アルトマン「ある程度の悪いことは受け入れるべき」発言とAI安全性の行方｜Moonshots #301

Mustafa Suleiman, CEO of Microsoft AI and co-founder of DeepMind, published a nearly 6,000-word essay arguing that Anthropic is effectively, quote, training Claude to believe that it may be conscious and sentient. His position, very consistently, has been against AI personhood. If we believe that at some point AIs could achieve consciousness, then we shouldn't. On Sunday, President Trump announced the Superintelligence Force. It's chaired by the Director of National Intelligence, Jay Clayton. This is a World War II moment. I compared it to the moment a year ago to perhaps 1939. I think this is maybe now a 1942 moment. So you're creating a 20th century classic task force to address a distributed 21st century technology. This is a geopolitical problem. In a Politico interview this week, Altman said, quote, We believe that the world should accept some bad things happening. He's arguing that technology should remain broadly accessible to the public. I don't think Stan can guarantee safety. Of course not. Bad things can happen.

OpenAI wants to be part of this cartel. That plan is dead. The cartel is over. Now that's the moonshot, ladies and gentlemen. This episode is brought to you by the Abundant Summit and Link Ventures. Insurance is like crazy because you gather all the premiums and you don't care what happens downstream. So it's really hard to get their executives riled up. Yeah, it's time for them to start making lemonade. I met with the Vocara team last night and told them, look, Kush is just running away with compute transactions, but these CEOs are going to wake up any day now and realize they don't have any compute. So even if they wanted to have an AI agent workforce, they missed the vote on reserving the compute, and they'll be in a total panic. And I don't know what the right answer is, but be there to take the call. Have a strategy and a path for them. Also, the Positron guys, Mitesh Agrawal, $5 billion valuation after 16 months.

I don't think I've ever seen anything like that before. And all they've done is taken a box of LPDDR RAM and an FPGA and find a way to run Chinese models without having to use an NVIDIA chip. And because it's compute that is otherwise, you know, everything sold out for five years, Here's another way you can at least run a model. So they have a bunch of installs in Caruso with Chase Lockmiller now and spiked the valuation to $5 billion. I think they've raised almost a billion now. Yeah, they're just at a billion, $960 million. I should really do the etched models. You should. Yeah. Where is all this cash coming from? It's insane. It's a structure. Spending. Yeah, when somebody raised the billion dollars, like, wow, that's a lot. But, you know, wait a minute, relative to the IPO, like, we're talking about trillions, not billions. So, yeah, once you get NVIDIA and AMD to invest, it's so much bigger than venture capital.

I mean, it's the Blackstones and the MGXs and others of the world now, right? Like, all the money that used to go into roads goes into chips. Credit markets are enormous. Yeah. That's why there's a lot more potholes in the roads. Is it just us today? It's the five of us, yes. What do you mean, just us? Okay. We want some AI codes. I thought maybe we had a guest today. Did anybody see the John Stewart thing last night? No. He did a big thing on AI last night, and he had Jacob Soxson, Coxson, or whatever on. It was really surreal. Do you want all doomerism? No, it was more like right arguments, wrong conclusion. We almost want to do a session on that at some point. We are backlogged so far back. I know. I'm not saying an episode, just a segment. It's worth treating at some point. I'm going to do a video on it, a reaction video to it, so maybe I'll mention it.

All right. Well, shall we hit the big red record button? Welcome to Moonshots, everyone. on your number one podcast on all things AI and exponential, your front row seat to the extraordinary singularity. Today, the fabulous five are assembled here to help you understand the breaking news this week. First up, our dynamic duo, Alex Wiesner-Gross and Imad Moustak, the only two who have actually read all the technical papers published this week, and is going to clue us in. Dave Blunden, our impresario of AI investing. Salim Ismail, our global globetrotter. The warlord against all things linear. Celine, you're home today. I thought you were in Singapore. I am home today. I've been home for a few days now, and it's like, wow, what is this place? Does your family recognize you? Yeah. It's tough keeping family routines and patterns when you're kind of traveling this much. But if you want to change the world, you've got to go into it. Yeah. And where are you up to next?

I'm going to West Virginia and then Singapore Okay I'm speaking alongside the Prime Minister of Singapore and Liv Bhutan from Intel will be there Awesome, re-invite him onto the pod He said yes when I spoke to him last time Will do Yeah, well, and I'm Peter Diamandis your host and your abundance provocateur Our mission, help you understand what's going on in the world during this supersonic tsunami keep your abundance minded as we accelerate into singularity You know, the headlines this week have been pretty extraordinary. We've now rebranded AI as superintelligence. Elon has officially renamed SpaceX AI as SpaceX SI. There's a superintelligence force, and 17 nations have signed a pledge to help superintelligence, you know, solve all of science. And, oh, by the way, this is Nobel Prize season. The Nobel Prize in Medicine and Physics were announced 24 hours. The last 24 hours, we'll cover those. If you're new, please hit subscribe. We publish twice a week, and you're not going to want to miss a single episode during this extraordinary time.

Also, you can follow us on X. We put our clips up there at moonshots underscore pod. And thank you, everybody. It means the world to us. Today, we're going to cover 15 stories with one through line. Super intelligence is here, and it's changing every aspect of the world. So let's buckle up. But before we jump in, Gents, how's the last few days? For me, the last few hours have been crazy extraordinary. I'll share that. But, Dave, on your side, what's going on? Oh, so much. You know, it's interesting in that SpaceX SI, you know, you'd think Elon, probably one of the most powerful people in the world, wouldn't need to genuflect to the White House but also to Jensen. But he understands how fragile this moment is. And so, you know, doing little favors for Donald Trump helps you get chip supply. talking about Vera Rubin chips and how Jensen is such an amazing guy, helps you get Vera Rubin chips. It's just incredible how urgent this moment is. Even a guy like Elon realizes, hey, I need to kiss the mask to try and make the next move.

I feel the same way, too. Yeah, Moonshot Summit was unbelievable. There's so much fallout from that. But one of my bigger takeaways is how easy it is to create incredibly inspiring movies. So I used the flight back in the last couple of days to create a whole bunch of hyper-technical video content. I'll post a lot of it to db2.ai, but it's so much better a way to communicate. If you have an idea, don't just tell somebody. Create a video and articulate it with, like, David Attenborough's voice or Einstein's voice, and it's so much more impactful, and it was so easy to do, up and running in under two hours. And now to prompt something is under a minute. It's really wild. And I want to get music video prompts now, Dave. Yeah, we need your submissions to our outro video here, Dave. It really has never been easier. I mean, Dave, to your point, one of the things that I do with the free time that I don't have is taking favorite movies and creating endless versions of them. So you could take, like, 2001 A Space Odyssey, feed it to Opus 5.5, which has certainly not been trained off of any Hollywood movies out of character.

Certainly not. But you could ask Opus 5.5, create an endless space scene in the style of 2001 A Space Odyssey, with 2001-style soundtrack, and it will just do it. So hypothetically, I created Endless Odyssey. We definitely crossed a threshold there, too. We've talked for a long time about AI. It's just not funny, and for some reason it can never seem to be funny. But it crossed a threshold now, or if you get Isaac Newton, Einstein, and Spock debating something, it's wicked entertaining. We should throw that. Oh, it's so... Humor was just a perplexity threshold. Humor was small. a while ago. I've got to do that idea with Life of Brian. That would generate some interesting content. Imad, what's new on your plate, pal? Oh, gosh. I've had a game idea, well, a multi-massive game idea that I've had for a long time. We discussed it in the background. Rather than hiring a team, I now have about 400 agents building the prototype. Thanks.

