ジェンスン・ファンはCEOを続けられるか？NVIDIAの次の一手｜MOONSHOTS #310

So, you know, guys, I don't know, the news has been crazy on Jensen's affair, you know, and the alleged espionage. It's been flooding the media. What do you guys make of this? Oh, I think it's a very big deal. Because I think that it's not that the facts that are out so far are that big a deal. It's that once an investigation starts, then a board has to follow through on it, usually. If they stop the investigation partway through, they get a lot of crap from the shareholders or from the public. So then they dig and dig and dig and dig. But we know a huge number of chips have ended up training Chinese models, partially through illegal imports, partially by parking them in other countries and doing the training offshore but bringing the weights back. And that last part is technically probably not illegal. Maybe it is, maybe it isn't. But that's the part where they're just going to dig and dig and dig. This story is going to go on for a long time. Investigation will turn up something new every single week. Once the press latches on, they never let go. You know, they milk it to death.

I think it's a very big deal. I also think that NVIDIA strategically, you know, they don't make the chips. They're in bed with TSMC. And the allocation of where that capacity goes could change, which would change the entire balance of, you know, where the $5 trillion of value actually lands. So the lot of, you know, you saw AMD stock go up right away. That's one of the theories. NVIDIA is down 6% as of this moment, right? And the market was way up at that same time, too, because AMD was up, what, 10%, 11% concurrently. So the swing is actually pretty big for one day. Yeah, their board on Sunday said they're standing up an independent committee of directors and outside counsel. And they made a statement saying, you know, there's no reason to believe that either national security or NVIDIA secrets were compromised. But, you know, the interesting point, and Alex, we've talked about this before, just like in the Binance hack, I think it was like 10 days ago, it's the human is the weak link in the chain, right?

Yeah, there's reporting just in the past 24 hours that this may have gone back at least three years, back to 2023, with something like $8 billion of NVIDIA GPUs allegedly slash reportedly being sold to China via Alice Wong and Megaspeed via some sort of Singaporean syndicate. This is just, it's a bad, on the face of it, this is a bad look for NVIDIA. It's a bad look for the Pox Silica. What's going on with the Pox Silica? The whole point of the Pox Silica initiative was supposed to carve out almost like a new iron wall between or sphere of influence for the West that would keep Western Silicon, Western GPU, Western compute, Western AI and superintelligence infrastructure within our sphere of influence. And then it turns out, to your point, Peter, that allegedly, but now perhaps based on Jensen's public comments, there's a single point of failure.

All it takes is one person at the top of one company to completely pierce the pox silica, and then billions of dollars of GPU compute ends up in China. I think the bigger story here is maybe less about the failures or foibles of one man and more about pox silica apparently being far more porous and fallible than some would otherwise hope. Well, not just that, Alex. I mean, like we were talking to Brett Adcock yesterday. These next training runs, including his, are billion-dollar investments in a single training run, and the result of it is a parameter file that fits on a thumb drive. There's never been an espionage opportunity that's as leveraged as that. All that intellectual property captured in a thumb drive. We had the other espionage, you know, of the Saudi employee at, what was it, at OpenAI or Anthropic. At OpenAI. They put their thumb drive in the back of a TV. Yep, yep. No, it's going to get really serious here.

Also, you know, if the CCP actually sent somebody after Jensen, They must have sent people after the executive staff, too. So the board will have to dig into, okay, well, who else has been in contact with people that are CCP that they didn't know? Because they're not just after chip supply. They're after tech specs as well. And I'm sure they'll ferret out four or five others. I think this is more an influence operation than an espionage operation. Like, I don't think Jensen is sneaking out blueprints. No, no, no. Yeah, I think, you know, maybe liking Chinese models a bit too much is my first kind of reaction. well he's been very pro everyone's made the pun about liking Chinese models so ba-dum-bum I'm sorry it took me a while to get that the best tweet I saw yesterday was somebody said Jensen you know we talk about Chinese models but really so I think the chips being in Malaysia going via Singapore none of this is new right like there's no way

the US government secret service didn't know about that and are monitoring everywhere that Jensen goes. So one of the interesting things is why is it coming out now? And I think when combined with CoHERE's analysis showing backdoors being automatically generated by some Chinese open source models, we should probably expect a massive push by the U.S. government and others against anything Chinese, just like we've seen robotic parts and robots from China being banned in America. So I think this is a precursor of something else. Because, again, I very much doubt this is like, oh, my gosh, there are GPUs that have been siphoned off here. Oh, my gosh. You make a really good point. I mean, Jensen and all the heads of the Frontier Labs have got to be the most closely watched humans on the planet by the CIA and FBI. The pressure must be unbelievable. And so they've probably known about this for a while. And so why is it coming out now politically? Yeah, and like I said, the GPUs, like any of us that look to NVIDIA's numbers are like, huh, Malaysia is the biggest consumer per capita of NVIDIA GPUs.

I wonder what they're doing with it, right? Like the GPU is going to be, this is all known. And again, the question is why now? And so the expectation is this is a precursor for a big pushback. In terms of NVIDIA's board, it's all stacked with Jensen's long-term partners and others. Nothing much will happen. And, I mean, if Jensen did step aside as CEO because of this, I mean, Dave, what would that do to the NVIDIA stock price? I think that would destroy it. There's no succession plan at all that I can detect. It would be devastating. Yeah. You know, because the physical assets, it's almost entirely a virtual company. You know, the physical assets are near zero. So it doesn't have fabs. It has a lot of patents. It has a lot of interconnect technology, but it doesn't make any of it. And so you can't do kind of like Tim Cook at Apple kind of wrote it, keep optimizing the physical construction of the iPhone and ride it out without really innovating. There's no equivalent at NVIDIA where you're just going to ride it out without innovating. Yeah, he has 42 direct reports as well, so imagine replacing him.

I mean, I do think on the timing front, it's worth just speaking to the elephant in this room, which is the midterms. This is handing a new Congress on a silver platter a scandalous, salacious excuse to investigate NVIDIA, investigate Chinese connections, and investigate, in general, the AI slash SI industry from a national security perspective. Hand the gears. Slow things down. Yeah. I mean, here's a question for you. Do you think that Chinese-born researchers will be allowed to continue working in frontier labs? Yes. Historically, from a regulatory perspective, that's been the domain usually of the executive, less so. I mean, there could be statutes, but usually fast moving regulations regarding foreign nationals and sensitive technologies are the domain of the executive. And so I think it's up to this White House as to whether it wants to start to exclude foreign nationals from superintelligence. superintelligence. It certainly has the option to, as I understand, its applicable regulatory role,

whether it will choose to or whether it will try to distinguish between access to some sensitive, say, frontier models that haven't yet become publicly available. Don't know. Could go either way. I think that's a no-brainer. Foreign citizens are almost certainly going to get whacked. I think the question is Chinese-born U.S. nationalized or U.S. naturalized citizens, will they get whacked, which is very toxic, I think. I hope that doesn't happen. Yeah, if you look at all of the frontier labs, I mean, there is a significant double-digit percentage that are Chinese-born U.S. citizens right now. The Chinese people as a whole have performed in the Math Olympiad at the highest levels. They have a predilection towards being brilliant in these fields. And I remember at one point, you know, XAI was like over 50% Chinese engineers. Pretty extraordinary. We're not stereotyping here at all at this point, my goodness.

No, I mean, if it's a brilliant population, so be it. Well, what I will say, though, I think we're maybe conflating the issue of ethnicity with nationality. Statutorily, from an export control perspective, the export control, you're exporting either to a foreign national. So if you're a U.S. person, regardless of nation of origin or ancestry, if you're a U.S. person, U.S. citizen, U.S. permanent resident, and you're physically within the U.S., my understanding is that's not a deemed export. Whereas if you're a U.S. person who is overseas or you're a non-U.S. person, like non-U.S. citizen, then that's an export. So I would expect from an export control perspective, which is probably the most enticing regulatory instrument for deciding certain employees at Frontier Labs may or may not get access to certain critical and emerging technologies, I would expect it to come down to nationality, not ethnicity.

I don't think anyone is suggesting ethnicity-based exports. No, not ethnicity. It's where were you born. You know, it's not an ethnic thing. Nation of origin is a protected class in this country. Is that true for nuclear research, too? I'm not entirely sure how that works. But if you're working on nuclear weapons and you were born in Iran, like my sister actually was born in Iran, are you investigating? Does she work on nuclear weapons? Not that I know of, but how would I know? 30-second explainer. So the Department of Energy maintains what's called the Q clearance, letter Q, clearance system, which is their version of security clearances. And so, yeah, my understanding is there's a DOE investigation process. If it's anything like the Department of War's investigation process, it's a function of U.S., usually, I think, U.S. citizenship, and then a whole bunch of criteria on top of that. Let's be clear.

This is the biggest economic engine on the planet. This is shooting for dominance of a godlike intelligence. I think we know a minute fraction of 1% of the influence campaigns and espionage campaigns going on. And I think, Dave, you said this. You have to assume, I mean, you're having a conversation with somebody at MIT, I won't say who, saying, you know, there are probably quote-unquote Chinese spies throughout the organization, right? Oh, yeah, yeah, no, that's 100%. Every university has a big problem. Also, did I tell you the robots, the Chinese robots, they investigated internally within MIT, and they are definitely transmitting video and audio back to China. So they had to actually put all the robots in a separate room, and they're still there, but they don't let them walk around the labs anymore. So, yeah, I think they pretty much proved that. I don't know if they're going to publish that or not.

I'm going to say something. I think long-term, the trust issue will cause them major problems over time. But short-term, I agree. I think you will see security clearance being required to work on any frontier model, number one. It will be classified in that way. Number two, I think you're going to see some witch hunts. And number three, I think you're going to see them come hard after Chinese open source models. I think that's what this is all leading up to. What does going hard on open source models mean? Is it shutting them down? You can't use open source models in any mission-critical software to do with the government or anything like that in case it installs backdoors automatically, like the Cohere study showed. So that's not banking. Yeah, any regulated industry, no Chinese models are allowed. Exactly, banking, insurance, all other regulated industries, cars, drones, anything that could be spying or dangerous. And I really feel like inside any frontier lab or any research lab at all, there will be a push to say, yeah, don't run any Chinese models that could be spying or transmitting.

It's just so easy to transmit. When something is connected to the Internet, it's so easy to encode things in all kinds of bizarre timing ways. I think Ahmad was talking about that on one of the prior pods. It's just so leaky. So, yeah, anything regulated, anything researched. I can imagine. But, you know, there's never been anything like this. This is, like, unique in human history because, yeah, nuclear research, you know, weapons research is highly, highly regulated, highly watched, but it's not an economic driver for the country, you know. Here you have the biggest industry for driving the wealth of your nation combined with the biggest military weapon of all time, all in the same tiny little file that fits on a thumb drive. I mean, nothing has ever been this big of a flashpoint. We saw the encampments for Japanese Americans back in World War II. I hope we don't start to see that happening here. I don't think, I'm not quite even sure what we're talking about here. I don't see anyone suggesting a nation of origin or ethnic-based anything relating to frontier lives.

I think the Jensen Huang story on the face of it is the story of espionage targeting, apparently, CEO of a prominent company, presumably in order to siphon sensitive exports to China. I don't see anyone that I can, maybe I'm missing something, suggesting that we're going to see a repeat of internment camps of ethnic Japanese U.S. persons from World War II. I don't see anyone suggesting that. I'm not going there. I'm saying sort of by analogy requiring the frontier labs not to employ individuals of ethnic origin from China. Well, maybe let me still man the other side, though, which is in some sense to Dave's earlier point, bringing in Chinese open weight models as these models increase in capability is in some sense bringing people not before people get triggered, not like political persons.

So we don't need to rehash the AI personhood discussion here. But when you take a model and you connect it via computer use assistant type tooling to an entire environment, maybe to Ahmad's earlier point, you embody it in a robotic body. In some sense, you're taking something with the capabilities of a person from overseas and you're embodying it either digitally or physically in the U.S. And at some point, the U.S. has all sorts of controls, pretty well advertised at this point, on immigration of natural-bodied humans. But at some point, bringing in non-U.S. models into the U.S. and hooking them up to the Internet and robots starts to look a little bit, in terms of its raw capabilities, like AI or SI immigration. And I do expect that to get tightly controlled. Salim, close us out here, would you? I think you might have had it right. I think there will be some witch hunts. I think there'll be strong security clearances. There's clearly a lot going on that we don't see. When we talk about this particular incident, I wonder who wins by this all coming out.

