mean, today's star show, engineering review went a bit longer because there were more topics to discuss. They're trying to figure out how to scale to a million first times to over per year is quite challenging. Can I ask a question, so you said about optimists and AI that they're going to result in double-digit growth rates within a matter of years. Oh, the economy? Yes. I think that's right. What was the point of the doge cuts if the economy is going to grow so much? Well, I think I've wasted four to not good things to have. I mean, I think in the absence of AI and robotics, we're actually totally screwed. Because the national debt is falling up like crazy. Now, our interest payments, the interest payments to national debt exceed the military budget, which is $1 trillion. So, over a trillion dollars, just the interest payments. Yeah, that was like, I was like, okay, pretty concerned about that. Maybe if I spend some time, we can slow down the bankruptcy of the United States and give us enough time for the AI and robots to help solve the national debt. It's the only thing that could solve the national debt. Like, we are 1,000% going to go bankrupt as a country and fail as a country without AI and robots. Nothing else will solve the national debt. We just need enough time to build the AI and robots to not go bankrupt before then. I guess the thing I'm curious about is, when those starts, you have this enormous ability to enact a reform. Not that enormous. But totally by your point, it's important that AI and robotics, dry product improvements, drive GDP growth. But why not just directly go after the things you were pointing out? The terror of some certain components, or rather, it's like permitting. I'm like the president. And very hard to cut things that are obvious waste and fraud, like ridiculous waste and fraud. What I discovered, that is, it's extremely difficult even to cut very obvious waste and fraud from the government. Because the government has to operate on who's complaining. If you cut off payments to fraudsters, they immediately come up with the most sympathetic sounding reasons to continue the payment. They don't say, "Please keep the fraud going." They say, "You're killing baby pandas." Meanwhile, there's no baby pandas are dying. They're just making it up. The forces are capable of coming up with extremely compelling, sort of heart-wrenching stories that are false, but nonetheless sound sympathetic. And that's what happened. And so it's like perhaps I should have known better. And I thought, wait, let's try to cut some amount of waste and fraud from the government. Maybe there shouldn't be 20 million people walked as alive in Social Security who are definitely dead at the age of 115. The oldest American is 114. So it's safe to say if somebody is 115 and walked as alive in the Social Security database, there's either a typo. Or somebody should call them and say, "We seem to have your birthday wrong." Or we need to mock you instead. One of the two things. Are you intimidated and called to get? Well, it seems like a reasonable thing. And if they say their birthday is in the future. And they have a small business administration loan and their birthday is $21.65. Maybe again, have a typo or we have fraud. So we say, "We appear to have gotten the sanctuary of your birth incorrect." Or a great platform movie. Yes. When I'm about ludicrous fraud, that's when I'm about ludicrous fraud. Were those people getting payments? Some were getting payments from Social Security. But the main fraud vector was to mock somebody as alive in Social Security and then use every other government payment system to basically to do fraud. Because what those other government payment systems would do, they would simply do an "Are you a live check to the Social Security database?" It's a bank shot. What would you estimate as the total amount of fraud from this mechanism? My guess is, and by the way, the government accountability officer has done these estimates before. I'm not the only one who's coming out of this. In fact, I think they did the GAO did an analysis of a rough estimate of fraud during the Biden administration and calculated roughly half a trillion dollars. So don't take my word for it. Taking a report issued during the Biden administration. How about that? From this Social Security mechanism? It's one of many. It's important to appreciate that the government is very ineffective at stopping fraud. Because it was a company like stopping fraud, you've got a motivation because it's affecting the earnings of your company. But the government just print more money. So you need caring and competence. And these are in short supply at the federal level. When you go to the DMV, do you think, "Wow, this is a bastion of competence." Well, now imagine it's worse than the DMV because it's a DMV that can print money. So was it not possible? At least the state level DMVs need to, the states more or less need to stay within their budget when they go bankrupt. But the federal government just print for money. Was it not possible? If there's actually half a trillion of fraud, why was it not possible to cut all that? Because when essentially we did, we actually, no. You really have to stand back and recalibrate your expectations for competence because you're operating in a world where you've got to sort of make ends meet. You've got to pay your bills, you've got to buy the microphones. Yeah, exactly. So it's not like there's a giant, largely uncarrying, monster bureaucracy. An accuracy computer that are just sending payments. Like one of the things that the board seems, and it sounds so simple, that probably will save, let's say, a hundred billion, maybe 200 billion a year. It's simply requiring that payments from the main treasury computer, which is called PAM. It's like payment accounts master or something like that. There's five trillion payments here. Requiring that any