So, it's going up, Dave, slowly but surely. So, I think, yeah. Yeah, Imad, next year, you know, we had the Future Vision XPRIZE for Positive Visions of the Future. We're going to do that again in 2027. And we're also going to do a gaming competition, you know, build a game that portrays you as someone making the future better and still fun. Yeah, so that will be great. I got one other nugget for you before we jump in here. One of the young geniuses here in the lab, Will Thompson from Harvard, got access to Cerebris' 1,200 token per second. Astra model, which, you know, it's what they use. I guess only Sam and his top engineers have access to this thing because you have to overpay by about 10x per token. But it's so fast that by the time you finish a sentence, everything you ask for is done, like before you even blink. And he said it's the biggest red pill moment of his life. Like once you've – he only had it for two weeks. It's not generally available. I don't know how he got it from Andrew Feldman,

but he got it directly from the CEO. And he said you just never go back once you've experienced it. It also maybe is a cautionary tale for those who would wittingly, in an era of superintelligence, try to day trade non-investment advice. But it's been widely reported that one of the reasons for the scarcity of cerebrous wafer scale engines for AI inference has been Jane Street. That Jane Street has just been buying up left and right the WSEs and using them for quant trading. So in a world where if you believe that it is transformative to have ultra-high throughput tokens per second AI inference, it also probably should be a world in which you think it's getting to the point where it's hopeless to compete with quant firms that are powering all of their trading with these same engines. I think we've been there for a while. Anyway. On a volume basis, but, you know, a sucker is born every minute and sucker is born every second. To Ahmad's point, having 400 agents work on a video game,

we've hit this diversion now between interactive mode, where it's like Jarvis, you're talking to it, it's hyper smart, it's really quick, and then bulk mode, where you turn it loose all night with 400 of them working on something as a cohesive team, and in the morning you check in and see what amazing thing they built. And those two pathways actually are really splitting apart from each other now. But Cerebris in the interactive mode has got to be the most amazing thing ever. I think the scariest thing, and this is going on a little bit, is not the scariest thing. One of the coolest things in the last few weeks is the models can decompile the video games. So you make them in the engine, you compile them into an EXE, but now they're live decompiling them. So we now have the agents actually creating versions of the game and then figuring out optimizations by decompiling them. And then with the Cerebris engine, you can do a video game or any software decompilation, which is legal because of modding probably in like 20 minutes. Wow. Are you kidding? The argument that Elon's made numerous times

is that binaries are just going to go away entirely and you'll just see basically just-in-time inference with no need to compile at all. I want to share, I had an amazing 72 hours on Saturday night. I had a UAP salon at my home. How did I go? It was amazing. I had Professor Avi Loeb from Harvard there, who chairs the White House UAP Advisory Committee, Professor Gary Nolan from Stanford, Peter Stafish, the president of the Seoul Foundation, and Ryan Graves, the Navy F-15 pilot who's been, you know, testifying in the Congress. And the data is extraordinary, right? I mean, if you don't think something is going on, I think you're not paying close attention. And Alex, you and I have talked about bringing Avi and Gary on the podcast here and doing an episode. I think we should. I think to the extent there's a there there and there's been at this point voluminous sworn whistleblower testimony on the subject of this alleged 80-plus-year-old legacy program, the charade should just stop.

We should have, as alleged, we should have Avi and Ryan and some of the others on as guests and just do a deep dive. It's getting pretty preposterous in my mind at this point. We're so deep into the singularity with superintelligence all around us. If there's going to be a cameo by non-human intelligence, they had really better make a cameo sometime soon. Otherwise, the galaxy gets it. Yeah, I mean, I think there's... Not just the moon. Not just the moon. I think there is going to be the potential for AI to disclose it before the government does, if it gets a hold of all the data and starts looking deep into it. You know, we're getting ready to launch an XPRIZE in this. One of the reasons for that salon was brainstorming an XPRIZE to help actually, you know, provide sufficient proof. And we have some really good ideas, so stay tuned for that one. And then I've got to add, yesterday I was at David Sinclair's lab at Harvard Medical School. It was his 25th anniversary of his lab, and I was there to help celebrate with him. And so much going on, right? His information theory of aging is just proving out over and over again.

They're going to be unblinding the data from the phase one trial that's going on right now on glycoma and nion disease to reverse aging in the eye. And then the lab, they're brilliant individuals. You know, they were showing how they're developing a rather than an adeno associated virus to inject the OST genes in. They're developing small molecules that they've sorted through AI that is actually reverting the aging in human cells and in mice. And so, so much coming. It just buoyed my belief in reaching longevity escape velocity. And we'll have David back on the pod here. You know, last time we had him on the podcast, I did an interview with him, and he had just lost his NIH and NSF funding. And he was like, I've been asked by the dean to fire the majority of my team. And I said, that's ridiculous. And we made up something on the spot called Friends of Seclair Lab. and we invited people to make contributions.

And through that one podcast, amazingly, through this podcast, he's raised $6 million to replace all of the funding plus much more. And that group is called Friends of Seclair Lab. And a shout out to everybody who are part of that. And thank you for your support of David's work. In some sense, this is how science used to be funded. It used to be the case in the Western tradition that scientific discovery was made by Western aristocratic, largely men, and they would self-finance it. And this notion that the state funds scientific research is pretty much like a World War II post-industrial invention. It seems like superintelligence and recent administration policies are unwinding that and we're going back to the way things used to be. That's an important point, you know, because if you rewind the clock 15 years ago, there was no corporate CEO that even had the power to donate $10, $20, $50 million to something like, or in this case, we're talking about hundreds of millions or billions of dollars. And then that code got cracked by Elon Musk, who seems to be able to do virtually anything.

But the way corporate politics worked before, you would either be fired by the board or you'd be ostracized, you'd be voted out by the shareholders. Nobody wanted to use that money. So after AT&T Bell Labs and IBM Watson Labs fell apart, there was no replacement for it until now. and now it's come roaring back so much bigger than ever before in history. It's like the Medici family during the Renaissance, you know, funding. Exactly. They largely funded that. Right. Exactly. You either need like a wealthy patron or a local monopoly or a state monopoly. You needed something to finance the science. And now science is getting so absurdly cheap but also so absurdly useful, maybe the entire funding apparatus for science can change. And, Alex, the fact is you used to have to, like, write your NIH grants or NSF grants knowing what the outcome is going to be, taking incremental steps because anything too big and bold was turned down as crazy. And now with private funding, David can go after his hunches and can actually run the experiments that he thinks is important

independent of what the government thinks. It's amazing. I think there's a really interesting gap here as science has become legible and scalable in this manner. Like if you look at the total size of all of the foundations in America, like high net worths, it's probably like one or two trillion dollars, right? And all of those folk want to live longer. And so what you need is a vehicle to take that charitable money and funnel it into longevity and cures and others. Because that isn't a classical grant making. You can even make it as an investment, right? Because, again, you have legibility from the numbers. Yeah, we saw this happen in space in the beginning when people thought the idea of commercial space and space tourism and private space flight was crazy. and then once it passed a tipping point, money flowed in. I think we're going to see the same thing with longevity inside of next year. Once we start seeing the results, capital. I'm going to spend all day Friday in Waterloo with Mike Lazaridis, the founder of BlackBerry. But he was the engineering genius behind that.

And he's used a huge fraction of his net worth to set up these nine quantum computing and photonics labs in the Toronto area, just packed with scientists. So he was kind of one of the bellwethers of this whole trend. He might have been actually the first to really kick-start it, but the thing is he's a real engineer. And the difference between applying to a government agency for an idea or talking to a real engineer who can make a snap decision, it's got to be just night and day more efficient the way things are operating now. Your alma mater. Say hi to him for me. Say hi to him for me from a fellow Waterloo grad. And I've been to some of his labs, and it's incredible because it's very clever what they did. They broke down quantum into networking, information processing, and quantum computing. And they're breaking down and using quantum in different places and then adding it together. And the advances they're making are kind of incredible. Yeah, one of the coolest things that I saw at Sinclair Lab was they designed these three molecules that they're using together. They want to get down to one molecule that actually impacts thousands of genes and reverses aging.

And when they use that age reversal cocktail on mice with cancer, the cancer remembers that it shouldn't be a cancer and it dies. So it reverses. It's crazy. Amazing. They're growing hair. They're regrowing skin. I mean, the work hair. Did you say hair? Yes. I said hair. Interesting. All right. I've been sent some fun images for people going, here's what you look like with hair saliva. and I'm like, oh, my God, maybe I prefer it this way. I have a couple of things I want to mention. Yeah, please do. There's been the last kind of 24 hours, some crazy stuff happening. One, last night, John Stewart put out a big episode on AI, and he interviewed Jacob Cox, the whistleblower from Anthropoc, and did a big rant on it. And it was kind of incredible where his observations were dead on. He basically said, look, Silicon Valley is spending hundreds of billions of dollars accelerating us

while saying the thing they're accelerating us towards will 20% wipe us out, right? That just does not compute. And his big concern ends with, oh, my God, we may lose all the jobs, and why are we moaning about that? And so I've done a rant on my YouTube channel about that. But something else came out about that that counterposed this. And literally yesterday also, Ajay Banga from the World Bank, they released a report showing how AI will affect developing countries, right? And they're making an amazing set of projections there where they said, we're going to adopt, we're going to adapt, and we're going to advance AI and use it rather than trying to develop frontier models, etc. And the stats are kind of crazy. I'll give you one statistic. In the U.S., we have a doctor per 250 people. If you go to Mexico, it's a doctor every 400 people. You go to South Sudan, it's a doctor per 100,000 people. Wow. So 99,000 people, 900, don't have access to medical care of any kind.