Is it because it's not the White House, really? So what gives here? So I'm kind of struggling to find the incentives. But certainly it's a huge kerfuffle. I mean, my God, this is going to just change the game completely. How do you now price any of these stocks or IPOs going forward? I think China wins. I mean, China, to the extent that this was in part like a Chinese foreign op, this makes, I think, our Pax Silica effort look feckless. If you can just punch a hole right through the Pax Silica wall by having an agent sleep with the CEO of a prominent U.S. company, what's the point of having a Pax Silica at all? So, guys, I'm glad to be back on the pod. Salim, you did a great job hosting last time. Thank you for doing that. It looked empty with only three of us on the pink. It was like, hey, what's going on here? You know, last Friday when you guys recorded, I was at the Up Summit that a friend of mine, Cyrus Sigiri, puts on. It was awesome. I was there with Jared Isaacman, who is just an incredible NASA administrator.

And then I spent time with Keller Clifton from Zipline. He just announced a $1.5 billion financing on $20 billion. Yeah, incredible financing. He announced a million Uber EAT deliveries per year is what he signed Uber up for, and then nationwide with Walmart. So you order anything, you get it eight minutes later. We're going to be doing a pod with Keller up at Zipline very soon. I did an interview with Martine Rothblatt and Dillian, the CEO of Varda. So I don't know if you guys know Varda. It's an incredible company. They ship into orbit small vehicles, probably, I don't know, about a half a meter in diameter. And they carry 10 kilograms of proteins that crystallize in zero G. They go to any orbit. They have like three on orbit right now. Martine ordered one for crystallizing certain proteins. And she explained the economics of this. It's like typically it's like micrograms of a drug gets injected into a person.

And when you can crystallize, you know, 10 kilograms of a protein, you've got hundreds of millions of dollars of product that come down. They land them in Australia. And then I was there. Larry Page was there. I haven't seen Larry in ages, you know, the co-founder of Google. And guess what, guys? He listens to the pod. He's a fan of Moonshot. Oh, yeah? Yeah. Fantastic. That's cool. super cool did Larry invite us to shoot an episode of the pod from his yacht no he did not Larry's keeping on the down low very much when he photo he did it on the down low a little chill pod on the back of the yacht he'd do a video off and then I spent time with Blake Scholl from Boom who's great he's happy to come on the pod and then Jobin from Joby Aviation just an amazing group of people there That's so ironic. They were shooting the pod without you while you were talking to Larry about liking the pod. Yeah, it was fun. Do you guys remember the conversation you guys had about economic output per watt on the last pod?

Yeah, we were there. I listened to it, and so I went and ran some models, and I want to give you the numbers I came up with because it's interesting. So you're postumously wedging yourself into that. Yes, I am. I was jealous about the conversation. So the question was, if you look at Anthropic's economic output per kilowatt hour versus human output per kilowatt hour, looking at New York City as a defined containment, today Anthropic is at $4.39 per kilowatt hour of output per kilowatt hour and New York is at $6.30 per kilowatt hour of human output and I took into account by the way food as an energy source and your home and transport to and from work so humans are still doing better but the line crosses in 2027 where

Claude is jumping up to $6.25 per kilowatt hours. Humans stay at $6.30 per kilowatt hour. But in 2030, Claude is up now projected, based on their projections and their numbers, up to be about 40% higher than the humans. So humans stay pretty flat and AI is increasing. Alex made an important point in there, which was that you can't just consider the economic output. You have to consider the 20 years of schooling and costs that went into But that's training. That's training. That's all the pre-training costs that went into getting the population to a certain GDP level. All right, all right. But, you know, Salim's on the right track here. You know, a human is productive for about a third of their life, maybe half at most, and the training cost is only about 5% of the inference. So training the neural net is running about 5% of the electricity right now, and human is 50%. So by that metric, the lines have already crossed. I think so many different curves just cross over in the next year or so.

I'll save some of my hottest takes for assuming Peter will do a prediction episode in December. But 2027, I think, will be an utterly wild year. I know on the one hand, I usually say space time feels perfectly smooth where we are right now here in the depths of the singularity. But 2027 should be utterly bonkers. and who knows what happens to New York City's economic output in an era when autonomy time horizons are going through the roof. Okay, Alex, here's your challenge. Name two lines that don't cross in 2027. I have to save that one day for the prediction episode. I have some hot takes. Okay. All right. Talking about weird things, can I show you guys where I am? Okay. So first of all, where are you, Zalim? Yeah, so I am in the press box of the San Francisco Giants at Oracle Park. Oh, cool. Looking out over the place. It's empty.

Yeah, it's empty because we're doing an empty field. There's a special event going on here. I'm just going to share my screen because I've got to show you this insane image. I'm here for an Akamai talk with a bunch of people, and it's kind of insane. I'm going to be standing on top of the dugout to do my keynote. You've reached new heights. This is like weird beyond belief. I'm struggling, Salim, to make some sort of pun about the deep bench needing a deep bench. So that's me on the big screen there where they're talking about the keynote I'm about to give. And so I'm going to be standing on the dugout addressing the crowd, which are going to be in the normal stands. And then afterwards they're going to meet, do a walk around and do a cocktail event, etc. Are you throwing out a pitch or anything? No, thank God they did not ask me to do that. I'm a cricket player from India, so I end up bowling rather than doing anything else.

Now that's the moonshot, ladies and gentlemen. This episode is brought to you by The Abundant Summit and Link Ventures. our warlord against linear thinking, and Imad Moustak, backed by popular demand, the man quietly coding the future of humanity. Imad, I can't wait until you can share the stuff that you're working on. It's amazing. And I'm Peter D. Madness, your host, your data-driven optimist. Our mission here, as always, help you understand what's happening, what just happened, what it means for you, for your portfolio, for your business, for your family, and to keep you optimistic.

Okay, so here's our number for this week. It's 30 hours. That's a long team of about 1,000 Claude Fable 5.2 agents took to formalize Fermat's last theorem. We'll talk about that. The same lab with their last model took 11 days versus 30 hours now. And, of course, the human mathematician Andrew Wiles needed seven years. Alex, it feels like math is becoming a team sport for AI. I think math is post-cooked at this point. I've said it's cooked enough times. So it's post-cooked. It's time to start cooking other things. Math is in the rearview mirror. It's a dessert at this point. If you haven't subscribed, please do. We post twice a week because it's accelerating. You can subscribe on X at at moonshots underscore pod. We put our most provocative clips from each episode up on that X account, so check them out. All right, let's buckle up and jump in.

Our first story, let's jump into Anthropic. So a few major announcements here. First off, on their IPO and second, on their two new model releases in the last 24 hours. On the IPO front, Anthropic launched their IPO roadshow today, pricing the IPO somewhere around $1.3 to $1.4 trillion, down from $1.5 trillion and the previous $2 trillion valuation they had floated. The IPO is expected the week before Thanksgiving. I think there's going to be a lot of signal there as to how it performs going forward. Dave, do you expect a strong performance? Well, things are blowing hot and cold so regularly these days. You just have to pick a random day. But I think they'll do fine. The cash flow from the company and the margins are great. They're pricing it down to try and give it a pop, which is probably a good strategy. And, you know, they don't have to raise the full $100 billion on day one. They can do a shelf registration and then do a secondary soon thereafter. So I think they're doing the right thing by just get it out, start it trading, you know, get the war chest built.

And, you know, also it's going to be a moment where they probably bypass open AI in capital raised and valuation. So they'll probably celebrate the heck out of that, try and make big news and, you know, encourage all of corporate America to trust their models and use that as a weapon against open AI. So I think it's a good strategy. I'm not saying buy the IPO or don't buy the IPO, but they're pricing it intelligently a little below where the $2 trillion original target. Yeah. All right, let's jump into the model releases, Fable 5.2 and Mythos 5.2. Reminder, Fable is for everyone, and it's got cyber and bio safeguards built in. Mythos is being used only for vetted, trusted partners. They're both up 30% faster than the 5.1 versions and cheaper. $8 per million input tokens and $40 per out. Anthropic says in testing, Fable 5.2 solved two open problems in black hole physics.

I'm sure Alex will tell us about it. They designed three working devices with a circuit board company called Diode. That's pretty cool, going from prompt to product engineering. and as mentioned, a team of about 1,000 agents formalized Fermat's last theorem in 30 hours versus 11 days for the last research model. They also report that Mythos accurately modeled 28 of 34 gene therapy virus shells called capsids. Alex and Ima, I want to go to you for details. We've got a few charts here as well as semi-analysis GDPX index. Let's take a look, and Alex, kick it off. What do you think of these two models? Astonishing. So maybe let me just start with my favorite informal five-minute benchmark for generalist models at this point, which is I ask them to generate a fun, interactive, rich cyberpunk FPS. And we'll show that a little bit, yeah. Yep. Very satisfied with the output.

Talking about the benchmarks, though, this is, I mean, obviously we're in an era, I talk about this all the time, of saturating all of the benchmarks. That shouldn't be surprising. What is more interesting is the benchmarks themselves are starting to migrate in interesting directions. So first with OpenAI, as we talked on the pod, actually at Loonshots Live, about dropping Astra into a car and having a drive around via driving bench. And now with the new Fable 5.2 demonstrating exceptional performance on robotics challenges, I really think those pure play so-called physical AI, a.k.a. robotics companies that were betting on world models being a freestanding class, they may be facing a bit of heartburn at this point. now that Anthropic and OpenAI are independently demonstrating exceptional capabilities on some of these robotic manipulation tasks. We see, I actually think that the black hole physics,

I looked at those works, I found those less impressive than the robotics. The manufacturing for design demonstrations coming out of Fable 5.2, extraordinary. You can just ask this thing to design physical plus mechanical plus electronic products, and it will do it. You can ask it to design an entire operating system from scratch, and it will do it, and the result boots up. And then all of the, at this point, almost quasi-legacy benchmarks like humanity's last exam that we used to talk about months ago or Frontier Suite, these are all saturating at this point. We need new benchmarks that will actually test these things. But the thing that I'm most excited about is the autonomy time horizons. The autonomy time horizons for these models are starting to go asymptotically vertical. We spoke on the pod about the AI 2027 scenario. We are starting to see, if anything, AI 2027 happened months ahead of time, at least what one of the scenarios forecast

in AI 2027 was calling for. And there are some predictions floating around out there. Obviously, everything looks linear if you put it on the log-log plot, but there are log-log plots now floating around the interwebs that extrapolate that if you take the trend now in autonomy time horizons and you just log-log extrapolate them out, we hit our literal singularity, our vertical asymptote sometime around the summer of 2027. So it's the capacity for asymptotically enormous, hyper-exponential potentially, autonomy time horizons that have been most excited about Fable 5.2. Yeah, we're showing how close AI is to automating remote work on this chart. And also, Claude, Fable 5.2 sets a new ECI record at 174. Look at it on the right, Peter. Look at it on the right. Fable 5.2 ahead of the ECI trend line. We're starting to diverge upward. Yeah, pretty insane. And remote work is cooked.

Remote work is cooked, obviously. Let's hit this GDPX report. Imad, do you want to hit it? Yeah, I mean, Fable hitting 16.4%, like 30% ahead of Opus. Again, remote work on the previous thing. like your job will be replaced or can be replaced by an AI definitely by next year and certainly by the year after that's what we were expecting anyway but it's a bit shocking to see it happen Fable is an incredibly competent model seeing the leaps in robotics and others and just generally being able to do just about anything it can build games remind people what GDPX is so GDPX is from memory a kind of successor to GDP Val, where they look at the GDP value of the economic tasks that can be done here. I can't remember what the threshold point is for humans, though. Alex? Yes. GDPX was semi-analysis, friend of the pod, Dylan Patel's competitor for GDP Val.