parent that goes out, have a payment appropriation code, make it mandatory, not optional. And that you have anything at all in the comment field. Because you see, you have to recalibrate how dumb things are. But you can pay also being sent out with no appropriation code, not checking back to any congressional appropriation, and no explanation. And this is why the Department of War, formerly the Department of Defense cannot pass an order because the information is literally not there. Recalibrate your expectations. I want to understand this how much trillion number, because there's an IG report in 2024. Why is it so low? Maybe, but which found that over seven years, the Social Security fraud, they estimated, was like, 70 billions over seven years, like 10 billion a year. So I'd be curious to see what the other 490 billion is. Federal government expenditure is a seven and a half trillion a year. What percentage, how competent do you think of mages? The discretionary spending there is like 15%. Yeah, but it doesn't matter. Most of the four is non discretionary. It's basically a fraudulent Medicare Medicaid, Social Security disability. There's a zillion government payments. A bunch of these payments are, in fact, their block transfers to the states. So the federal government doesn't even have the information, in a lot of cases, to even see know if this fraud. Let's consider, let's like reductio out of sodium. The government, the government is perfect and has no fraud. What is your probability estimate of that? I mean zero. Okay. So then would you say that four and a waste that the government has, is 90%. That also would be quite generous. But if it's only 90%, that means that there's 750 billion dollars a year of waste in fraud. And it's not 90%. It's not 90% effective. This seems like a strange rate of first principles. If you want a fraud in the government, just like, how much do you think there is? Anyway, so we don't have to do it live, but I'd be curious. You know a lot about fraud at a strike. If you will, of course, you try to do fraud. Yeah, but as you say, it's like a little bit of a, we've really grounded down. But it's a little bit of a different problem space because you're dealing with a much more heterogeneous set of fraud vectors here than we are. Yeah, but I mean, I mean, that's right. You have high confidence in your tri-hog. You have high confidence in high carrying. But it's still fraud is non-zero. Now imagine it's at a much bigger scale. There's much less confidence and much less carry. You know, back in PayPal, back in the day, we were trying to manage fraud down to about 1% of the payment volume. That was very difficult. Took a tremendous amount of confidence in carrying to get fraud merely to 1%. Now imagine that the organization where there's much less carry and much less confidence. It's going to be much more than 1%. How do you feel now looking back on kind of politics and doing stuff there, where it feels like, the outside and the two things have been quite impactful, one, the American pack and two, the acquisition of Twitter at the time. But also it seems like there was a bunch of heartache. So what's your grading of the whole experience? I think those things needed to be done to maximize the probability that the future is good. So politics generally is very tribal and it's very tribal. And people lose their objectivity usually with politics. They generally have trouble seeing the good on the other side or the bad on their own side. That's generally how it goes. I guess one of the things that surprised me the most is you often simply cannot reason with people. If there are one tribe or the other, they simply believe that everything that tribe does is good and anything the other political tribe does is bad. And persuading them is otherwise as almost impossible. So anyway, but I think I think overall those actions acquiring Twitter, getting trouble actually, even though it makes a lot of people angry. I think those I think those actions are good for work, good for civilization. How does it feed into the future you're excited about? American needs to be strong enough to last long enough to extend life to other planets and to get, I guess, AI and robotics to the point where we're going to show the future is good. On the other hand, if we were to descend into say communism or some situation where the state was extremely oppressive, that would mean that we might not be able to become multi planetary. The state might step out of progress and AI and robotics. How do you feel about Optimus, Groc, etc. are going to be leveraged by, and not just yours, any revenue maximizing companies products will be leveraged by the government over time. How does this concern manifest in what private companies should be willing to give governments what kinds of GAR real should, like, should, you know, should AI models be made to do whatever the government that has contracted them out to do, ask them to do. Should, like, should should, should Groc get to say, like, actually even the military wants to do X. No, the Groc will not do that. I think probably the biggest danger of AI, or maybe the biggest danger of fail for AI and robotics going wrong is government. Interesting, you know. The way people who are opposed to corporations or worried about corporations should really worry about the most about government, because government is just a corporation in the limit. It's a government, it is, it is, it is, the government is just the biggest corporation with them and awfully unviolence. So, I always find it like a strange dichotomy where people would think corporations are bad, but the government is good when the government is simply the biggest and worst corporation. But people have that dichotomy. I somehow think at the same time that government could be good for corporations bad, and this is not true. Corporations are better morality than the government. So, I