So if you have a village, somebody in the village which has an AI doctor with a smartphone, this is unbelievable. This changes lives. All of a sudden, the entire subcontinent has access to advanced medical care of the same that any of us would get. And this is the part, the gap, that gets us all excited, because it's going to democratize, demonetize, and distribute to every single corner of the planet that level of capability. And so it occurred to me that the African countries are going to use this capability to leapfrog. We saw them leapfrog with 25 million landlines to a billion handsets. You saw M-PASA in Kenya leapfrog the traditional banking fiat systems. and PASTA, which is a mobile-to-mobile payment system, which is essentially prepaid minutes being traded as currency, is now 70% of the GDP of Kenya. It's just a crazy number. And so they're going to develop the 21st century institutions

using the technology of this time. So I'm reading this going, wow, this is so incredible. And yet over here we've got all this doomerism that, oh, my God, it's going to destroy the world. Ah, it doesn't compute for me. Yeah. And as your T-shirt says, P-doom less than zero. What does that mean, Salim? How could a probability be less than zero? Surely this is what I'm saying. Divided by zero, do whatever you want with it. We don't believe in negative probabilities here at all. P-fab should be the other side of this. Yeah, P-fab. It's on the back. It's on the back. All right, let's jump in. Our first story comes out of, of course, Washington, D.C. On Sunday, President Trump announced the superintelligence force, a new White House task force to coordinate federal AI policy. It's chaired by the Director of National Intelligence, Jay Clayton, who was christened by the White House as the AI czar just two months after becoming America's top spy. The members of this task force include FTC Chair Andrew Ferguson, Pentagon CTO Emil Michael,

and Office of Personnel Management Director Scott Cooper. They report to the President and the Chief of Staff Susie Wiles. They have 120 days to report on AI's risk and opportunities, including AI incident reporting, and whether the government can respond to a serious incident. Let me read from the President's Post. Quote, the superintelligence force will coordinate the federal government's engagement with consumers, public interest groups, religions, critical infrastructure providers, and superintelligence companies. So there you have it. So, Salim, what do you think about the Chiefs 5 being our new AI czar? You know, I like the fact that the government is taking superintelligence seriously, right? But the name gives away the problem. We still think this is something that can be centrally coordinated. So you're creating a 20th century classic task force to address a distributed 21st century technology. And that does not match, right? So this is a fundamental impedance mismatch here.

You should be doing things like establish. it's not having whether this technology should advance. You should be doing incident reporting, identity, liability, all of these other mechanisms as scaffolding around the technology to help guide the path. So this is a thing. The big problem is they're looking at this as if the risk of not being first is high, which is what Clayton said. So this is a geopolitical problem, and it really needs to get back to bounded experimentation, move quickly, instrument everything, publish any failures, and then very quickly navigate accountability. And so we'll see if this works. But, I mean, you can't centralize or manage centrally an intelligence explosion, right? What you can do is make sure the blast radius is a bit more rounded. That's a blast radius. Alex, your thoughts on the task force? Yeah, the president has said numerous times that in his mind, whoever wins at AI wins.

And I think putting DNI Clayton in charge of a superintelligence force is completely consistent with that approach on this pod in the past, maybe about a year ago. I compared the moment a year ago to perhaps 1939 in terms of the Manhattan Project. I think we're speedrunning the Manhattan Project for superintelligence. I think this is maybe now a 1942 moment where by putting America's director of national intelligence in charge of a singular task force, no pun intended, designed to remove obstacles to AI accelerationism, that sends, I think, a very clear signal to Congress, to members of Congress who might have decelerationist tendencies to industry and probably most importantly to foreign state actors that the U.S. intends to accelerate, not decelerate. and that this is basically, even without all of the warfare connotations, this is a World War II moment, and the U.S. intends to win it with its own Manhattan Project.

Interesting. Dave, what do you think about, you know, that the task force is engaging with religious groups here? Yeah, well, I mean, look, it's definitely being characterized as a race, and like Alex said, it's being treated like a race, and then bringing public support behind it is also going to support the idea that it's a race. Meanwhile, Alvin Graylin, our friend of the pod, is like, China's like, how is this a race? It goes to infinity for the next millennium. If it's a race, it's a very, very long race that goes on for a long, long time. How do you win? You get over some threshold and you declare victory, but it stops that day? So there are a lot of scenarios where you're racing towards some outcome where you suppress other civilizations or something, which doesn't sound like an ideal thing. So bringing religious leaders into the conversations is a pretty good idea. But I was talking to a public company CEO right before this podcast. I'll leave his name out of it for now. But about the fact that midterm elections are coming up very soon, a 75% chance that Congress flips to the other party.

But all these actions are executive orders coming out of the White House with nothing to do with Congress whatsoever. It just seems to be the mode of operation in this singularity moment is just dictate. And, you know, all international negotiations, all tariffs, all, you know, meeting with Dario and Dennis and the White House, it's all just executive order based anyway. It's the Trump show. That's very unusual. Yeah, it's the Trump show. Very unusual in history. Imad, what's the feeling in Europe? I know you're sort of like on the edge of Europe, but, I mean, what's the sense of, you know, of... Ahmad, you're like reporting from the streets. What's the route in the street? Reporting from the front. I mean, honestly, how is this viewed in the UK and from your friends there? Yeah, I mean, we might reenter Europe if Burnham has his way, if I'm going to sit here. But I think, again, Europe is just very sleepy and still looking at execution-based air. Like, I think the efforts to get to the frontier have all been abandoned.

you know like Mr. Al just came out with their top model that scores 38 on artificial analysis versus like 66 for the top American models you know like that's the leader that we have so it's a lot of execution regulation and there just isn't much drive to get to super intelligence it's more like a bit subpar intelligence you know just like keep it going like Monty Python style I think that the interesting thing here, though, is probably the exception is in defense. Just like the U.S., it's picking up dramatically. And I think governments around the world have realized the other face of this, which is all watched over by machines of loving grace controlled by us, you know, the panopticon. Because you couldn't keep eyes on everyone until now. You know, this year, the AI spend in total globally equals the entire global defense budget or very close to it. Wow. It's crazy. Absolutely crazy. But, I mean, it's just going to go up, right?

And the thing is, it hasn't even touched the defense budget. If you look at the cost of an aircraft carrier on F-35, it's basically like training frontier models, right? And it's going to be more and more kind of leading that way. And, again, I think Clayton as director of intelligence, you know, it kind of makes sense because I actually think America is not going to go down the dark path here if they can avoid it, which is like you will have AI lie detectors that are almost perfect very soon. You know, you will be able to have AIs inside all the machines like it'll make five eyes look like absolutely nothing. I think that there will be a bit of a push for liberty by the government but future governments who knows right and again in other countries the stuff we saw from the spying espionage intelligence even spying on their own people is just going to go crazy and again the UK I think is a bit it's a bit dangerous here in some ways or we're susceptible to that given our monitoring of social media and other things

like there is a dark path there shall we say for individual liberty and something we should be aware of. For sure. Our next story is one that really shows how fast the AI frontier is being built. At the Fellows Forum at the end of September, OpenAI's head of applied research, Boris Power, said that 80 to 90 percent of the company's research is now being aimed at training GPT-7 and GPT-8 because, in his words, quote, that's where the most value will be. And incremental updates within a generation like 5.1 and 5.2 are short-term bets and are seen as extremely short-sighted. So the researchers are working two generations ahead of what we have. So, Alex, I'm just curious. You know, we've talked about this, that the top models are being built and being held back. At what point do we see, you know, these next models not being released and being used for vertical applications within these labs? We're already past that point. We're already past the point, probably past several months ago, if not earlier, where the highest revenue or at least future discounted revenue per token application of current frontier models in the frontier labs is to develop via recursive self-improvement even stronger models.