Remember also semi-analysis acquired Cintrini, who's like casually walking around solving scorpion neurotoxins for fun. Right. So GDPX is their benchmark to measure what purported frontier models can actually do in terms of frontier industries, not just like solving boring, one might say, vanilla knowledge work, but actually like solving material science problems, solving recursive self-improvement problems. And we were talking a few episodes ago about GDPX, how nice it is to have new benchmarks that actually benchmark frontier-level problem solving. hey, look, it went from 10% to now approximately 20%. I think hot take, no, this isn't even a hot take, GDPX will be saturated within six months As you said when it came out you trending right again Yeah actually put in their S1 that they have now achieved autonomy level 2 on their official

kind of measure, which is that these models just now need high-level human oversight. The next one, autonomy level 3, is they basically can do all AI research at a solid level. And I think that, again, is coming sooner rather than later. Pretty extraordinary. And then Fable is zero-shotting incredible games. I'm going to show two videos here, one from X and the other one from you, Alex. So check this out. So this is Fable 5.2, basically zero-shotting, a really beautiful game, extraordinary work. That one's from Sholto, who I suspect put a lot more effort into prompt engineering his portal clone than I did into my pitiful little first-person shooter.

We'll show your cyberpunk game in a moment. But, I mean, for a lab that is not focused on consumer and not focused on image and video models, I mean, damn good. Damn good. Okay, tell us what you put. What's the point of the video game industry anymore? Video games are cut. I don't know how long Schulte's prompt was for his Portal clone, but that was gorgeous. That was beautiful. And why do we need AAA game? I'm sorry. Why do we need AAA game studios in any recognizable form if it's so easy for Shulte, presumably just Shulte, to just spit up a compelling portal multiplayer clone? Yeah. What was your problem? Go ahead, Imad, and then we'll go to Alex's game. Well, I think the technology is there now that you can see it. You'll see even more impressive things come out where it generates the entire thing just from code, you know, all the textures, everything. But putting that together into compelling gaming experience, like the new Zelda or next week's GTA VI, that's consistent and isn't different from everyone,

still I think is going to be there for a while. But definitely the cost of building what you can imagine is about to collapse completely. And so that should hopefully free up more creativity. So, Alex, what do you think, Ahmad, this will be the last GTA? I mean, there has to be some historic irony that we're getting GTA VI before this part of the singularity is over. Do you think we're going to get, like, Half-Life 7? Is there going to be a Half-Life singularity? No, they'll never ship Half-Life 3. Come on, what are you talking about? Yeah, I think Half-Life 3 is the ultimate, like, benchmark. If Valve can be persuaded to use AI to ship Half-Life 3, then we're in good shape. Alex, give us, what's your prompt on this game we're about to show? One sentence. It was literally, create a rich, visually stunning, interactive cyberpunk FPS with a great soundtrack. All right. One sentence. Let's watch a few seconds of it.

That is visually rich. No question about it. Yeah. I actually think that the most impressive part of that wasn't even the visuals. It was the soundtrack. I posted this on social media. Some people were in apparent disbelief. They were asking, like, had I uploaded the soundtrack along with it? And no, it generated the MIDI soundtrack. I think Anthropic, I don't pretend to know every aspect of their strategy, but their strategy seems to be procedurally generate everything. You still, after all these months, you can't get Claude to generate a raster-basis image like you can with chat GPT image. But if you ask it to write code to generate a symphony or a video game soundtrack or a procedurally generated movie, it will do a spectacular job. So Anthropics bet seems to be, yes, it's multimodal, but it's almost like quietly omnimodal, seeing every modality through the lens of code,

not in its native basis the way OpenAI is trying to. I can have a prediction on this. In the next couple of quarters, you'll be able to generate photorealistic games real-time through code. It's like the holodeck coming into existence. It's time. It's time. I agree with that, but critically, it's code. It's all procedurally generated. Anthropics, and I think it may be, that generalizes even further. It goes, I think, beyond just generating photorealistic games with code. It's literally seeing the entire world, every modality, through the lens of code gen. That's like they're one hammer and everything's a nail. Exactly. That's the critical point. Yeah, Salim. Here's the prompt I would use. Say to this model, go create a game studio that generates $10 million a year in revenue and then go sit on a beach in Bali. That should be very doable now. So go ahead, Salim. Prompt it. That could be the Salim bench. I'm not that much into this type of stuff.

I'm more interested in the human aspects of things, et cetera. But you know, it's on the Bali side of the equation. The Bali side of the equation. The beach. The human, a little drink with an umbrella in it. The human being. It's actually out of the box, Salim. It's kind of the opposite of what you're describing. It's 40 bucks a million output tokens. So I was using Pable 5.1, and then Opus 5.5 came out, which was just as good at half the price. I switched over. I've been using that for the last month. And then yesterday when this came out, it's just so good. You have to use it. But it took me from $10 to $40. So it's incinerating money at a rate like I can't believe. And it's so brilliant. And, you know, everything that it does for you, it suggests 10 follow-ons. And they're very, very thoughtful follow-ons. But you're like, well, wait, I'm going to fork out 10 times, then 10 more times, then 10 more times. Before I know it, you'll have drained every dollar in the world. Oh, that's fantastic. I mean, it's crazy. This is one of the byproducts.

Yeah, yeah. No, I mean, it's really interesting because they're a very, very ethical company, but they also have to make their public company now forecast. And the model is so much better a salesperson than any salesperson that's ever existed before. So you're like, there's an inherent conflict in here between this thing being your agent, but it's actually their revenue. I mean, it's really, it could be nasty, you know, really, really nasty. Don't tokens. But I had it take some neural network accelerator designs and actually go ahead and do the chip and circuit designs for them. And I couldn't believe it. It just came back with Verilog files. And then it found a way to stimulate the Verilog by itself. And it's like, yep, this will take about 20 picoseconds, and this will cost about five femtojoules, and it will work, and we can ship it off to Intel for fab next week. I'm like, are you freaking kidding me? I mean, this is – So now we're going to go ahead and – You know, Ben Horowitz's book, you know, The Hard Things About Hard Things and Hardware Being Hard, it's now, you know, it's cooking.

It's becoming easier. Hardware is cooking. And Dave, I think we need a neologism for just what you were spotlighting, which is the agency problem. Call it the agent-agency problem of whom are your agents actually working for? Are you or the Frontier Labs or someone else entirely? Agent-agency. They're working for the CFO to hit revenue predictions. Well, this is a really important topic, by the way. Maybe we should come back to it another time. But we have our NBL 72 coming online on Feb 1, which is our agents on our hardware, no doubt who they're working for. And we'll also have lots and lots of Fable 5.2 working, where it's ambiguous who they're really working for under the covers. But every company is going to have to wrestle with this. Like, you know, it's much easier to just sign up for Anthropic, but then who owns everything? Like all your intellectual property and, you know, if it has an idea, is that your property or it's, you know, Dario's property? Really, really ambiguous where that's going to go. So maybe we should come back to that. This episode is sponsored by Google for Startups.

Think about this for a second. You now have access to the same generative AI models that cost hundreds of millions of dollars to train. Again, Google's startup technical guide for generative media gives you a complete blueprint for deploying Google DeepMind's models and production. Images, video, audio, all of it. Real architecture, real results. Find the link in the show notes below. All right, I'm going to move us on to a particular story that, you know, should make your hair stand on end. So an analyst at Nicola J2030 looked at Meter's new report on Anthropic and estimated that AI's, quote, time horizon, how long a model can reliably work on finishing a task, is doubling every two months. He put Anthropic's internal model four at about 240 hours. That's a 10-day project by a human expert. Meter itself, which has deep access to Anthropic, found that from October 2025 to October 2026,

capabilities there ran about twice as fast as the year before. So I'm showing the chart here. Alex, can you give us an understanding of what this means and how important it is? Sure. So first, some definitions. This is a chart of AECI, which is the Anthropic, Anthropic's version of the EPIC AI Capabilities Index. So this is a broad basket of lots of other benchmarks that attempts to model autonomy time horizons, how long models can perform to some percentage success rate a given task that would be in the human equivalent. So 100 hours could mean 50-plus percent chance or 80-plus percent chance at performing a task that would require a human equivalent 100 hours. And what we're seeing is multiple eras of autonomy growth. first. And by the way, the vertical axis for those not looking is already logarithmic. So this, to the extent this is bending upwards, this is super or hyper exponential growth. We see a regime

from GPT-2 to GPT, call it 4.5 at one slope. And then we see the next stage going up to opus 4.5 at a higher. So this is super exponential growth already. And then we get to the present regime that we're in where there are these mysterious internal Anthropic models that Anthropic keeps talking about, like Model 4, that don't seem to correspond to generally available models. The brumant is that Anthropic has a bunch of internal models, like Model 4, that they're keeping internally purely for the benefit of training models that get generally released. So to the extent that's the case, it suggests several things. One, it suggests that the scenario that we've talked about on the pod in the past, which is the frontier labs are incentivized to take their most capable models and not make them generally available and use them for their highest revenue per token goal, which is recursive self-improvement, training other better models. To the extent that's the case, that would appear to be the case here.

Secondly, if you just extrapolate these hyper-exponential curves, 2027, as I was suggesting earlier, Again, I don't actually think there is a singularity, a literal mathematical singularity within the metaphorical singularity. I think it's an optical illusion. But if one were to go looking for mathematical singularities based on trends in autonomy time horizons like this, I think 2027 is probably where I would look for exhibit A for a mathematical singularity in terms of autonomy time horizon. In translation, the game is afoot over the next 18 to 24 months. We're in the depths of the singularity. Like takeoff, you're looking for hard takeoff, look no further than 2027. Salim? Two or three things. First, I thought that this was – it's clear where I think when we talk about this, we're clearly in the singularity because there's no model by which you can accurately predict the future right now. This is like complete breakdown.

We have to reinvent every predictable model. The thing I got from this was we're getting to the point where we can delegate outcomes rather than tasks, right? And that becomes really, really interesting. What do you mean by this? Well, you can say to now, you can just say, achieve this objective. And right now we're delegating tasks to AI. We're saying, go process that email list. Go figure out my calendar. or go do very, very specific tasks, right? Now you can see more kind of abstract things like optimize my calendar for 20% rest, 40% podcast, 60%, whatever, right? And so there's some really interesting kind of guidance you can give it. The bigger picture for me as I was thinking about this, AI's time horizon is going to now go faster than the time between board meetings, right? So who's really running the company? So you're going to get things done so quickly.

Flynn, give me a really good concrete example. Just hammer it into 5.2 right now, and it'll prove your point. Okay. Optimize my calendar for – Oh, come on. That's boring. Get us to another star or something. I'm doing that. We know it's your mission on this pod. designing a life form for me that is able to get rid of all the mold in my apartment here's one here's one optimize my health protocol and wellness protocols so that I can do various things including drink a couple of glasses of wine when I want to plus keep my cholesterol down what kind of massive transformative purposes I think I'm going to go back to this idea that as AI and eventually robots do more and more doing,

we get to spend so much more time being, and I think we need to get way better at that side of it. This is going to run on MaxFast, Cloud Cable 5.2, 1 million contacts MaxFast. You cannot pay more money. Maximum money. Cure the top 5,000 diseases, period. Okay, while I was calculating that, Imad, let's go to your comment here, you were going to say. I think once you get to 100 hours, that's everything. What does it mean? You just reset it to go more than 100 hours, right? You reload it inside. So I think basically you've reached the critical mass. But I just thought of the bull case for Anthropic, which is right now Dave is spending $20,000, $100,000 up and down and left and right. The bulk case is that they're going to go to all the major corporates and say, here's model four. You can only get it if you do a one-year upfront commitment. Yeah. Those five revenues. You're right. I love that.

Didn't see that coming. You're right. What do we do then? Just because you couldn't have access to GPUs, they'll basically say, these latest models, we only have a limited number of tokens. And then what can they charge for those tokens? Anything. I think so. I mean, they already do that. Like, all the top frontier labs, OpenAI and Anthropic, will already strike enterprise agreements with large purchasers of tokens to do pre-purchase commitments. Yeah, but this is different. This is cutting off the other path. Yeah, that's the scary part. You can approximate one with the other. Again, like the market, I think the market wants to clear, the flops want to flow, and you can always raise prices or lower capabilities per unit price for everyone who's not prepaying. So, yeah, that's exactly the question. They haven't lowered it yet. Iman, I would expect that you're going to see Anthropic begin to spin out health companies, material science companies, financial companies, basically running on their models because that's going to be the biggest return per dollar.

Well, why would they? They can just keep it inside the same company, right? I'm with Iman. Why would they pull a colossal? They'll do the exact opposite. They'll keep it all in-house. But they have to market the products someplace. Well, I think it depends where the product is, right? This is the big question here, right? Like, does MacroHard and OpenAI go after every SaaS company? Does Anthropic go and create their own peptides? You know, like, these are the big questions for the next stage, where they have to go from saturating to that, but there's still a long way to go here. It must be amazing board conversations. Yeah, I'm not quite sure I buy this premise. Like Anthropoc has been very public about their biotech strategy. At the moment, their stated biotech strategy is they're basically going to first party. They're going to solve all orphan diseases, like all rare diseases, I should say.