actually think it's, you know, that's, that is the thing to be worried about. It's like, if the government should not, the government could potentially use AI and robotics to express the population. Like that is a serious concern. As a guy building AI and robotics, how do you, how do you like, how do you prevent that? Well, I think that like if you have a limited government, if you limit the powers of government, which is like really what the US Constitution is intended to do is intended to limit the powers of government, then then you're probably going to have a better outcome than if you have more government. So, electronics will be available to all governments, right? And not all governments. I mean, it's difficult to predict the, like I can say, like what's the end point or like what is what is many years in the future, but it's difficult to predict the sort of path along, along that way. Like, if civilisation progresses, AI will vastly exceed the sum of all human intelligence and, and they'll be far more of its than humans. Along the way, what happens? It's very difficult to predict. I mean, it seems like one thing you do is just say, you are not allowed to, whatever government acts, you're not allowed to use Optimus to do X, Y, Z, just write out like a policy. I mean, you, I think you treated recently that Grock should have a moral constitution. And one of those things could be that we, we limit what governments are allowed to do with this advanced technology. I mean, yeah, what we can do, what is what? I mean, it's technically, I mean, if the policies just pass a law, then, and they can enforce that law, then it's hard to not do that law. The, the best thing we can have is, is, is limited government where, you know, you have, you have the appropriate cross checks between the executive judicial and legislative branches. I guess the reason I'm curious about it is this like, at some point it seems like the limits will come from you, right? Like, you've got the Optimus, you've got the space GPU, you've got the, I think I'll be the boss of the government. Or you will get the, like, the, I mean, already, it's the case with SpaceX that for things that are crucial to the, like, the government really cares about getting certain satellites up in space, whatever. Like, it needs SpaceX. It is the, it is the unnecessary contractor. And you are in the process of building more and more of the technological components of the future that, that will have analogous role in different industries. And you could have this ability to, like, set some policy that, you know, suppressing classical liberalism in any way. My companies will not help in any way with that. Or, you know, some policy like that. I will do my best to ensure that anything that's within my control maximizes the good outcome for humanity. I think anything else would be short-sighted because I would say I'm part of humanity. So, I like humans. Pro-U1, Pro-U1. You mentioned that Dojo 3 will be used for space-based compute. You really read my, what I say. I don't know if you know Twitter, but I know you a lot. I don't know if you have a lot of followers. How do you discern my secrets? I personally. How do you design this chip for space? What changes? Well, I guess you want to have designs to be more radiation tolerant and run at a higher temperature. So, roughly, if you increase the operating temperature by 20th set in degrees Kelvin, you can cut your radiator mass in half. So, running at a higher temperature is helpful in space. There's various things you can do for shielding of the memory. But neural nets are going to be very resilient to bedflips. So, most of what happens for radiation is random bedflips. But if you've got a multi-trolling parameter model and you get a few bedflips, it doesn't matter. It's much like curiosity programs are going to be much more sensitive to bedflips than some giant parameter file. So, I've just designed it run hot. And I think it pretty much do the same way that you do things on Earth, apart from making it run hotter. I mean, the solar array is most of the weight on the satellite. Is there a way to make the GPUs even more power dense than what Nvidia and TPUs and et cetera are planning on doing that would be especially privileged in the space space role? Well, I mean, the basic math is, if you can do about a kilowatt per reticle and then you'd need 100 million full reticle trips to do 100 gigawatts. So, depending on what your yield assumptions are, that tells you how many shifts you need to make. If you're going to have 100 gigawatts per hour, you need 100 million chips running, that are running a kilowatt sustained output per reticle. 100 million shifts depends on, yeah, if you look at the die size of something like black, old GPUs or something and how many can get out of the wafer, you can get on the order of dozens or less per wafer. So, basically, this is a world where if we're putting that out every single year, millions away for a month, that's the plan with tariff app. Millions away for a month of advanced process notes. It could be some number north of a million, I think. You're going to do the memory, too. Yeah. You're going to make a memory, five. I think the tariff app's going to do memory. It's going to do logic memory and texture. I'm very curious how somebody gets start. This is the most complicated thing man has ever made. And obviously, if anybody's up to the task, you're up to the task. So, you realize this is a bottleneck and you go to your engineers and like, what is the next, what are you telling to do? I want a million wafers a month in 2030. What is the next, like, what do you, do you call ASMR? What is the next step? That's so much to ask. We make a little fab and see what happens, make our mistakes at a small scale, and then make a big one. Is a little fab done? No, it's ours. I've done it. I mean, George, we're not going to keep that