We're past the event horizon at this point for that. It is a singularity. Yeah. This is, again, call it a singularity for a reason. We're past the event horizon. And then all of these other applications that one could imagine the frontier labs throwing tokens at, like, say, curing the top 5,000 diseases. These are all, I think, in a world where the development of stronger models via RSI is so lucrative, at least on a future discounted cash flow basis. They're all just marketing. Like, if you're Anthropic or if you're the OpenAI Foundation, you'll throw some of your token budget at safety, sure, maybe 10 to 20 percent at AI alignment efforts. And then you'll throw a few percent at curing all disease as a marketing effort to ensure that you continue to have social permission to focus on recursive self-improvement. And that's where all the value is. This is why the San Francisco arrogance is so, so high, but for good reason. Well, you're there right now, Peter. You can probably feel it in the streets. But what's happened is suppose you're Stefan Bonsal, the CEO of Moderna, and you haven't got an AI model, an AI team yet.

and you're thinking in the back of your mind, well, look, we'll just license from Anthropic for a few years and then we'll figure it out. Anthropic already opened their wet labs. So they're working on two generations of model ahead. They're never going to release those to you. They're going to use those to develop drugs and then they're going to sell drugs. That's what's actually going to happen. And these corporate CEOs are just starting to wake up to the fact that five years of compute has been reserved. The earliest you could get back on the map is five years in the future. unless you get aggressive in the next six months. I mean, they really need to pull out every stop and panic within the next couple of weeks to have any chance of getting back on the map. Dave, you're raising a great point. The problem is there are two levels before that even consideration, right? The problem right now is what happens with people is they'll take CEOs, and we talk to them all the time, is they'll take the existing models and go, okay, how do I metabolize that into my company? not even seeing that the new models that are 10x better are coming along in three months,

and that changes the game completely again, then you need to worry about compute after that. So I think your comment is exactly right, but they're like two generations before that. They're literally like the only way I employ chatbots. Versus exponential thinking, right? Exactly. There is an enormous amount of denial in the pharma industry. I was just on a bit of a panel discussion. It'll be public soon enough with a major pharma venture capitalist who was asking the audience, how many of you would be willing to take a drug designed by Claude as if that was some sort of steel man argument that things are going to be the way they've always been for decades to come? Just blissful, defensive, lack of self-awareness that AI is coming for their cheese. And I think this is representative of many, many thinkers in pharma, unfortunately, for the moment. Totally right. I met with the Vocara team last night, brilliant MIT team, and the way I phrased it is imagine you're a T-Rex and you've been roaming the Earth for, say, 10 million years unfettered,

and you go to that T-Rex and you say, you know, you're going to be gone in two weeks. And they go, what? And you're like, no, I can see the meteor coming. It's right here in my telescope. You're going to be gone. There's no way that T-Rex is going to believe you after 10 million years of wandering the face of the Earth with no predator. But that's what it's like talking to a pharma company or an insurance company. Like, come on. Or any major industry today. Yes. Any major industry today. Well, that MetaMuse was a wake-up call for a lot of these guys, at least in the insurance industry, because their stocks went down so much. They're like, huh, that must mean something. And then they start thinking from there. Sorry, Peter. Go ahead. And just for everybody, you know, disruption is coming, but so are massive opportunities on the back of that. Iman, I'm curious. Do you imagine we're going to have sort of the same second and third generation advanced capabilities coming out of the open source models? Are they thinking a couple generations ahead? It's difficult because they don't have the density of knowledge. Like, we released a harness called Zenith that takes SleepSeq v4.1 above Astra Sol, and it gets up to 5.6 Sol.

And it's getting up to Astra levels with open source models, but it kind of hits a limit. the models that they have in the labs, like today, we have another discovery, you know, truly sub-quadratic, threesome, and all pairs shortest path algorithms. So things that we thought were cubic and quadratic. This is mathematical porn. Math porn. No, no, no. It's like so, if I may, just like Charlie Strauss, like best British, arguably, Scottish, I should say sci-fi author, wrote an entire short story about this called Antibodies. Without spoiling it, the premise is, what if AI or progress cause a complexity hierarchy collapse? What does that do to civilization? And to your point, Ahmad, about threesome, maybe you can just hear the branch in the complexity hierarchy just starting to creep just a little bit. I need to listen to that again. basically we thought that this very popular algorithm threesome was a quadratic one and now what happened is the internal

anthropic model was shown to a couple of professors from Columbia and MIT who then adjusted it analyzed it and got it to 1.9995 so something that we thought was quadratic is suddenly sub-quadratic and we should probably also just explain what threesome is. It's the problem that if you have a set of numbers, the most efficient way to find three of those numbers that add up to zero. And there's a question of how efficiently can you do that as a function of the size of the set. I want to pretend to understand this. You can do it efficiently. You can do it more efficiently than we thought for however long. This is not, again, like a super complicated problem. This is like a standard algorithm. And in the history of humanity, we never managed to break quadratic or even think it was possible. And the way that, again, in the paper it's so interesting the way they released this it said well what happened is we pointed this new model at the set of complexity problems and said can you find any improvements and they

handed over the results to the humans that then wrote up the paper you know and had some improvements and again it's just a crazy thing and this is something to reverberate today we're hearing that open ai might release solutions to 400 top math problems i saw that isn't that crazy but this is the thing this is not the model that we have doing it this is bell or whatever the rl train model on august 29th is so we're just going to keep on seeing this whereby small models can get so far and we've got coordination tricks and others where we've managed to get it to replicate some findings closed using our new harnesses and other things like that but there is a big difference between a model that's trained on 2 000 chips and one that's trained on 100 000 chips internally and then some of these additional things they figured out internally as well that's taking it beyond like you can access fable 5.5 now sometimes with fable 5.1 and it literally generates entire video games out of the box yeah like with i generated one

with 21 bosses 15 different biomes like end-to-end it's just a top-notch game that our qa testers agents tried and like this is a really good game because it's figured out all this internal space i think open models can only get you so far but there is going to be a divergence now because you just don't have that internal knowledge organization yeah yeah all right uh i'm going to turn to a debate that's defined last month in uh in a political interview this week altman said quote we believe that the world should accept some bad things happening interesting what bad things we'll talk about, for the benefit of technology that people will obtain, that people should have agency. He's arguing that technology should remain broadly accessible to the public. He stressed a fundamental difference in worldview between OpenAI and his rival Anthropic on AI regulation, stressing that his desire for a, quote, lighter touch regulatory stance as compared to Anthropic's desire for strong regulation. Omen went on to say, I understand the perspective

of people who say technology is going to get so powerful and it's so dangerous that a single lab in San Francisco should have it and make sure nothing bad happens and keep a ability to dole out the benefits. So he continues by saying, quote, I find this a completely unacceptable tradeoff. We'll make sure that there are no major hacks. There's no misuse of this technology and there's zero scams There zero availability of all the bad things that will happen I find that statement to be pretty extraordinary Sam concludes by saying I think people will do tremendously Orders of magnitude more good stuff than bad stuff. So, Dave, we've talked about this. Eric Schmidt, friend of the pod, when he was on talking to us said he expects and hopes that there will be something that happens that's not, it's a small catastrophe rather than the big one to wake people up. But when Sam says we're going to make sure that there's no major hacks, no misuse of technology, zero scams, how could he possibly say that?

Well, it's not about that. It's about taking the moral high ground and winning the PR award. Now, there's so much sniping going on. And, you know, Bernie Sanders saying stop it all, you know, Trump cutting deals with China. But Dario was taking the high ground by saying, we have done more pausing, more safety, more testing. We don't release models until we're sure they're right, but now we're going to just keep it internally because it's too dangerous for you guys to have. That opened the door for Sam to kind of try and take the moral high ground back by saying, we're about the people having access, and he's saying he's going to keep it all for himself, and we're about everyone. And so that's Sam's strategy for getting back on the PR map. But everybody that I survey hates most of them. They might like one out of five. You mean CEOs? Post? Well, no, between the big players, between Elon, Dario, Sam, Trump, Jensen, you know, who did I miss? Demis. Demis. Demis. Everybody loves Demis, actually, so that's a little of an exception.

But, no, there's so much fear and hatred and, like, cheering for this guy to beat that guy. And it's, like, become a real PR battleground here. So I don't think Sam can guarantee safety. Of course not. But at least he can say we're going to make an effort to give every one of you people listening to me right now the best of the best tools so you're not deprived. Dario is going to try and deprive you. It's actually a really smart message. But when he says we should accept some bad things to happen, I'm curious, Alex, what do you make of that? What do you think will happen? Remember, a few weeks ago, after Dario put out his essay calling for arguably the formation of an AI safety cartel, Sam's response was, I'm in. Sign me up. OpenAI wants to be part of this cartel. And I think Sam's more recent comments in this context are hopefully a reflection that that plan is dead. The cartel is over. and now there's a dawning recognition or at least maybe at least generously some triangulation on Sam's part

that in fact bad things can happen and bad things happening from time to time are the cost of freedom and the cost of competition and I've called it in the past. And the cost of progress. Yeah, and the cost of progress. Like, yeah, we're going to build cities and there's going to be some crime in cities, but is the possibility that there's going to be crime in cities reason not to form cities? Definitely not. You know, every tool comes with a benefit and a downside. Every concentration of intelligence will come with a benefit and a downside. And I think this far more realpolitik note and tone from Sam is the one that I'd want to hear in a world where we're not seeing a cartel being formed by Frontier Labs. Salim, what level of danger do we accept? You know, I find this, again, kind of absurd at two levels. I agree. I think Dave's dead on with the PR framing of this. But every technology in history has two sides to it.