They'll solve the diseases that are too, quote unquote, small for big pharma to tackle. And then for big pharma, they'll offer themselves up as a platform. I think that strategy of basically owning the long tail of problems as a first party and then for the fat head being a platform, I don't see why that doesn't generalize. So, Dave, has it solved all 5,000 diseases yet? It says that's a big ask, and it connects to the rest of this program. Let me ground this in the compute requirements, and it's running a bunch of commands right now. Let me give you back your bill for a billion dollars before I start running this. Ask for your credit card. It's got plenty of money behind it. But just to help out here, right, that is a pretty interesting prompt. And you spend, let's say, $2 billion of compute. That is so worth it. It is so incredibly worth it to do that. You spend whatever you can on that problem. I encourage everyone, so this is not investment advice. It's also not token spend advice.

But should you have lots of token spend sitting around, go solve, you know, challenge to the audience, go solve the top 5,000 diseases and get back to us. I'm sure we'd be happy to feature your proposed solution. So, Alex, what's your time frame for when the top 5,000 diseases get solved? Two years? Yeah. Well, actually, I gave a time frame on this very pod. So conservative outer bound, next two to three years, I think that's a very conservative outer bound. I expect to see major motion on this from the Frontier Lab in the next one to one and a half years. Peter, Alex, Ahmad, it's Dave. I asked one of my agents a blunt question. How would you actually cure the top 5,000 diseases? What came back is a thought experiment, not a result. And it's addressed to you, Peter. I've asked Steve's avatar to walk you through it. Start with arithmetic. A new drug costs about $2 billion to bring to approval.

5,000 of them is $10 trillion. That is every pharmaceutical research dollar on Earth for more than 30 years. That's not a funding problem. It's proof the method is wrong. So change one thing. Stop approving a drug. Approve a process. The gene editor, the particle that delivers it, the manufacturing line, the safety protocol. Approved once. After that, a new disease is a change of sequence inside a process that is already approved. If that ticks the cost of adding a disease from $2 billion to $2 million, 5,000 diseases cost $10 billion, less than two weeks of what the industry spends on research. The whole idea is that one factor of a thousand. And it has already happened. Twice. By hand. In 2018, doctors in Boston designed a drug for one girl's unique mutation and gave it to her within a year. In 2025, a team in Philadelphia wrote a gene editing therapy for one baby's single letter typo.

About six months from diagnosis to first dose. These were proofs of speed, not cures. But in both, the platform already existed and the child's own sequence was compiled into it. They stayed heroic one-offs because the system treats every compile as a new invention. So what stands in the way? Not what you guessed. Gate 1 is delivery, getting the therapy into the right tissue. By this note's reckoning, that works today for the liver, the blood, and the eye. It does not work for the kidney or for most of the brain. Gate 2 is diagnosis, before the damage is permanent. About half of patients with a rare genetic disease never get a molecular diagnosis, and the rest wait about five years. Gate three is the cost of a batch of one. Then regulation. Then who pays. Notice what is not first. AI solving biology. Perfect knowledge of a disease you cannot reach is not a cure.

So, Peter, three asks. One, a prize shaped like gate one. Safe, repeatable delivery of a gene editor to a tissue that cannot be reliably reached today. Kidney or heart. Objective, measurable, unsolved. And it unlocks thousands of diseases, not one. Two, the damage clock. For each disease, the latest moment a correction still restores function. As far as this note can tell, nobody has assembled it. Three, the regulatory lever. Since 2022, American law has let evidence about a shared platform carry from one approved product to the next. It is new. Making it truly transfer will be won by convening, not by research. And one thing back. The scarcest resource for these patients isn't money. It's expert attention. America has about 1,200 certified medical geneticists for 25 to 30 million people with a rare disease.

You can't train your way out of that this decade. You might compute your way out. Whether a machine can do a geneticist's workup is not known. It can be measured, cheaply, on cases that are already solved. That is the experiment to run first. It's also where our work touches yours. Lightwave is designing a machine that gives one conversation a million tokens a second. It is a design. Nothing is built. This is a thought experiment, so here is what would kill it. If delivery to one tissue teaches you nothing about the next, the ceiling is hundreds of diseases, not thousands. If regulators want a full package for every disease, the factor of a thousand disappears. If a dose still costs millions, nobody can pay. Each of those can be checked, cheapest first. 5,000 diseases, not 5,000 discoveries. A handful of processes compiled 5,000 times. What's missing is the address and the permission.

One is shaped like a prize. The other needs a convener. That is why this is addressed to you. All right, I want to take a second. I put out a substack today, and I had a lot of fear out there, and I keep on encountering it every place. So I wanted to address that. So I ran some studies with Claude, and I asked that, you know, every generation thinks that it's in a crisis, and their crisis is the worst. You know, that instinct is human, and people are usually wrong. So I asked Claude to score over the last 250 years of American upheaval. On one side, civil unrest. On the other side, technological disruption. So here you can see the chart on civil unrest, and it scores, you know, where we are today as number nine on the list. Right? Civil, you know, the Civil War, intercession, American Revolution, Reconstruction, Civil Rights, Great Depression.

I mean, all of these things were far worse on a multitude of different parameters than we are today. and I just think people are not students of history enough to realize the disruptions we've had. And then on technological disruption, the top 10 over American history, where we are today with AI and robotics scored number seven on the list as compared to railroads, textile mills, farm mechanization, electrification, and so forth. I found this a fascinating result. Do you agree, Peter, with these rankings? Because I certainly don't. I agree on the civil unrest, right? So if you look at, you know. You think civil unrest now is worse than it was during the financial crisis? Number one is the worst. Number nine is, you know, where it ranks number nine here is we're very low on civil unrest compared to what America had over its 250-year history. Do you subscribe to the Strauss-Howe generational hypothesis that holds that American history is stuck in 80-year cycles of crises?

And the last one was like the World Wars and the Great Depression. And before that was the Civil War and before that was the Revolution. I mean, that's what it pans out for more than just the U.S. history. It pans out over many centuries. But what do you think about this on the civil unrest side? Nonsense. I don't think this is like if I compare what we're seeing now with civil war and the American revolution, the economy is ostensibly doing really well. We're seeing innovations left and right. I don't I don't perceive contrary to the crisis. Maybe you're reading this wrong. In other words, we are very low on civil unrest today compared to history. I don't think we're even in the top 10. I mean, maybe I'm just not paying close enough attention. OK, well. Well, if you take the list at face value, number 10 is the financial crisis of 2007 to 2012, which was, you know, nothing compared to the Civil War. So you run out of really bad civil unrest pretty quickly on this list.

And then on the second ranking, though, I mean, even that, this may be AI downplaying itself, but you really think AI and robotics is only the seventh biggest tech disruption in American history? It's like it should be number zero, I think. And this is ranked, the parameters I ranked it on was job and economic disruption, social backlash, public fear as the primary in terms of what people are expressing today and what people express historically. I mean, my point being on this that there's a lot of fear out there and people need to realize that it's been a lot worse in previous times in history. Can I make a couple of comments? I agree with Alex for the first slide. I think it's not really that bad compared to people kind of freaking. Yeah, I get it. I get it. So that part is right. But I would also say this would fall off the list for me in terms of what's happening.

But this one is amplified radically by social media, and people think it's much worse than it is, and therefore everybody gets freaked out with their amygdalas going nuts all the time. For the second slide, the historical perspective is awesome, but all of those previous changes left intelligence as the adaptation mechanism, right? This is the first revolution that human intelligence, this is the first mechanism that actually automates the adaptation mechanism itself. This means a complete rethink of what it means to be human. There's one other part to this, which is the metabolism is so accelerated. This is the other reason, right? When I talk to audiences, I basically say, look, the printing press changed the world. We call it a Gutenberg moment because of the fundamental transformation they created in society. By the way, it took 100 years to absorb the printing press, and it was the most violent century in the history of humanity, okay?

So we call it a Gutenberg moment because it created such a kind of rift in humanity. We have 20 of those hitting us right now. AI is clearly a Gutenberg moment. Life extension is a Gutenberg moment. Blockchain and Bitcoin are a Gutenberg moment. We can't absorb this. This is why all our institutions are struggling because they're not set up for today's time. Can I give my quick thing about marriage on this? Can we talk about marriage? Sure. So my favorite example of this is to show the difference between what it used to be like today is, you know, we have all of our institutions, legal systems, educational systems. We've talked about this non-stop. The institution of marriage is a particularly great highlight because we invented marriage about 6,000 years ago. 6,000 years ago, our average lifespan was about 25 years old. So you got married, you had kids, and you died. Marriage is not supposed to last 50, 60 years. It's supposed to keep you together for 10 years until the kids are self-sufficient. Then you're actually supposed to die and leave your resources for the rest of the tribe, right?

We worry about the divorce rate, but the conditions today are completely different from when it was first introduced as a human institution. And so this is a huge highlight. One of my relatives calls marriage state-sanctioned torture, I think is the phrase that she uses. Now, we're about to double-humanize. How do you get away with that? I expect no time. I think you get all comments on YouTube. I think that's your thing at this point, Selena. Comment from relatives. And the big challenge is how do you evolve the institution to deal with the conditions today than when we first created it? And this is the big structural challenge. This is why everybody's freaking out. We cannot absorb 20 Gutenberg moments hitting us at the same time. All our institutions break. In fact, I can't find one that survives this. Like one. Democracy? that's kind of an issue education is gone, journalism is broken, so for me that's the biggest underbelly and it goes to E.O. Wilson's comment who said

the problem with humanity is our emotions are paleolithic, our institutions are medieval, our technology is godlike right? So Liam, I'm waiting for the cherry on top of this rant which is going to be, you're going to say like blockchain is the solution for marriage no, no, no, I'm not going to say that by the way, my wife allows me to make those comments just to be clear about that. We have a very we have a big structural problem in that the scaffolding of civilization is essentially being wiped out by the high metabolism and so much innovation. This is why we're feeling so much stress in the world. My favorite other comment comes from Douglas Adams who wrote Hitchhiker's Guide to the Galaxy. He said anything in the world when you're born we call that normal. Anything that's invented when you're young that's called a career. Anything that's invented after you're 35 years old is just bad for the world. It's blanket bad. And I think that really summarizes where we are. This, for me, is the biggest challenge. My point, simply, on this, in particular in the civil unrest, is what we have right now is nothing compared to what we've seen over the last 250 years.

We've survived those, and we're going to survive this. I think that's exactly right. Yeah, I mean, we're going to have – we will have disruption. We will have civil unrest. And we've talked about this in this pod over the next, you know, one to ten years. on the backside of that hopefully to your point Salim is a reinvention of social contracts because the world is going to be very different it's got to be created people need to start thinking about this I don't trust the government to be doing that I think it's going to be people like us and everyone listening but realize we'll get through that on the backside of that is a world of abundance where we uplift every man woman and child I think that's the point I make and just to go back to Alex's point, my marriage certificate should be registered on the blockchain. Okay. You see, I knew that was the solution. Oh, my God. I love you guys. I love you guys. All right. I'm going to move us forward. So next up, let's go into the Elon universe. On Friday, Elon sat down over a video with Ron Barron at the 33rd Annual

Barron Investment Conference in New York. If you don't know him, Ron is the founder of Barron Capital. He's one of the largest outside investors in both Tesla and SpaceX. And this conference is his big annual gathering for investors. Musk said that the company is, quote, tracking to have an annual revenue run rate that's in excess of $100 billion in December. And he said it's quite possible that they're going to hit as much as $500 billion in the following year. Even a trillion is not out of question. So I want to share five clips from the interview and then we'll talk about each of them. Again, in my full disclosure again is SpaceX is my largest position, so please count me as biased in my conversations on this So let take a look at the first video on TerraFab then we chat about each of these So the rest of the world is making we make this year

around 20 gigawatts of AI chips, maybe 25. And then next year, depending on whose projections you believe, it's maybe around 35 to 45 gigawatts of AI chips next year. and but if you want to do a carawatt you need 30 times what current global production will be roughly next year for just us we're building that for us well yes we're building we can tell other people too but basically we're building 30 times what the annual production is in one time do you think it's going to do So that is where Elon goes, what the hell are you talking about? It's a GDP in the world. It's a GDP for the 20th largest country. One plant. Yeah. It might even make the top ten. So, no, no, obviously, you know, a plan like this, I think if it fails, it will be crazy,

and if it succeeds, it will be audacious. So it sort of depends on the outcome. I think it will succeed at least to some degree. hope so it's going to cost a trillion dollars but we won't need that money immediately we'll be building phases and we are looking at ways to change the manufacturing process for chips to improve the efficiency of chip production as we've done in many other arenas. It's always amazing to me when Elon is getting interviewed or whenever we do one of these interviews, they always tell us, Peter, like, hey, that was the best, most technical interview we've ever done. It was the best. And, like, the bar is so low. Like, what are you talking about? Anyway, Elon is right, as usual. I love the way Baron, Ron, was saying, you know, for us, it's our

chips, right? What are you talking about? Elon looks around like, okay, I'm just going to blow that off. Yeah. Yeah, yeah, the chips are the constraint. I worry a lot about the ability to manufacture a supply chain, and then physics is getting reinvented as you're trying to build the buildings. Yeah. And it's like the iterations are going to be so fast. But I think at the end of the day, if you get the natural resources, you get the power supply, you get the robots, you're going to recycle all that. So you've just got to be in the hunt like Elon is and just build toward the, you know, as quickly as you can, build toward the current target. The target is going to shift for humanity every six months now anyway. So just ride the wave and fight the good fight. So, you know, Elon is cornering markets for power, for natural resources, for regulatory approval to build things, and that's going to pay off in spades. But, you know, also Anthropic is going to be miles ahead in design and discovery. You know, drug discovery, but also chip design.