cat in the bag. They're just going to come out of the bag room. They'll be like, drones hovering over the bloody thing. You know, you'll see it's construction for us on X, right? In real time. So, no, we, I mean, like, I don't know, we could just flounder and failure. There is like not success is not guaranteed, but since we want to try to make, you know, something like a hundred million. We need, we need, we want a hundred gigawatts of power and a hundred chips that can take a hundred gigawatts. And it's a cool, you know, but yeah, by, by 2030. So then it will take as many chips as I was applying us, we'll give us. I've said this to, I've actually said this to TSMC and Samsung and my colonists, like, please build your more fabs faster. And we will guarantee you to buy the output of those fabs. So that they're already like moving as fast as they, as they can, like it's, it's not like, to be clear, it's not like us to, you know, it's not like either it's, it's not like. It's us plus them, you know, there's an idea that the people doing AI want a very large number of, you know, chips as quickly as possible. And then many of the input suppliers, the fabs, but also, you know, the turbine on the factures are not ramping up production very quickly. Yeah, the explanation here is that they're dispositionally conservative, you know, their Taiwanese or German as the story maybe. And they just like don't believe they say, like is that really the explanation or is there something else? Well, I mean, it's the reason what's like if somebody's been in say the computer memory business for 30 or 40 years, they've seen cycles. They've seen like boom and bust like 10 times. Yeah. So like, that's a lot of layers of scar tissue, you know, so it's like, it's like during the boom times looks like everything is going to be great forever. And then the crash happens and then like desperately trying to avoid bankruptcy. And then there's another boom and another crash. Are there other ideas you think others should go pursue that you're not for whatever reasons right now. I mean, there are a few companies that are saying like new ways of doing chips. But they're just not scaling fast. I mean, within AI, I mean, just generally. I mean, like people should just should do the thing that where they find that they're highly motivated to do that thing. As opposed to, you know, something something of some idea that I suggest. But they should do the thing that they find personally interesting and motivating to do. But you know, going back to the limiting factor, you know, use that phrase about a hundred times. The current limiting factor that I see in the timeframe, you know, in the sort of 20, 29 20 like in the in the three, three to four year timeframe, it's chips. In the one year timeframe, it's it's energy power production electricity. It's not clear to me that there's enough that use electricity to turn on all the chips that are being made. Towards the end of this year, I think we're going to have real trouble turning on the chip outward will exceed the ability to turn chips on. What's your plan to deal with that world? Well, we're trying to accelerate electricity production. I guess that's that's maybe one of the reasons that. I will be maybe the leader hopefully leader is that we'll be able to turn on more chips than other people can turn on faster. Because we're we're we're good at hardware. And and and and and generally the the innovations from the corporations that mess that call themselves labs. The ideas have to flow like it's it's weird to see that there's like more than about a six month difference between I like the ideas. I traveled back and forth with the people so. So I think you sort of hit the hardware wall and. And then whatever whichever company can scale hardware, the fastest will be the leader. And so I think actually I will be able to scale hardware the fastest and therefore wants likely will be the leader. You you you jokes are you know. Or self-conscious about you know using the limiting factor phrase again. But I actually think there's something deep here. And if you look at a lot of things we've touched on. Of course, but maybe kind of a good note to end on. Like, if you think of a senescent low agency company. It would have some bottleneck and not really be doing anything about us. You know, market reason had the line of most people are willing to endure any amount of chronic pain to avoid acute pain. And I feel like a lot of the cases we're talking about are just leaning into the acute pain, whatever it is. It's like, okay, we got to figure out how to you know work with steel or we got to figure out how to run the chips in space. Or like we'll take some near term acute pain to actually solve the bottleneck. And so that's going to be unifying theme. I will have the international. That's helpful solve the bottlenecks. Yes. So. You know one thing I can say is like. I think the future is going to be very interesting. And. And as I said, the dollars have only been to especially dollars, I think it was on the ground for like three hours or something. It's better to be it's better to earn the side of optimism and be wrong than earn the side of pessimism and be right for quality of life. So, you know, your happiness will be you'll be happier if you are on the side of optimism rather than earning the side of pessimism. And so I recommend earning on the side of optimism. That's that. Cool. Thanks for doing this. Thank you. Great stamina. Hopefully this encounter is a pain in the pain tolerance. Hey everybody, I hope you enjoyed that episode. If you did, the most helpful thing you can do is just share it with other people who you think might enjoy. It's also helpful if you leave a rating or comment on whatever platform you're listening on. If you're interested in sponsoring the podcast, you can reach out at twerkash.com/advertise. Otherwise, I'll see you on the next one.