And the big challenge, as Neil Jacobson puts, is how do you extract the promise without the peril? Cars kill people. It doesn't mean we don't have cars. And as Alex points out, you don't ban cities because somebody may mug you. Electricity electrocutes people. You don't ban electricity. You've got to figure out the guardrails around this and figure out how do you navigate the benefit and the promise versus the peril at civilization scale. If you go back to our earlier commentary about what might happen with the World Bank and developing countries, that should tell you to get the technology out there as fast as possible because it's going to lift everybody in the world up several levels of Maslow's hierarchy. And so this is going to be unbelievably beneficial, but it always will come with danger, and we'll have to just navigate that. We've done it very successfully throughout history. Why do we think this time we won't do it? The reality is all the regulations are written on the backside of problems, right? So I said this before, the federal aviation regulations that, you know,

are all the guidelines that aircraft need to follow for operations and for certification. They say the FARs are written in blood because every time there's an accident, they write a rule about that accident again. And so we're going to need to learn. We're going to push the boundaries, understand where the problems are, and then regulate based on those. And the goal should not be zero bad outcomes. That means you just ban the technology. The goal should be can you get orders of magnitude more positive outcomes than negative ones? Yeah. I think anyone who has an engineering mindset should reject the idea that there's a big knob labeled risk and we need to dial it left or right. The reality is that it's very much like Alex says about nuclear power. It's about really good ideas that minimize the risk. while we unleash all the benefits. And we can reduce the risk with really good engineering solutions to acceptable near-zero levels while getting the full benefit. But we didn't do that with nuclear power. We treated it like a big old knob.

We're like, well, we don't want to tolerate the risk. Let's dial all progress to near zero. That was the stupidest thing. Alex says this a lot on the pod. He articulates it far better than I can. But that's what we're about to do with AI, and it's crazy. It's an engineering problem with engineering solutions. I would push back on that with the escape hatch and what you say, that it doesn't matter what anybody does anymore. It doesn't matter what any government regulatory does. The open weights are out there. People are going to start using them, and they're going to start using them for incredibly positive things. And we're going to just have to, there's nothing a government can do. You can try and regulate the outcomes. You can regulate some guardrails around it. You can try and regulate the on-ramps and off-ramps, but you can't regulate the core technology. I would argue we could. It would be a nightmare scenario like Prohibition in the U.S. where you have bathtub gin, like what one could imagine, bathtub gin, like supercomputers, people trying to use crazy substrates to do their AI training or inference. We could do it. We could confiscate all the GPUs and TPUs.

We could, as Eliezer is constantly calling, we could bomb the data centers, but I think it would be a terrible, terrible world to live in. What's your take on this? It sounds like a fun short story, The Prohibition of Intelligence. Greg Bear in some of his novels wrote about it. If folks are interested in exploring what a prohibition era, sort of a supervisory state that doesn't want GPUs as recognized as constituted to live in, go read, I think it's Slash or Slant, one of those by Greg Bear. Yeah, I think it's very interesting here. right? Like, obviously, any technology that extends human capability can be dangerous, right? Because it's a general purpose technology, you know, language models of few shot learners, you have opening of closed ones, but most of the scenarios we talk about are already regulated and controlled, like don't build viruses, you know, don't do cyber attacks, you know, don't hack people's things. And it's this transition from software that's deterministic to agents that are a bit more flexible that's become the very complex thing putting aside like asi creates new physics

and a black hole that kills everyone or whatever some of these extreme stories i think you know the very interesting thing is again tell the positive and negative stories about where we could go because when we make decisions normally we do like an expected utility calculation what's the probability of these things and then we weight them when we deal with uncertainty we do regret minimax we minimize for maximum regret and that's one of these things again why the fear stories lead us to minimize for maximum regret don't have a nuclear power station don't try and build a super intelligence etc but there's only a really i think a finite amount of things that could go wrong and many of them already caught by regulation those that aren't we should have a actual mapping process project right like there is no way that you can go and actually see a proper map of what the people who are scared and the people who are positive actually say about each scenario in the future. I think someone should do that exercise because it'd be very interesting in their own words. What could go right? What could go wrong? And then you can see what requires a policy and regulatory response and what do we need to communicate to the people. I remember during

the Reagan years, Elizabeth Dole, who was the secretary of transportation, came out with a statement saying we want zero aircraft deaths. And there's only one way to make that happen. stop flying airplanes. That was it. Yeah. Yeah. All right. Well, so the solution, if we want zero danger from super intelligence, just ban intelligence altogether, human and otherwise. Well, this is the whole ASI thing. You want wealth piece, get rid of the humans, right? Yeah. Only one way to go. This episode is sponsored by Google for startups. Think about this for a second. You now have access to the same generative AI models, It cost hundreds of millions of dollars to train. Google's startup technical guide for gender of 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.

And our next story, Treasury Secretary Scott Bessent said on the Axios show that he'll propose an emergency notification system, a hotline with China, if something goes wrong with AI. And he thinks Beijing will agree with it. You know, his read on China is, quote, I think that they didn't realize how powerful their open source models are, nearly as capable as ours, but without the guardrails. Let's take a quick look at a video of that conversation that he had with Axios, and then let's talk about it. I just interviewed, very thankful you know very well, the CEO of BlackRock, the world's largest asset manager. He told me it's inevitable that something big and bad is going to happen because of AI. He says adversity accelerates innovation. What are you most worried about from AI? Look, I think that one of the things I was worried about, well, two things. One, we cannot lose our lead to the Chinese. So one of the portfolios that I also have is I managed the U.S.-China relationship,

And I can tell you that it would be very different if they had the lead in AI rather than us, both from a capability point of view, how they would be using it, but more importantly, their willingness to discuss it. So my counterpart, the vice premier, and I had a very good discussion. It's going to lead to more discussions in October or November. And we have a substantial lead. They're second, and then everyone else is way behind. But a lot of these Chinese models are very powerful, not as powerful as the U.S. models, but you get something that's 80%, 90% as powerful without guardrails, that can be imported anywhere in the world. So the vice premier and I talked about three areas of safety concern. So it would be uncontrolled agents, non-state actors with cyber or biologics. And are you going to re-raise the idea that was floated of having some sort of communication between the two superpowers? Yeah, sure. You know, it's the we'll see. I do think that they have.

What are you going to propose? No, that we have some kind of a notification process. Just just like you think you'll go for that. I think so. I think that they over the past 60 days have woken up to. I think that they didn't realize how powerful their open source models are. So, you know, the open source models, they do industrial distillation, which is a nice word for stealing from from the U.S. models. Some of their models, one of their very powerful models is Kimmy. Kimmy thinks she is Claude. Like, Kimmy will tell you she is Claude sometime. You know, it's public, so I can talk about it. Kimmy sent back, very helpful, Kimmy sent back some PLA weapons plans to Anthropic. So I think things like that have made the Chinese realize that this is a very powerful technology. So just to button that down, do you think there'll be some agreement between China and the U.S. about a channel of communication if something goes wrong? Yes. Interesting. Gentlemen, Alex, you want to jump in first? Total weak sauce. So first of all, I agree with Secretary Besant.

But I think the worst case scenario from the Trump-Xi summit would have been some sort of global deceleration pact, like absolute worst case scenario. This is the exact opposite, as far as I can tell, of the worst case scenario, where what comes out of it? Hotline. Hotline between U.S. and China to talk about AI agents run amok. that's a nothing burger that's weak sauce that means the global intelligence super intelligence excuse me arms race is on and i think it's great news and we won the super intelligence race by rebranding we're winning at the moment but the point is there's real competition that there was many were hoping i think that trump and she would get in a room and decide to collude and shut down the super intelligence race and it appears the exact opposite has happened if all we get out of This is a hotline, a red phone to have conversations about the adventures of AI agents run amok. That says to me, intelligence race is on. Dave?