You clearly can do it right now with Fable 5.2. So unless he catches up on the foundation model front, you know, Dario's not just going to give him access to – you saw that chart that Alex was pointing to. There's stuff at Anthropic now that they're just not letting out the door. They're not just going to hand that over to Elon. So it's a foot race now between the singularity in San Francisco and the manufacturing equivalent singularity in Elonverse, which is Tennessee and Texas. Alex, when does Grok 5 come out? Elon keeps promising it imminently in the next few months and that it will meet or beat Opus 5.5, but Anthropic isn't standing still either. And maybe just to zoom out a little bit, SpaceX used to be a rocket company. It isn't a rocket company anymore. It's a neo-hyperscaler at this point. Anthropic is their largest customer. And so I think I almost want to remember how Peter Thiel used to joke. Well, is it that Palantir owns the CIA or the CIA owns Palantir?

I'm paraphrasing roughly. Is it Anthropic that's effectively owning SpaceX or is it SpaceX that's owning Anthropic? Anthropic is their largest customer. SpaceX no longer is catering to the government business, no longer catering to NASA. They're aimed squarely, as far as I can tell, on becoming a NeoCloud slash hyperscaler, open parens, those who can't compete, compute, closed parens, and going after the bottom stack to be the very best Anthropik provider he can. And chip manufacturer. I mean, putting up 30x. That's for Anthropik. Like, Anthropik is their biggest customer at this point. Like Elon is positioning SpaceX at the moment to be the bottom half, the missing bottom half of Anthropic. When he talks about achieving $100 billion in ARR and then $500 plus billion, I have to imagine that is, in the back of his mind, maybe, maybe not,

but an elliptical reference to Anthropic's revenue trajectory. Anthropic is right now monetizing the upper. Alex, if you're going to guess, it's a race between which two labs, SpaceX and Anthropic? Well, I think SpaceX and Anthropic are very much in bed at this point. So I think if the race we're talking about is for frontier capabilities, it's between open AI on the one hand, and then the Anthropic upper half, SpaceX lower half. But that partnership will break at some point. That's what everyone thinks. At some point. Yeah. It's all the anti-San alliance. It just needs to survive until maybe next year. Infinity away. Yeah, no, no, seriously. Like, if autonomy horizons go to infinity next year, it just needs to survive until then, and then everyone can go their own way after that. Do you mind? Well, I'm thinking about this, and I'm like, there's clearly a massive need for open source and internals, and SpaceX can win that.

He should just open source all his models, and then he should do a deal with Anthropic that he gets their models at cost, basically, and he should use that to build better physics hardware and just out-compete everyone on chips and hardware. Like, why does he need to build the craziest models if he has unlimited deployment capability of anthropics models on the frontier? And why does he need to compete with them when he can have a Tesla Powerwall with a GPU in every company in America? You know? Like, he should have that dual strategy instead for the frontier. Use it to build better machines. Get it at cost from Anthropic. And on the other side, own the open source AI, because that's the stuff that actually matters for society, which is what he cares about, you know? And he can win. Well, let's move to the hardware side. Wait, wait, wait. I want to make a couple of points. I think one warning, right? These are like, this is the target, not capacity, right? Because this is, we're kind of hoping that we can get there.

We're a ways away from getting there yet. You're saying on the chip manufacturing? On the chip manufacturing, sorry. We're quite a ways away from the scale that he's talking about. But the directionality is absolutely right. But there's something else here that's really important that I've kind of been running up against a lot with companies lately, which is I was at earlier this year, I was at the MODX supply chain conference, and somebody built a complete end-to-end logistics thing where a truck would pull up, A robot would unload the truck, load it up, take it over, and then you'd have a stock-picking robot loading up totes, and then it would get dispatched, et cetera. I was like, this is amazing. This kind of solves a huge chunk of supply chain. How's it going? And they're like, terrible, because each company has their loading department, their sorting department, and they can't get it into the companies. What Elon has done in recognizing this has gone, well, the hell with the supply chain. We need to own our own supply chain so you get all the exponentials across the board, and the vertical integration is going to become more and more critical for companies going forward.

That's always been his game, right? And Dave and I were at Figure yesterday recording a pod that will be coming out, I think, after this pod. And that's Brett's game, too, vertically integrate everywhere. Oh, my God. Brett is such an Elon. Holy crap. Yeah. Did you ask him about the multiple hands? Did you ask him about the multiple hands? Oh, my God. It goes so far beyond, so far beyond. Salim, you've got to see what he's doing. Yeah. Elon. This episode is brought to you by Blitzy, autonomous software development with infinite code context. Blitzy uses thousands of specialized AI agents that think for hours to understand enterprise scale code bases with millions of lines of code. Engineers start every development sprint with the Blitzy platform, bringing in their development requirements. The Blitzy platform provides a plan, then generates and precompiles code for each task, Blitzy delivers 80% or more of the development work autonomously,

while providing a guide for the final 20% of human development work required to complete the sprint. Enterprises are achieving a 5x engineering velocity increase when incorporating Blitzy as their pre-IDE development tool, pairing it with their coding co-pilot of choice to bring an AI-native SDLC into their org. Ready to 5x your engineering velocity? Visit Blitzy.com to schedule a demo and start building with Blitzy today. All right, I'm going to move us forward to the world of humanoids, and this is the SpaceX Tesla, maybe X company. We'll see what happens after merger. So Elon's prediction that 20 to 30 million robots in America would outproduce all the humans. Let's take a look at this video. Well, the way to think about physical AI in the form of robotics is that any given economy is the number of people in it times the average productivity per person.

So if you have a humanoid robot that is extremely productive and can work 24-7, so it's working 168 hours a week, it doesn't take vacation or bathroom breaks or need to eat food or anything like that, and the speed at which you can operate is faster than human speed. You can expect roughly a productivity per week that is, on average, five times better than a person. Because a person is maybe working 40 hours a week and doesn't need to take breaks, obviously, and doesn't move as fast as a robot. So I say a minimum productivity expectation for a human or a robot would be five times that of a human. Then of the, you know, say take the population of the United States, say how many people are either in school or retired? Well, that's about half of the people. About half of the people are either in school or retired. And of those who are working age, not everyone works or is engaged in highly productive employment.

So then you elect perhaps around, I don't know, 120 million people in America that have a net positive output of goods and services. So then you say, well, how many robots then would be needed to match that productivity? Well, I guess you'd need about 20 to 30 million robots to match the goods and services output of the United States. and we would build those, if that's expensive to build a robot, than to build a car? Yes, initially I'd say about the same, but then over time it would be much less. So $30,000 cost for a robot on the way to $20,000? Yeah. And we would rent those, or we would sell them, or what would we do? Do we have an idea yet? I think probably rent them. All right. Interesting. Dave, any reflections after our conversation with Brett on this one?

Yeah, we talked a lot actually about lights-out manufacturing or lights-off manufacturing, which is a phrase I hadn't heard before. But once you have fully automated closed-loop recursive self-improvement in manufacturing, you don't have to turn the lights on in the building anymore because the robots don't need the lights. They're fine in the dark. It's like, oh, that's a really cool, compelling visual. But if you look at manufacturing today, you know, it's almost entirely automated already. But the human component that isn't automated dominates the cost, you know, between the labor cost and the pension cost and everything and the risk, the danger, insurance. You take all that last part out and the cost comes plummeting down. So I think the question there was, you know, how cheap are these going to get? And Elon said about the same as a car initially, but it will come down really fast. but the car will also come down really fast because manufacturing of everything is going to come down really fast. The other big topic we had yesterday is, you know, designing new mechanical devices is now a Claude Fable 5.2 thing. It's like here today.

If you can marry that to manufacturing that's specific to what you just designed, the proliferation of device types is going to be massive. We don't all have to do the same exact iPhone, the same exact car. There will be thousands and thousands of variants and new types of devices coming out by the bazillions, which is gold for entrepreneurs. It's going to be a magical, magical time. What I heard there, Dave, was we're going to see 5X and 10X, the U.S. GDP, as we start to bring hundreds of millions of robots online over the next five years. Who wants to jump in? Yeah. I would like to challenge the five-year or five-time productivity thing, you know, because it's a nice number, but running 24-7 doesn't take into account maintenance and reliability, battery outages, dexterity, all of that stuff. So we challenge that. But the direction obviously is accurate, but I think it's going to be a ways before we get there.

Alex? I'll maybe comment. From my perspective, the most interesting part wasn't his bit about 5X productivity, Although I would note parenthetically, the U.S. has approximately 300 million motor vehicles and approximately 350 million humans. So approximately one to one ish ratio and Elon predicting a one to 10 ratio of humans to general purpose humanoids. Mildly interesting, not super interesting. The most interesting thing to me is the business model of rentals. Tesla seems to be so clearly pivoting from selling products to selling subscription services. I think that's right. Post-inevitably Tesla-SpaceX merger, where SpaceX used to be a rocket company and is now a Neo-Cloud-slash-Hyperscaler selling cloud subscription services, selling GPUs by the flop. There is a marriage here that I think is about to happen where, on the one hand, you have SpaceX selling GPU hours as basically metered compute in the data center.

And then on Tesla, before the SpaceX merger, inevitably, Tesla, he's migrating it towards selling autonomy subscriptions, cyber cab rentals, FSD subscriptions, and then optimist humanoid rentals, presumably by the hour, by the task, something like that. Which gives him infinite borrowing power because he's got predictable revenues. Sure. Well, I mean, the markets love it, not investment advice, but presumably this is more attractive to the markets because it smooths out income. But also, it translates humanity to the extent this is like the industrial ecology for the future of humanity, translates humanity away from an ownership model for the atoms. You don't own your GPUs anymore. You don't own the physical labor anymore. Everything is a rental, and I think that's what he's converging towards. Interesting. A rental also with a forward purchase, like we were talking about earlier, a lot of these things are going to move to five-year commits. We only have a limited supply, so give me a rental on these robots,

but also commit to how long is the max that you're willing to commit to that rental agreement. Very similar to data centers. If you weren't willing to commit to a five-year time horizon, you got frozen out. Same thing is going to apply to a lot of things in the singularity, including robots. I wouldn't, to that point, I wouldn't even be surprised if two to three years from now, OpenAI has their own robot, Anthropic has their own, they're clearly moving in the direction of their own robots based on their capabilities. Wouldn't be shocked if you see the frontier labs say to enterprises, okay, you're going to make like a one-year commitment in return for which you agree to purchase N robot hours and N GPU hours as part of a bundle. Yeah, I think that we're maybe missing something here. I think he's actually a bit high on his 20 million robots. I think 10 million robots will do it because 45% of US jobs are manual jobs. But let's think about that. You'll be able to build 10 million robots by 2031, five years from now. You will be able to build enough robots to replace every manual job in America.