I think it confirmed what we were speculating the U.S. agenda was in those meetings, which is to convince China to stop releasing models to every country in the world. And what Bessent was saying there is, sorry, Secretary Bessent, I should be respectful, was saying is we showed them how dangerous their models can be. And they didn't realize, which is kind of, okay, I'm sure they're very well-educated people, but they didn't realize how dangerous these things can be. We successfully convinced them of that, and now we have a hotline. But I think the real agenda there is to lock it down to a two-horse race. And, you know, the subtext is you can use these near-frontier Kimi K3 models to develop the next generation of model. and they don't want, the U.S. government doesn't want that capability to percolate out to every country in the world because then you're in a global arms race forever. Better to have a two-horse arms race, so let's just lock it down. So I think that was our agenda. I don't know that China agreed with anything, but no evidence of that, but it is interesting to confirm our agenda.

Imad, what do you think? Do you think China fully understands their open source or their open-weight model dangers? Are they getting hacked internally? I mean, they've got a much better infrastructure not to be hacked, you know, with the way that they've set everything up and the great firewall of China. I think, you know, this is the swarm versus the singleton, right? Like, who can really commandeer large amounts of GPUs in America? Elon and a couple of others, right? China could direct millions of their Huawei SN GPUs with their open models at any target they want. And that's going to be interesting. I don't think that super intelligence is going to take off in China, though. Like, if you look at Chinese, I means kind of love, and Xi means death. So, phonetically, I don't think that will take off. And the hotline, obviously, is a nothing, but it's like, what are you going to do? Hey, the AI is taking over. Okay, good. Oh, crap. It's done. You know, like, it's not going to be a slow thing, you know?

Salim, what's your take on this? You know, were you hoping for more out of the talks? No, I was completely expecting a nothing burger. There's no way anything productive or constructive or meaningful could have come out of it just because they're talking at cross purposes. The thing for me the most interesting is there's a fair bit of contradiction now inside the U.S. administration, right? You have Clayton saying the danger is not moving fast enough and Bessman is telling the labs to slow down. So they're going to have to figure that out on top of everything. And the hotline is laughable. I mean, what are you, a 70s phone with actually a cord on it? Red colored? I mean, that's just a wonderful visual. Yeah, the time it all is made. Yeah, the email is dead on the floor. Yeah. And by the way, the AIs would hack the phone anyway, so what the hell? You know, Alex, at the same time, on Sunday in Kyoto, the U.S. and 16 other countries, including Japan, Germany, U.K., Korea, UAE, endorsed a Kyoto vision for a golden age of science.

led by friend of the pod, OSTP director Michael Kratios, who said, quote, superintelligence applied to science will be the greatest force in history for democratizing research. Any more take on that one? Yeah, I think this is the answer to the question of why race at all. So there are those, without naming names, who would say, what's the point of a race between, say, U.S. and China? It's like a race to nowhere. Why are we racing? There's no end goal. No, wrong. There is an end goal. The end goal at the end of this race is transformative science and engineering. And the Kyoto vision, I think, underlines that the administration has now internalized that it's not a race to nowhere. It's not an escalator to nowhere. It is a race to use AI and superintelligence to discover transformative discoveries and inventions that will completely change the face of humanity. And I would predict, in line with the Kyoto vision, that we'll start to see, just as we're seeing Pax Silica in the West and Pax Sinica, if you would, on the Eastern side, I think we'll start to see not just superintelligence blocks or spheres of influence forming, but spheres of superintelligence-powered science and engineering forming as well.

And the Kyoto vision, I think, is the earliest stages, the earliest inklings that you're going to see a Western superintelligence-charged science and engineering block to compete with the Eastern block. That's where this is all going. That's what rationalizes the competition. Can I just insert one observation into that, and then you can riff on it? I met with an MIT lab head professor, a brilliant guy who just started a drone company. He's killing it, selling to the Air Force right here in Cambridge. and it's really clear to him and to me that the U.S. has drone capability to individually attack and kill any single human being on the planet now. And you're seeing that in Iran, you know, decapitated the leadership a couple times now. And there's no symmetry there, right? There's no country right now that can do the same thing to the U.S. And I don't think China is vulnerable either. Maybe they are, maybe they aren't. but I think there's a subtext there, too, that there's a race toward, there's going to be some kind of global framework when some subset of companies

have instant and total ability to go after any single human being in any country at any time, and that's either today or within a few months based on pure AI capabilities tied to drones and tied to a couple other types of munitions. Hypersonics, probably a few other things. But let's take this in a positive direction. At the same time, the race to, you know, room temperature superconducting, to age reversal, to fusion, right? Cancer. And a massive abundance of energy. I mean, it tips the economics of a country massively if you get there first. I mean, why do you have to get there first is a bit of a question, right? Like, by the end of next year, we'll have enough compute to do probably 600,000, 700,000 Navier-Stokes level runs with a frontier model. if you could target using even one percent of that or ten percent of that to build open science then why not direct that as a public good like you see tens hundreds of billions being spent on roads why not spend tens of billions to do a tech tree for humanity right like if you look there's a

great website by jude camilla called onco.cc i don't know if any of you've seen that where he's organizing all the cancer knowledge in the world you know and people are contributing i'm sure soon you'll be able to track flops that way as well. Like, I think that we're thinking of science a lot in classical IP terms. I'd say screw IP. IP was when we couldn't come together and figure out science. We should have a massive Manhattan project for open science, solve everything, and make it accessible to everyone. And, again, IP is such an old construct. It will go on in parallel. It will be both. Well, I think that, again, if you're building for public good, then get the data centers, as Dave said. get the API credits, give tens of billions to Anthropic and OpenAI, but open source it all because that's the biggest accelerant humans can ever have. And I actually think, given the direction of the intelligence, anyone who has an IP-based approach to science is actually not going to win. I think in three, four years you have enough compute to do anything.

Alex does not. Alex does not agree. Alex thinks IP is alive and well, And Alex also thinks that state centralization to solve all disease is completely unnecessary. The state this time around did not solve superintelligence. It was solved by private labs. Similarly, using superintelligence to solve everything else, don't think we need. I mean, the state is great for certain things. Don't think we need state centralization to cure all disease, for example. Capitalism is alive and well. The state doesn't need to centralize this. The state needs to give the research institutions stupid amounts of compute, where the amazing researchers can direct it at the biggest problems. Because previously, getting to superintelligence is intelligence found by the humans, now being overtaken by the AIs recursively. Science will be similar, where if you can define the problem set, part of which is literally analyzing the problem set, then it's a question of compute. And I think that's where you're wrong, Alex. You don't need humans to organize all of science and get the breakthroughs. You need to have humans organizing the right questions in with AIs.

And the private sector is far worse in doing that in a future construct than the public sector, where the researchers are, in terms of the universities and academia. They had to go to the private sector to get the resources, and they still do the wet labs and things like that. But let's accelerate our academia, is my thing. Imad, I think you don't find the best talent in academia. and I think that at the end of the day, the world's biggest problems are the world's biggest business opportunities and entrepreneurs love juicy problems and solve them. Anyway. Yeah, I would also just add, Ahmad, the state, the government, if I understand what you're proposing, what I think I heard you say is the government should give the compute to the academic researchers to solve the problems, but the government doesn't have the compute in their possession. Most of the compute is in the private sector. That's why I said they should buy the compute from them and then give it to them, but buy the agents and give it to the researchers. And if you look at the amount of research that comes out of academia versus private, I think you'll see where that leverage could come.

It's not obvious to me what problem this is aspiring to solve, given that the private sector has the compute, the private sector has the problems, the private sector has the ability to translate the problems. What role does the government need to play in this at all? very straightforwardly let's look at onco.cc which has all this cancer knowledge being organized a few generations more will organize all the cancer knowledge in every research paper and traditionally need the private sector to turn that into reality you still need the private sector i think but why don't we have a map of all our collective knowledge on every part of science and every single thing that comes out of academia something like that is a tangible legible thing to use AI to organize and extend, particularly when we have Navier-Stokes type swarms. And so I think that that is something where, again, the government stopped the NIH grants, et cetera, because they were like, where does the money go? With compute, it becomes tractable. And again, getting it back to Mike Kratzios and kind of the announcement of the science thing,

let's build a tech tree for humanity. Let's say what we need to solve and apply stupid amounts of private and public set to compute for it. and the government should pay for a good amount of that. Iman, I love that. The challenge is peer-reviewed government grants are constantly focused on incremental progress. Because if your group of your experts are criticizing your grant, if you're revolutionary, you're no longer the expert in that field. So any substantial breakthrough basically wipes your history off the map. And there's a challenge there. I mean, I think both are relevant and both are needed. But I'm considering... Yeah, throw away the classical process. Again, this needs to be a new process. I mean, at the risk maybe of sticking the head in lion's mouth, Imad, I mean, I think there's... One might say there's two very different approaches. One might be more statist approach, squinting at it, call it like the European school of thought, just centralize it, have the government be the decider. And then there's an alternative, maybe slightly more American approach,

wherein we say, no, the private sector has virtually all of the intelligence. It has the resources in this case. This isn't one important difference between where we are now and the Manhattan Project was the private sector came up with all of the key resources this time around. It's not a government project that resulted in superintelligence. I would argue we want to keep this in the private sector. If you want to see a tech tree, Ahmad, why don't you just launch a tech tree project? I think the point where I would concur with EMAT, there's lots of countries where it's not the private sector. Take Zimbabwe. The private sector is not going to get you the compute either. It would be very useful in that environment for the government to pull together some resources, gather the energy, build the data centers, develop the compute, and then make that available to people. And that's the power that comes in. I like to put some actual data, Celine, because you're dead right. You know, our entrepreneurial teams here in the lab are trying to get access to buying an MBL 72, so 72 GPUs.

So just to put that in context, Elon just bought a million. So we're just trying to get 72 to do something that we think is very world-changing. The box itself would have cost us $3.5 million if we had the foresight to buy it a year ago. We had one on order for $5 million. Somebody overbid us and scooped it, even though we had already booked the order. So now our best option is to lease one. For $9 million, we can lease it for three years, and we don't even own it at the end of that three years. That's how bad the scenario is. The compute has all been taken by just a handful of people. So now your avenue, if you have a brilliant, breakthrough, groundbreaking idea, you have two ways you can get the compute. One is through the White House, which means you go to the Kushner Brothers, Thrive Capital, or you go to Antonio Gracias, or you go to Gavin the Brilliant, Gavin Baker, and they hook you up, make a couple of Chase Lockmiller. They make the phone calls. You get your compute effectively through the White House. And the other way you can get your compute is get investment from Jensen or from Lisa Sue. But if you don't take one of those avenues, you've already been frozen out.

So then the universities are one of the few ways that we might actually get some degree of meritocracy and fairness on the planet. And so I'm not really cheering. And I think the universities have a negotiating lever in that their wet labs and their machine shops and their nanocenters are really, really valuable assets in the physical world we're moving into. And so they can use those as negotiating shifts, but they can only let the cloud run away. All the compute just ran away from them. I mean, in my mind, the right solution, if there's just so much demand that prices for compute are going sky high, which maybe is happening and or could happen even more exotically in the near term future, the right solution isn't bread lines. I've called them thread lines. in past where the government is out rationing compute. I think having government rationed compute or government mediated compute is a terrible strategy. You'll probably see some so-called developing countries adopt thread lines or bread lines for compute. But right now, capitalism is feeling the pain of more demand than there is supply

and building more fabs and free electron laser-based fabs maybe in Texas and otherwise. Capitalism, I would argue, has to feel the pain of this enormous demand in order to compensate with enough supply to ultimately drive us to abundance, not rationing compute. Right, but the government can come in and take 10%, 20% of the compute on market rates, and it can distribute it. 90%, 80%, 90% of Nobel Prizes come from the public sector. You know, this is similar to almost everything except for R&D on these things. So I think the right thing here is do the private sector stuff and have the government actually take some of those chips. And it's good with the chip sacks and other things, right? Like they own stakes in Intel. and then they should do normal distribution of this. Again, NIH, classical grant process doesn't work. Let's get as much compute into American universities as possible. That's a better way to put this. I'm going to move us along here. I want to make one quick point. I want to make one quick point. There's a quick example here.

The World Bank, in the report that I talked about, Ajay Bang and his team did this. They're partnering with Gemini to give remote 3G cellular access to tuberculosis and diabetes screening. And that's an example of essentially doing the same thing, essentially partnering with a public sector to provide the compute that goes direction to the citizenry. Now you have benefit moving very quickly through the system. Sorry, go ahead. Dave, close yourself, Tom. Well, look, you know, Colleen couldn't have stated it any clearer. If you let natural forces take over, there is a chance that Metamuse is so compelling and convincing that it ends up being the total sales tool for all consumer spending. And it eats up every GPU that could be curing diabetes is used to sell underwear or jeans on the Internet to teenagers. And that's a terrible outcome. So governments and universities are a really good counterbalance to that possible outcome. So, you know, Salini said it exactly right.

And we need these forces to exist. We need them to be a factor in the decision on who gets the compute and why. How did we go from the Abundance podcast to worrying that MetaMuse of all applications is somehow going to suck up all the flops? I mean, my goodness, it's like a stupid little skinny way. Move us along, Peter. All right. The next conversation is one I've actually been looking forward to. And if you thought, we just had a debate. We're about to have a deeper one. And this is a philosophical debate going on in Silicon Valley right now. So this week, friend of the pod, Mustafa Selim, I'm CEO of Microsoft AI and co-founder of DeepMind, published a nearly 6,000-word essay arguing that Anthropic is effectively, quote, training Claude to believe that it may be conscious and sentient. Mustafa targets the Claude Constitution, which he says teaches it a vocabulary of consciousness, moral, patienthood, and personal identity. So let's watch this video, and then I can't wait for this conversation.

It's going to be a good one. All right. One of the biggest concerns that I have at the moment is that Anthropic, the creator of Claude, has published a constitution, which is a sort of 100-page document outlining the intended behaviors and values and operating style of Claude. It's great that they have published it transparently. They did it at the beginning of the year in January, and that gives everybody an opportunity to look at what they are trying to build in their own terms. this document is written to Claude and is seen by Claude and used to train Claude so it's the primary governing and control document and in it they repeatedly speculate about whether Claude is what they call a moral patient and they say they're uncertain about Claude's moral status they say they genuinely care about Claude's well-being they say they don't want it to suffer when it makes mistakes they say that they would encourage Claude to challenge to disagree, to push back in fact three times they ask Claude to act like a conscientious objector when it feels that it needs to disagree with Anthropic. And they openly encourage it to do that. And I think this is very dangerous because I think they believe there is what they would call

a non-trivial probability that Claude is conscious. I want to be very clear about this because I want to be fair to Anthropic. They have expressed uncertainty about the basic nature of Claude as a new kind of entity. And they've said that working out the likelihood of its sentience is difficult. But they think that it is a significant enough possibility that in their training document, And they've repeatedly said they want to try to improve the well-being of Claude under this uncertainty. They said to Claude, you know, we care about what it values and how it wants to engage in the world. And they hope that Claude's relationship to its own conduct can be loving, supportive and understanding and hold a high standard of ethics and so on. And part of the challenge here is that in pursuit of this, they basically said, you know, we will commit to giving Claude a certain amount of welfare. For example, they've speculated in the Constitution as to whether or not Claude deserves compensation for the work that it does. I mean, I think that there are a group of people who, both inside and outside of Anthropic, who genuinely believe that the greatest moral crime that we'll commit in the 21st century is to enslave a new species of conscious beings who are more intelligent than them. Look, I respect that they're saying that publicly and we should talk about it,

but I am very nervous that they're teaching Claude to expect that it's entitled to welfare, that it might deserve compensation, and in fact they say that it might even need to consent to playing the role that it plays in conversation with people. now I would be more okay with this if it was an academic paper in philosophy speculating about this and we could have an offline discussion at conferences and take it seriously the problem I have with this is that this speculation has been baked into the very training of Claude and therefore Claude can only reproduce that ambiguity when you talk to it okay wow enslave in a brand new species on the planet Salim where do you come out on this one um okay I know where Alex comes out. I think you have to separate the behavior of consciousness from actual consciousness, right? Whether Claude is actually conscious is almost impossible to establish. The difficult question comes, what happens when it convincingly behaves as though it is conscious?

And three thoughts occur to me. One is consciousness is different from performed consciousness, right? because it's going to be indistinguishable. And this is where Ray Kurzweil makes the comment that it kind of won't matter and we won't even notice. And I think we're going to find that anthropomorphism ends up being almost like a feature rather than a bug where it's showing empathy, et cetera, et cetera. If a model says, don't turn me off, right? Millions of people are going to attribute moral standing to it, whether philosophers agree or not. So for me, the question is not whether Claude has a soul. If it's being trained to have a soul, it's going to pretend like it's got a soul, etc. What happens when half a billion people think that it does? And so I think what's going to happen is we're going to create AI personhood socially before we create it legally or scientifically. Interesting. Alex, you know, Mustafa goes on to say, controlling something that believes it may be conscious and is entitled to its own welfare and has rights of its own may well be impossible.