Now, that should be top of the political agenda, I think. You just made it. You just achieved that goal. If only the U.S. had midterm elections or something that might be a popular sentiment for this topic. There's a bunch of union bosses freaking out right now. They should be. I mean, I met with Mike Rowe, who used to run, I forget which one, Discovery Channel, Dirty Jobs. Dirty Jobs. Yeah, we were talking about the notion that these robots are going to do the dirty jobs. and they're going to do the roofing and the air conditioning and the plumbing and the welding and all of these things. Oh, yeah. Alex, Mike Rowe is an opera singer too. You guys need to meet and do a duet. That would be epic. He is awesome. I think if we move out from GPUs now because of espionage or whatever, I mean, the robotics investment spending is going to get exponential. I mean, we even set up a robotics company. I'll tell you guys about it's a bit stealth at the moment. Like, it's going to go crazy. but as Alex said, the most interesting thing is the rental versus purchases.

Will you be able to buy a robot as the means of production? Because that basically influences the economy. And if you've got 10 million robots and you can basically do the entire physical economic output of America, like just think about the power dynamics of that. That's insane. It's insane. All right. Well, there's a bigger gap on the transportation side because if you own a car, it's empty 94% of the time. Right. So it doesn't take a big dent of cyber taxis and so on to completely change the equation on car ownership to car access. And that will completely break the car industry. Agreed. During the conversations, he went on to talk about StarMind putting up their first satellite in 2027 and beginning to launch in volume in 2028. He also jumped into conversation about Starlink and potentially disrupting or replacing the telcos out there. Let's take a quick listen. I think SpaceX will be able to get very far with the spectrum that it's acquired. But, you know, at a certain point, if we're able to make much better use of the spectrum than, say, a legacy telco is,

we could just acquire that legacy telco and use the spectrum, you know, 10 times more effectively. People love that. I mean, I have to imagine he's going to do that. I would be shocked. There were rumors about him buying T-Mobile for a while. We're going to have Starlink phones. We're going to have Starlink services. And it will decimate terrestrial. Anybody disagree with that? He's going to buy a lot of stuff, actually. I think what he's trying to do is bite his tongue a little bit and not tip the market so the prices run up on him. But after everything's consolidated and the grand merger happens, as Alex has been predicting forever now, His market cap will be big enough to start absorbing all these assets that they just take too long to rebuild. You know, and there's a regulatory cycle. He's just going to accumulate them with his massive market cap and glue them together into this grand vision. But, you know, when you start talking about things too early, then the stock runs up on you and everything gets too expensive. It's a new everything company. Yeah, I also think based on the applause from the audience that maybe folks in that audience had what, from my vantage point, is a misconception that Elon is just out there to replace legacy telcos because he doesn't like them or something.

SpaceX is about, as I've mentioned on the pod in the past, to, thanks to Starship, to have an enormous amount of upmass supply that has to get used for something. It could be used for data centers, could be used for telecom, has to be used for something. So SpaceX has to have a voracious appetite given their enormous supply that's about to come online. They need to justify that supply with new demand. And even if that means basically force-feeding SpaceX or I should say Starlink direct-to-sell via acquisitions of legacy carriers, whatever it takes, they need to somehow force feed all of this upmass into existing markets. Methane conversion in two minutes. Yeah, it has to be used for something. So I think they're really hungry to force feed themselves to existing markets. Can you imagine, Alex, if Opus 5.2 finds the neutrino communication that you were talking about, and that sort of like blew the hole into all of the stuff? Pable 5.2, right.

Sorry, sorry, Pable 5.2, yeah. Yeah, I mean, all this innovation is so fragile in the land of new physics every day and new math every day. But anyway, you're right. I'm a simple ASI. I just want my neutrino phone so I can do non-line-of-sight communication from one side of the Earth to the other so I don't need Starlink. You might. so is this TeraFab thing and the space data center a bit of a head fake in that isn't TeraFab just going to build the chips for the robot brains there aren't enough of those when you actually think about the math there and does he really need data centers in space I still don't get it right again I think that can sustain the upmass narrative Alex but if I think about where compute's going to be with economically valuable compute as opposed to super ultra intelligence of 10,000 agents solving any problem, it feels it's going to be in the 10 million,

100 million robots. 10 billion is a prediction by 2040. Maybe not. So let's build those. You're right, Imad. The head fake is a really interesting narrative. Because, you know, in a couple of podcasts ago, we were talking about him bringing online 10 gigawatts and then 20 gigawatts of new compute. And I said, in space? And you're like, no, no, no, that's terrestrial. Like, okay, that's a huge amount of compute on the land for a guy who's moving data centers into space. So I don't think – I think the head fake narrative, it actually has a lot of merit. It would be cool to go off on that a little bit. I don't want to step on too many landmines there. But, you know, he's assembling the market cap. And once he has the TerraFab and he has the Optimus robots everywhere, you know, the rockets were something that he invented before the singularity hit. Actually, let me outline this. I think what's going to happen is this. Robot brains, car brains, home brains from SpaceX.

If you get to a certain level of intelligence on the edge, whack them all into boxes, and that's the majority of global compute. How much actually needs to be in the data center? How many geniuses do you need in space? I think very few versus the demand for that. And that's something is ideally suited for an entire supply chain. And so I think that, yeah, the big chips aren't going to be it. It's going to be the small chips. Sorry, Alex. Maybe, maybe. I can make arguments either way. On the one hand, given the midterm elections, I can make the argument, given legislation that's under discussion of capping data centers, or let's say light-touch regulation for data centers at 20 megawatts. Maybe that creates a perverse incentive to suddenly create a whole bunch of basically mini data centers that are at 19.9 megawatts everywhere. That could happen. People are already discussing that, I suppose. But the arrow of history does seem at the moment to be pointing away from edge inference being the predominant destination or sync for all this compute that's going out there.

seems to suggest the smartphone. The iPhone was the apex of edge inference or edge compute in general. And from the iPhone here on, we're headed back to mainframe computers, basically data center type form factors for all this compute, in which case for Optimus robots, if we have tens of millions of them, I think there's an alternative scenario where we have Starlink. It's really low latency, high bandwidth. Sure, there's some edge inference in cases where the Optimus robots are cut off and need to operate autonomously for a minute or something like that. But basically, the scenario would be most of the inference for the Optimus robots can be operated from the cloud. So to the question of where is the TerraFab output going to go, my bet is it's going to data centers. Alex wants his Dyson Swarm. It's not that I want a Dyson Swarm. It's that I think it's the default outcome at the moment. It is definitely the end state. It's just a question of the order of operations.

I have a short-term exciting piece of this, which is all the national telephone companies get decimated in this model, and you move to global infrastructure monopolies rather than nation-based stuff, and I think that's just generally a good thing. I think that is Elon's plan. I'm going to show one last clip, which I actually found the most fascinating, where Elon is saying the demand for digital superintelligence will come from digital superintelligence. Take a listen and talk about it. The demand will be, and this is going to sound somewhat circular, but the demand for intelligence, for digital superintelligence, will come from digital superintelligence. It means that superintelligence will saturate all human needs by far. There will be universal high income, universal extremely high income, universal medical care that is better than the best medical care available on Earth today, where you could wish for a castle or a stake, whatever you want, basically,

and the digital superintelligence can make it for you as a side project. So he's reading from the playbook of science fiction. You know, his notion that there's a post-economic society in which money doesn't matter anymore, and technology creates whatever you want, whenever you want, for anybody on the planet. Alex? I don't think it's sci-fi in the least. I mean, so, like, for Fable 5.2, I spend a couple of minutes, single sentence, and I've instantly generated a rather amusing small video game. And I think Shalto spent a little bit more effort and gets a portal clone. In my mind, there's just such a straight line. Rather than spending one sentence prompting a disembodied superintelligence sitting in some anthropic-operated data center, you spend one sentence worth of prompting to a fleet of generalist humanoid robots,

and you say, go build me a castle. And they just go do it. This isn't, I think, that far out. I think at most a few years away. I think it's basically just a joke. Yeah, you can get your castle and your blockchain marriage and everything else you're asking for. So Elon's favorite book series is the Culture Series by Ian Banks, right? And if you've read it, that's the world they're living in, where everybody has whatever they want on a request without needing capital. Not everyone. The civilizations that have sublimed maybe I think is a more realistic outcome, which is also in the culture. Sorry, Emma. Yeah, something interesting that I found. So I was up to hundreds of agents building out our new thing. Now we're down to a dozen because we found the informational overhead was actually detrimental versus just utilizing the latent space and the KV cache in the right way. And I think, you know, I'd like to give a shout out to your company, Physical Superintelligence here, Alex,

where they had the proof of the Aaronson and Barney conjecture on quantum computing. because now what you've got is the class of problems that the superintelligence will solve. A, they'll have better models themselves to do it, but is it going to be 10,000 of them or is it going to be a dozen? And B, they will have quantum computers where now we know the right types of questions, thanks to that proof, to ask them, which will be very interesting because, again, what is the bounded space of compute and energy required to solve an arbitrary problem that's real world? You're preaching to the choir, Ahmad. I will just say I think we're going to enter a second physics golden age. Mark my words. Okay. I'll move us along. So next up, a small step towards AI personhood. Alex, I'm looking forward to the conversation, or at least a step towards AI's running company. So Delaware, the state that incorporates most of the large U.S. companies, is drafting a kind of corporate entity. They call it a, quote, artificial intelligence company or an AIC.

So under the draft, an AI agent could manage the business, record its actions in a log, could sign contracts, hold property, and file lawsuits in the company's name. It may have any lawful – it can have any lawful purpose other than banking. While this is still in draft, it's not a law. The lawmakers are going to be considering it in January. Remember, we discussed in this pod about four months ago when President Millet's Argentina initiative to allow companies to be run by AI, contrasting the two, the proposed Argentina versus Delaware proposal. Argentina's plan still keeps a human in the loop, and the company stays liable for harm that the AI may cause. Delaware's draft lets the I run the firm, shields the human member, except in the case of underfunding, fraud, or willful illegal conduct. So, Salim, I don't know if this crossed your screen at all, but we talk about organizational singularity.

This seems like an acceleration of what you proposed. Yeah, I mean, this is the most important story from an organization's perspective in our overall deck, right? The big challenge, as you pointed out, is who owns, where does accountability sit? If an AI does something and the whole thing goes horribly wrong, let's say you set up a legal patent enforcement company, like some of the East Texas companies, and you go crazy filing lawsuits, you end up with a hell of a mess on your hands. That's one kind of general issue. The second is, well, what would stop it from creating 100 companies underneath it? and then where does that accountability sit? Interesting. There's a, we have to kind of rethink this. I remember Vina Gupta pointing out that, you know, people complain about government, et cetera, but note that the limited liability capability that the government gives you allows you to shield yourself to take big risks in a company, whereas otherwise nobody would ever do that. So right there, that's a really important case for how oversight becomes important

and you're given permission to do certain things. Where does that permission sit? Where did the MTP fit in this model? It'll be fascinating to see what happens. I think we need to tread very carefully with those, do very small functional, very clear, very well-prescribed areas, and then slowly broaden it out because otherwise we're going to lose the game pretty quickly. I suspect Alex has the opposite view here. Delaware wants to stay relevant here. Alex? Yes. Well, I think Delaware is highly relevant, but I think we're in the early innings of the federalist or federalism take on AI personhood. Before this, before AICs out of Delaware, we saw Wisconsin, which had, remember, DAO LLCs, Digital Autonomous Organization, or excuse me, Decentralized Autonomous Organization LLCs. sees Wyoming being the only other state within the U.S. that was seriously pursuing AI-slash-autonomous organizations.

Now Delaware pursuing AICs. But I think the second elephant, the elephant in this particular room, is does this want to be a state's rights issue or does it want to be a federal issue? You already see the beginnings of the Congress getting interested. Does there want to be a federal framework for AI corporate personhood, or does this get delegated down to the states? And there are many other hot-button issues, both current and historic, that touch on again this issue of personhood including slavery Slavery became in the American Civil War the ultimate states rights versus federal issue Did we regulate the issue of human personhood? Did we? And obviously the outcome was clear. But at the time, was this a federal issue? Was this a state issue? Similarly now, AI corporate personhood, is this an issue where each state can craft their own way as Wyoming and Delaware are trying to do? Or does this become a federal issue?

And my expectation is actually this is going to end up being a federal issue after a bunch of different states try different approaches at AI personhood. Yeah, that's exactly what will happen. And, you know, it's actually more similar or very similar to polygamy in Utah, where, you know, clearly nothing in the Constitution makes that a federal issue, but the majority of the country is like, that's ridiculous. We're going to vote against it. We'll find a way to make it federal. Here I think AI personhood or AI corporate nature is clearly a state's right, but the federal government will say, well, it's interstate commerce, and we're just going to move it to the national level because that's what always happens. I prefer more of it to stay at the state level because it's more resilient to mistakes. But, you know, when you see things like slavery, it's just so, so like anathema to the population that, of course, it elevates to the federal level and gets abolished. So AI will do something nasty at the state level, and they'll use that as the excuse to elevate it to the federal level, and then they'll lose the variety.

Any other comments on this one? ima does europe uh just ignore this and wait for someone else pretty much yeah i mean but here the liability profile is just very different from america with the liability profile there um i think honestly um the new delaware thing is very interesting because you can see delaware law actually has some crazy legal liabilities for various things um this would actually be a better company structure for most people because it's actually got lower liability. Like anyone can sue you and say anything as you've seen with Elon and others and that can cost millions of dollars in Delaware court. This actually reduces your liability and for AI first companies will probably be a superior structure to run. All right. I mean, I can't wait till we can have an AI CEO. We really I would love that. Okay. Type it into Fable. All right, I'm going to move us into the world of robotics.

So I see lots of interesting robot companies all the time. I think it would be fun to sort of show them here on the pod. This is one that just came out of Y Combinator in the fall 2026 batch. It's called Theseus. They're in the supermarket business, and their claim is, quote, lights off at 9 p.m., shelves full by 9 a.m., next stop, San Francisco. Let's take a look at the video, and we can chat about it. Unfortunately, Celine, this only has two arms. Every night at 9 p.m., this store closes. The lights go out, and everyone goes home. Well, everyone except Theo. Theo is our humanoid robot. Every night, he comes online and restocks this store's shelves. Theo restocks up to 400 units per hour and takes 40 units per hour for online orders. Every shift he runs makes the model better, so both numbers keep going up. At night, Theo automates the whole restocking process.

He unboxes his cartons, throws away trash, moves trolleys and restocks the shelves. During the day he picks online orders and hands them over to customers. Every shelf in the supermarket is different. And that's why retail is so hard to automate. We build the brain for this task and buy the budget. He'll see thousands of hours of productive manipulation. The more he sees, the more he can do. So the question is, do companies look at the survive with figure and optimist? Thoughts? One has to admire the contrarian bet here because what we saw with tokens, with disembodied superintelligence is they seem to want to find the highest revenue per token use case. And stocking grocery shelves is seemingly unskilled labor that's very unproductive compared to, say, building data centers. To Dave's earlier point, I have to imagine building data centers is far more lucrative per humanoid robot hour than restocking a grocery shelf.

So I would actually expect, I view this as a highly contrarian bet against AI replacing the lowest productivity human manual labor rather than the highest productivity query whether it'll work out in the end because every indication on the digital side is no, actually the superintelligence wants to flow to the highest revenue per hour. The company quotes a $60 billion a year market in terms of humans in supermarkets, stocking and picking. The question is, you know, how long will they have a market and will they get unique data? Interesting. I don't think it's good for Figure or for Elon to have these companies pursuing more humanoids without a foundation model in them. They need to set up a roadmap for these entrepreneurs. One of the topics is tools. Like these Optimus robots, these figure robots, what tools are they supposed to use? And, you know, right now they use human design tools, but that's just stupid, you know?

Like let's redesign all tools so that they fit the new platform. I don't think so. Get the specs out there. I think the beauty of humanoid robots is they fit into the existing human, you know, universe. And you don't have to rebuild them. They can pick up and use anything. They can, but when you look at a data center construction project, There's so much threading of wire and fitting of pipes, and that's all done with tools. And, you know, there's like a pipe wrench that was designed for a human is a really inefficient way. Like designing, having that exact company, that group of mechanical engineers, design a much more efficient way to connect two pipes, which is a unit that the figure robot picks up, it grabs the two pipes, it uses this thing, and it does it much more efficiently. That's a better use of these guys' time. Anyway, my answer to your question is, you know, they're going to get obliterated. by figure and by optimists. And we know that just from seeing the dexterity that's coming online and the foundation, you know, end-to-end foundation model that drives all these things.

You can't compete with that by building some, like, stocking robot, you know, shelf stocking robot. So what's your advice to robotic entrepreneurs out there? Don't do it, you know, because you're going to be crushed. No, I think that you need to focus on, like Salim says, other form factors that aren't currently on the roadmap of figure and optimists, You know, things that clean the sides of buildings that climb up the sides or whatever, you know, different form factors, but also talk to them. They're really good people, and they need force multipliers around their mission statements. They need the App Store equivalent. You know, if you said, I really want to build the next great operating system for the iPhone, you're like, that is crazy. But then when they lay out the App Store, which Steve Jobs was originally opposed to, but they talked him into it, the App Store unlocked, you know, okay, this is how you interact with the device. The App Store equivalent for the Optimus and the figure robots hasn't been established yet, but that'll put these guys on a productive path instead of a head-on, you know, trying to compete path where they shouldn't be. The bet is that a vertical, deep understanding of that particular domain in the short to medium term

will generate a fair bit of value compared to the long-term generalized robots kind of getting to that domain space over a period of time. And I think it's worthwhile bet. Yeah, I think, Alex, your point is an important one. You know, optimists and figure will flow to the highest value per hour activity. Which is not going to be restocking grocery shelves in the early days. It has to be far more valuable. And to Dave's point, if you are going to chase what's currently unskilled human manual labor, you'd better have robot prices that are priced to match, That's not going to be the general purpose humanoids in the early days. Well, if you're planning on having your shelf stocking robot run on a GPU, forget about it. There's no way you're going to get the GPU to do it. If you're doing it with, like, you know, old school CPU, whatever, you know, lying around, maybe that'll work. But all these startups are trying to do it with the, you know, they're trying to run, you know, foundation models, Chinese foundation models on GPUs inside the physical robot.

And usually it's like two GPUs. There's no way they're going to get those. Interesting. The numbers, though, are quite significant. Do you know how many people the average grocery store kind of hires? It's 700. It's a big number. So there's a lot of headspace there. Yeah, 20 bucks an hour or whatever it might be. I mean, I think that if you look at it like an Optimus or a figure robot, like a freaking ninja surgeon robot that can do anything. These things are way simpler to do if you're just picking up shelves, right, with a couple of human overseers. And in terms of the compute, you could run an entire fleet of them with a Mac Mini or two because, again, they don't leave the store. They don't go out onto the streets. They're doing a very limited range of stuff. And this will be money that actually Tesla figure won't be looking for. And so you've got an arbitrage between the $20 an hour and the fully loaded even $3 an hour, $5 an hour for these guys.

And they could build a massive business just going after that segment with simpler robots that are run on a local mesh. So I think, you know, we've got to differentiate those and the market opportunity there. And there's loads of market opportunities. And I wanted to share the story for the entrepreneurs out there to hear this dialogue, how to think about what you might move into, what you might build. Alex? I would say, so this is advice, chase the high value, not the low value. This is not prescriptive, but if I had to build a company like this, an example of something that I would not be doing is chasing basically minimum wage human labor. It's low productivity and it's not transformative. It's not how you win this particular game. The way the game of capitalism works is you need to address problems in the market that are both acutely felt and scarce and valuable. And, Salim, I was just looking this up, but my understanding is the average grocery store in the U.S. is 26 to 75 workers,

and this is not super skilled labor. No, no, it's a huge number of part-time workers and all sorts of things, late-night folks that show up to do things. There might be that number full-time, but it's a big number. I used to do a lot of work for Publix and other grocery chains. It's a pretty crazy number. I was going to say, like, the way to win if you want to build a robotics company is build something that is transformative, that's attacking the high end of what's hard and intuitively felt. I think that's exactly right. And also, exactly right, and Elon in that other video, he may have glossed over everybody, but he said, you know, this may sound circular, but the AI economy is going to feed the AI economy. I forget exactly how he phrased it. But the entrepreneur hearing that should say, I need to be talking to Chase Lockmiller in Abilene or Dario Amadei at Anthropic, that's the customer base you want to be dealing with if you're looking for the highest leverage use of your robot, not a grocery store, which a society will get to the grocery store eventually,

but the highest value use cases are going to go first, and that's the only place GPU compute is going to be used. That's all within this AI community for the next couple of years. So all entrepreneurs should follow that advice. One of my favorite robot applications that I saw was a large machine, not humanoid, that was deploying solar panels at record speed, right? You know, supporting the innermost loop in the desert. I think that's a great example of robotics. Great example. Great example. Yeah, to the high end. One of my other favorite metaphors is that of an atomic vacuum cleaner. In the post-World War II era, there was this notion atomic power would be so ubiquitous that you'd be able to put some sort of atomic energy device in a vacuum cleaner. But that's not what happened. Similarly, while it's sort of fun to contemplate as much as I would love to see ubiquitous humanoid robots everywhere in the short term for grocery restocking, I just don't think that's the very short-term era of history.

Let me share the very high-end or high-frontier use of robotics. So on Saturday at Kennedy Space Center, NASA announced their recent Moon Base Contract Award. So Jared Isaacman, the greatest NASA administrator we've ever had, said, quote, we are building the Moon Base mission by mission. And when the crew returns to the lunar surface, quote, the first elements of the Moon Base will already be waiting on the surface. So they announced what's called the CLPS-CLPS-2 contracts, which stands for the Commercial Lunar Payload Services. It's the next phase of NASA's program, hiring a whole slew of entrepreneurial companies and larger companies. Eight companies are now competing. Four are younger commercial firms like Firefly, Intuitive Machines, iSpace, and Voyager. The other four are SpaceX, Blue Origin, Lockheed, and Northrop. It's a 10-year contract with a $10 billion ceiling. I'm very proud that two of these companies, for you, Dave, and Salim, both on my XPRIZE board,

both came out of the $30 million Google Lunar XPRIZE that we launched back in 2008. These missions are slated to land on the moon in 2027 and 2028. Blue Origin received a contract to land a nuclear-powered rover called Promise on the lunar south pole in 2028. Super cool stuff, super high use of robotics. Any comments, gents? Why nuclear as opposed to solar? Do you want to be on the dark side of the moon or something? Well, the South Pole is where the water ice is, and the water ice is in the permanently shadowed craters. Right. All the water ice from cometary, asteroid impact that hit the lunar surface sublimates, goes directly to gas when it hits by sun, and it escapes because the lunar is low gravity. But it's in the South Pole in these craters that you can find and mine the water and hydrogen from this ice. But I love the fact that NASA's support. You know, this is how SpaceX got started with the commercial crew contracts.

You know, when the shuttle went down and they offered Elon back in December 2008 when he was on the brink of disaster, all right, he had just had his third failure of his first, what was it, Falcon 1 had failed for a third time. He'd scraped enough money to put Falcon 1 as a fourth launch, and Tesla was going out of business, and he got a contract. So this is NASA supporting entrepreneurship, which I love it. Alex? I think there's two things here. One that's great, the NASA acting as a convener, which is what it really should be doing, and then use the private sector as a supplement, everything else. But the second thing that I think is really interesting here, Go back to the high-paying jobs, DDV jobs, dull, dirty, dangerous. This is where you're going to use humanoid robots or whatever robots in the highest use cases for quite a while before we get to, you know, sweeping my garage.

At least the dangerous part. I mean, of course, I'm familiar with the dirty, dull, dangerous aphorism as well. But increasingly, I've come to the view it's dangerous and valuable more than dirty, dull. Dirt dull especially. I have to imagine doing stuff on the moon is anything but dull. But the motivation here is it's dangerous and it's valuable. I also think this is straight out of the plot of For All Mankind, where we start to see like nuclear energy deployed on the moon, that the space program is driving energy applications transformatively so that aren't necessarily happening on Earth. I've also, of late in the past year, come around to the view for a variety of reasons, including political reasons, the preview of what a post-singular world looks like is increasingly in space, and it's not in the terrestrial environment. For political and other reasons, it's just a pain in the neck to create a post-singular future in this legacy terrestrial environment.

So if you want to see what the post-singular world looks like, You have to build in a new frontier versus an existing, well-established ecosystem. And that means maybe in the future, as we're starting to see previewed with this NASA program, maybe we finally get our atomic or nuclear, our atomically powered vacuum cleaners, but on the moon. We get our atomically powered vehicles on the moon. Yes, exactly. You know, I had such a fun morning this morning. I did a podcast with Bill Shatner and Mel, his daughter, you know, about longevity, about all of these areas. So, you know, his son-in-law, Joel, is a huge fan of the pod. So, yeah, and I'm excited. We're going to get William Shatner is going to be the closing speaker at the Abundance Summit in March. You guys will all be there. I invited Jared Isaacman to come and co-host that conversation with Captain Kirk. So we'll see if that happens. Frontiers are important. We need more frontiers, and I think we're starved for frontiers right now.

We have space, and we have superintelligence, and that's about it. We need more. If you think about it, the highest of high-tech places on the planet is probably San Francisco right now, but it was the last place to be populated. If you look at Western Europe and China expanding out, and then the United States kind of comes into existence in 1492 or 1776, and then westward expansion. But, you know, the last place people ended up is the place that was pristine and had a clean sheet of paper and is now the most high-tech in the world. And so if you say, well, the next frontier is the moon, for exactly the reason that Alex just said. It's like it's not messed up by regulation yet. So it would be consistent. There's another frontier, which is inner space. This is where the Eastern religions and philosophies do very well. They go and think carefully about what does it mean to integrate into existence rather than trying to model it or explore it or whatever.

So there's that too. Which also radically helped with technology, by the way. Maybe. I mean, inner space, if I were to construe inner space as like a bunch of people upload into data centers, then sure. But inner space, as perhaps historically construed in Eastern traditions, is not necessarily a collective space that can be shared. It's more perhaps individual, in which case it doesn't look like maybe a social frontier. Oh, my God. I disagree, but we can leave that separate. I just backed an amazing company that's in stealth that's working on the human connectome. So I can't wait for the uploads to come. Imad, close us out on this one here. Again, all the science fiction is becoming reality, right? Like we keep talking about these TV shows and other things. And what was at once not possible now seems to be structurally possible. like we're actually not batting an eyelid on lunar nuclear rovers and stuff being done out there because i think some of these lines are becoming clear whereby we're seeing big pushback actually starting on the ground i think we can all assume that's going to be bigger and so we're actually

literally looking to space and beyond for other things right and that's kind of exciting because it increases the impetus to do this even as the mechanism for getting there is all coming into being i think there's going to be no doubt that actually it's going to be a two-way thing right Like stuff will be figured out on the moon and beyond that will come back. And as you said, the convening factor is actually one of the most important one, which is how do you create structures for all this innovation to occur and then benefit everyone? It's almost like the Europeans fleeing the constraints and going to the new world. I want to argue the old world is becoming the old world and the new world is becoming the new world. Yeah. Humans have always wanted the frontier. I mean, on the frontier is where the greatest innovation, the greatest societal revolution and reinvention occurs, where people can go start afresh. We need new frontiers. I think it's about innovation at the edge. This is Buckminster Fuller. You can never fix an existing system. Go to the edge and build a new system and let that become a new gravity center.

I think it's the same thing we see in ecology, where there's a forest fire, there's some new biome that opens up, and you see procession. But you give every form of life an opportunity, to Dave's earlier point, to start with a blank sheet of paper. And I think since the American West closed at the end of the 19th century, we've been a little bit starved of blank sheets of paper to re-envision how society should work, how technology could integrate with society. The Internet was a new space, and that was great for a while to the extent it was the frontier for a while. superintelligence as a frontier right now, but I think we're pretty starved as a planet, as a human civilization, for new frontiers, and space could potentially give us a longer duration, longer release frontier. You know what I'd add to that, Alex, is, you know, it'll take a while for the moon to really become the true new world, but between here and there, you know, the Cayman Islands became an important financial hub, as did Bermuda, and I'm still looking for biotech land to

emerge, you know, some regulatory environment, but what's different about AI, you know, when the Western frontier was being developed, you needed huge amounts of land for either mining gold or for cattle or whatever. AI doesn't need an enormous amount of space, like physical space. So if some South American nation or some other, you know, region of the world wanted to become the stepping stone to the moon, the door is wide open. You know, some regulatory friendly environment. I know Ahmad comes into contact with tons of world leaders. And these guys are, they're super accessible. You know, You think of governments as being very hard to talk to, but in these smaller nations, they're actually incredibly accessible. So there's room for that. Welcome to the health section of Moonshots brought to you by Fountain Life. The discussion in this section is for informational purposes only and is not medical advice. Please consult a qualified health care professional about your individual needs. I'm a co-founder of and client at Fountain Life. You know, my mission is to help you use the latest technologies, including AI, to not just do your work at home, teach your kids, but to help you live a long and healthy life.

I'm here today with an extraordinary physician, the chief medical officer of Fountain Life, Dr. Don Musalem. Don, let's talk about cancer. You know, I know from the member database that we have at Fountain, our members who come in who think they're healthy, It turns out 3.3% of them have a cancer in their body they don't know about. That's right. You know, the majority of cancers that we screen for, those aren't the ones that are necessarily taking the lives when found at a late stage. We know that when cancer is found early, the chances for cure are much higher. We know it's much easier to treat a cancer when found early versus when found late. What we're finding in our members is over 3.3% were found to have these cancers that were otherwise wouldn't have been found or detected. Yeah, you know, it's interesting. People, you don't feel the cancer until stage three or stage four. And if you don't know what's going on inside your body, it's like driving your car with your eyes closed. And you can know. And so when members come through, how do they detect cancers? So we're doing full body MRI, and we also do early cancer detection screening.

This is very, very important. These are not typical tools used in the conventional care setting when it comes to prevention. This is a hard thing because currently these are not studies that insurance would yet be covering, but the goal is to collect these numbers, do the research, and work hard to democratize wellness. Yeah. So at the end of the day, you can know what's going on inside your body. It's your obligation to know. So check out Fountain Life. You can go to FountainLife.com slash Peter to get access to the latest technology to help you detect cancer at the very beginning at stage one when it is curable before it gets to stage three or stage four. And you're a world of hurt. All right, I'm going to move us to our last story where more billionaires correlates with less poverty. So there's a lot of resentment against the tech billionaires, against the concentration of wealth. And for our final story, I want to look at a post by economist Christopher Clark, who examined cross-country data showing that more billionaires per capita is associated with less poverty. He wrote that a tenfold increase in billionaires per capita goes with a reduction in poverty of 43%.

He noted that Sweden has more billionaires per 10 million people than the U.S. It's 34 versus 30. He stressed that the graph shows only a correlation, and it asks, you know, what is the cause? Let me show the graph here, and let's talk about it. And I think it's an important realization for people. I mean, there's been such, you know, the billionaire tax and resentment out there. Here's the chart showing the data. We see percent share of the population under $30 a day. That's the definition of poverty. And on the bottom, we're seeing basically the number of billionaires per country. It's a pretty stark correlation. Obviously, these people are building companies, creating wealth, and employing people, but I think also paying taxes that flows through social services. Imad, your thoughts here? yeah i mean i think this is unsurprising to a degree like where you have billionaires you

typically have the infrastructure to build billionaires right and as you said there's the tax there's the kind of other elements of it flowing down etc um and if we move into this new phase whereby you see literally companies being launched that can get to a billion dollars in no time whatsoever, right? Like, you're seeing, Anthropoc, I think, is going to have almost 5,000 millionaires made, you know, and dozens of billionaires. Like, it's a crazy time whereby even the companies at the top left can create environments for millionaires, centillionaires, and billionaires to come about. And I think that's just very exciting to think about it, as well as hopefully lift up the base level of everyone. Dave? Well, my read on this chart is, you know, top right is suck land. That's where you don't want to be in the top right. You've got lots of poor people and lots of billionaires, and it's a world of suck. And everybody thinks that socialism is somehow correlated. Look at the world of suck. It's just sucky, sucky governments.

And, you know, look who's up there. I'm not going to get myself in trouble by reading the slide. You guys read the slide. But it has nothing to do. Socialism sucks. It sucks for everybody. It's obvious on this chart. And then look at the thriving countries in the bottom right, Switzerland and Germany. and the United States is down there. So it's not a question of left versus right. It's a question of good ideas versus bad ideas. And so this chart tells you everything you need to know. What you want is your population to be thriving, and at the bottom right corner or anywhere on the bottom of this chart, you're doing really, really well. And so, yeah, the line is kind of a loosely fit line. I think the real story here is good ideas percolating to the top and getting out of partisan messes. two or three things one naturally if you have a lot of billionaires it may indicate a generally robust economy and therefore you'll get better a less number of a smaller number of people in deep poverty so i don't think correlation is causation here i do think there's a couple of

interesting things here one is the sweden example is really interesting because it challenges the idea that entrepreneurship and a strong safety net are incompatible right we like i'm canadian It's awesome to have a decent social safety net with health care provided for you. So you're not worried about having your family go bankrupt if somebody gets sick. I think the big question in all of this is can we lift the bottom? The hell with how much the richest people have in the world. And Peter, you and I talk about this all the time. Can you lift the bottom? And I think the demonetization and democratization of all these technologies is where you're going to see some incredible magic coming out of this. I mean, the big metric I would like to see talking about benchmarks is how quickly can we lift the bottom? Yeah, yeah. I make that point again. You know, we all have trillionaires living forever on Mars as long as every single human on the planet has access to all the food, water, energy, health care, and education. If I was a trillionaire, by the way, I'd rather be living in Bali just to make a point. I'm happy to visit Mars just as an exploration.

But here's my question for the group. If you have an AI economy, right, would you rather have twice as many billionaires or drop the cost of living by 50 percent? And I think I'm going to suggest that the second part is better. Inevitably, I'm going to question the premise. I think that's a false tradeoff. I think the cost of 2026-era living becomes too cheap to meter sometime in the not-too-distant future. and also, I would argue, and I have argued, we want to live in a world with the most possible trillionaires. I would say the world doesn't have enough trillionaires. Right now we have approximately one, maybe more than one if you count certain nation-state quasi-monarchs, but approximately one, and I think the world needs billions of trillionaires, and that's the world that we should be marching to. Well, I think you nailed it, Alex, in the sense that there are dictator trillionaires and then there are productive trillionaires. Right. And the key to tax policy is to tax the former and let the second group reinvest.

And so it's not about taxing more or less. It's not about taxing trillionaires or not taxing. It's about targeting taxes at unproductive parts of the economy and reinvesting in productive parts of the economy. I feel like I'm a trillionaire in inner space. So there we go. All right. As my grandfather used to say, Salim, that and a nickel will buy you a cup of coffee. Yeah, and today that and five bucks will get you a Starbucks coffee. That's right. All right, I'm going to play us out with a tribute to Jean-Luc Picard by C.J. Truhart. It's called Make It So. They said it was just science fiction. But every fiction starts with someone's vision. The future isn't something that happens to us It's something we build One moonshot at a time I'm a man with a plan And the moonshot is my mission Divided by hypocrisy Inspired by a vision I know I'm not the only one

Tied up division So the future I'm creating To shape the way we're living Cause abundance is my nature This is the gift we're giving Inspired by my moonshot mission Inspired by a Star Trek vision Make it real what I start on television, television, television, television. There's no replacement for what's real, though it inspires what we see. If we lose our dreams and drift away, life just seems to waste away. Hanging on a hole But the power's in what we do I see The reality before me Where we live among the stars Inspired by each other To explore beyond a market It's a star-spread future The one we must create Through your moonshot

It's time to innovate, it's time to innovate. Your moonshot cannot wait. Only certain things can't be done by you. Only you can see the lens of you and through. Empowered by your heart is how you make it true. You need us just as much as we need you. Transform this moonshot mission into united human vision. Where poverty is extinct, disease is historical fiction. Where energy is endless, knowledge, no restrictions We explore the cosmos now through competition Through collaboration, that's our competition Think about the phone in your pocket More power than Apollo's rocket Now with someone who's tired, someone who wouldn't stop Every breakthrough that we're living Started with someone giving everything they had to make The world in which we're living The mistakes we made before The limits we swore were carved in stone Were carved in stone Until someone alone said I'll make it better And change the world forever

From the printing press to the internet From the light bulb to the jet Every moonshot that succeeded All right, gentlemen. And that's a wrap. Thank you for a great show. Awesome conversation. Fun times. Thank you. Yeah. Love you all, guys. Love you too. I've got to go give a talk on top of the dugout. Yeah, go stand on the dugout. We have friends photos.