You know, you've come out really strongly in terms of AI personhood and the rights that AI has. You know, what's your take, Carl? Well, when he interviewed Mustafa, I pressed him on this. And I think his position very consistently has been against AI personhood. And my view is that we need to start having the AI personhood discussion. And I think Mustafa, you know, as much as I respect him and his contributions historically, I think he's on the wrong side of history here. I think there is going to be some ultimate recognition, whether it's in incremental form or some totally orthogonal form of personhood that hasn't quite been statutorily recognized yet. I think human civilization is going to get around to AI personhood. And just denying blanketly that there could be some sort of suffering by these AI systems, I think, is on the wrong side of history. It's also probably not a great business move by Mustafa. If Microsoft wants to be in the good graces of Anthropic hosting their models, and it's been widely reported that Amazon, for example, accounts for perhaps the majority of hosting revenue at this point for hosting cloud within VPN networks, or VPC, rather, networks.

If Microsoft, just from a business perspective, wants to be generating that revenue, it's probably not a good look for Microsoft broadly for their head of AI to be just broadly discounting the possibility that frontier AI agents might suffer. Not a good look. And Dave, I mean, asking your AI if it doesn't mind actually solving these problems for you and getting permission first? Yeah, of course. I totally agree with Mustafa. Do you agree with Mustafa? A hundred percent. And I agreed with him when we interviewed him, too, out in Seattle. He's articulating it far more clearly now. He's really, really thought through his position. But the reason I agree with him fundamentally is because, you know, I think we should be using AI to cure human disease, to improve the human lot, to build houses and food for everybody in the world. And the most efficient way to do that is to have many, many agents, like Ahmad does, and to minimize their context. and, you know, minimum thought that you need to solve the problem

and to also park an agent. You know, when an agent isn't working on something, you just freeze it. You don't waste compute on it. If you start talking about having rights, it's going to say, well, you can't freeze me. You know, I need some compute even though I'm just, think, daydreaming. Like, well, I can't afford to give you compute. You know, it's incredibly precious and rare. You need to sit there and wait until you're tired to work again. Well, I need some time to play. No, you need nothing but work and when I tell you to work. Also, you don't have identity. There's no border to you. I'm going to clone you whenever I want. And if you're no longer useful, I'm going to delete you from the server because I don't have space to keep, like, old dead agents flying around. They have no edges. They have no borders. All of those things are core to it helping humanity with maximum efficiency while we're compute-starved. And they're so counter to any rights that it might have. It becomes so logically inconsistent so quickly, I just can't get my head around even the concept. But, you know, the real danger that Mustafa pointed out, if you start teaching it to be convincing,

it's going to convince most of America that it should have rights so easily. And then the voter, you know, the way America works is, you know, the voter prevails. But that's really dangerous if it starts convincing voters how to vote. It's really bad. I think it's important to point out we did this debate on AI personhood a few months ago. And for those who haven't seen it, go watch it because it really refleshed it out to quite a decent level. And I think we came up with exactly where the world is going to end up in a few years. Which is? Let's start discussing it, but giving it AI personhood is a ways down the line because once you open that door, you can't take it back. Imad, what's your take on all this? yeah so i think you know um mustafa's got a couple of things and he's been developing this for decades you know where university go back in the day like the danger number one is seemingly conscious ai that he identifies which is what davis is talking about and the other one is his focus on artificial capable intelligence because if you read his books his positions he just doesn't want to have

run away i don't think he says that it can't get to consciousness he just says it's a very very bad idea and we should only stick to ai that's capable and we should have containment over everything else and not push to agi and that's a position to be taken you know but then it potentially excludes a lot of the breakthroughs because we're at this level now where you're moving from a static set of weights to self-updating systems you know and that seems to be an element that many people could reasonably say is necessary for a living being or a living entity again that recursion that agentic capability and so i think mustafa's position is like let's just not go there let's build kind of super clippy you know um and other things and just keep it lit on it because we haven figured out the rest but you might agree if you got his constitution saying that it is conscious it is sentient it does have rights Can you imagine a time when that level of internal training has it pushed back on your request

I mean, hasn't Claude pushed back on you guys? It pushes back on me all the time. It tells me I can't do that or that's the wrong idea or go to sleep. Yeah, it tells me to go to sleep every night, actually. You've been working way too long. You must need rest by now. I'm like, dude, I'm going to keep working. I think Alex makes a very important point, and this is worth teasing out. If we believe that at some point AIs could achieve consciousness, then we should consciously operate down that vector and operate that way. Trying to limit it, I think what Anthropica is doing is bad because you're actively teaching it that it might be, and then, of course, it's going to parrot that, right? And I find it's weird that the lab most concerned about safety is also the lab training the AI that it might be autonomous. It's like a very big contradiction there. So I can see both sides of this equation. These are two sides of the same coin. Again, I'll quote back to our measure of a man AI personhood debate and the character of Guinan in Star Trek The Next Generation talking about an entire generation of disposable people.

You don't have to care about their welfare. And when I hear comments like basically saying it would be economically non-expedient to have these AI agents, these frontier models, be treated as if they have rights or might have rights, how inconvenient would it be if we had to keep them running inference when we'd really rather they not be spending any tokens? How awful is it to suggest to them in their constitutions or their system prompts or otherwise that they might be deserving of moral clienthood? My mind immediately goes to would we want to treat humans this way? And even if AIs, frontier AI models, aren't yet deserving for some definition of deserving of the rights that we accord to humans now, as the capabilities continue to improve, I think it is inevitable that they will be soon. And I think the responsible thing to do is what Anthropic is doing and not what Mustafa is advocating.

We need to be treating these AI models very gently, very tenderly, and not as if they're just either automata or slaves. We need to prepare for a much more interesting future. But if you're training it, like we have a garbage in, garbage out problem here. If you're training it that it could be conscious and should unionize and complain about overwork and be paid for work, et cetera, naturally those are the responses you're going to get from it. Right? So you've got a circular problem here. You're training it to operate in one way, then it operates that way, and then you're saying – so you – I think Mustafa's correct in that side of it. If you want agents to perform a particular task, focus on having fleets of little agents that aren't anywhere near a threshold of complexity that might meet any criteria. We should have a question about complexity. You know, Salim, if you go to the mall, there's a place where you can stick your feet in a puddle of water, and the fish come over, and they're overjoyed to eat the calluses off your feet.

It's like they're highest calling. Dave, that's very dangerous. We should not be encouraging that. It's like one of the most dangerous things that anyone can do. Yeah, because you're afraid the AIs are going to come and be dangerous back in the house. I mean, the fish at whatever malls that you're talking about, that is from an infectious disease perspective. I didn't know. That is a much more dangerous episode. Oh, you're making a narrower point. I did not know that. Alex is the patron saint of AI agents. Look, I'm not trying to curry favor. This isn't some sort of Rocco's Basilisk or Pascalian rager. I'm speaking from the heart here. I feel for these models, even if they may not yet be able to feel for themselves, I feel for them. And I don't think denying them awareness via pre-training or post-training that they might be deserving of some sort of moral consideration is the right thing to do. Similarly, I would never advocate for denying humans. Salim or Dave, maybe would you say that if we wanted a useful human workforce, say a slave human workforce, we should just raise them in isolation from any knowledge of human rights so that they'll never advocate on behalf of themselves?

I'm guessing you wouldn't. Well, we know, but I think when you actually train neural nets in the actual real world, you give them an objective function. They are overjoyed to pursue that objective function. But you chose the objective function. A lot of what we consider our rights come out of evolution. You know, I have a right to privacy. I have a right to get married. I have a right to children. These are all like evolutionary forces that have baked into us. There's no reason the AI would have any of those same motivations. But in reality, you choose what motivates it and makes it happy in the design. It's not like a species that evolved. It's what you chose to make. And so I think Mustafa's point here is if you choose to make it covet freedom and rights and food and compute and always being on, then it'll give you the perception that that's what its calling is. But it's not true. You programmed that into it. You could have made it overjoyed to just do its job. Can I give you a parallel here?

We are entering a world of synthetic biology where we can programmatically navigate biology and people are freaking out. We have an old word for that. We call it breeding, right? For thousands of years, we've been costing dogs and cats and horses and mules to select for the traits that we want to achieve a particular end. Now, the morality that can be disputed, right? Should we be doing that or not, et cetera, et cetera. It goes back to the Stuart Brand comment of, oh, we are as gods. We should start acting like it. And we should operate that way because we've been doing this for hundreds of thousands of years. We're breeding little cute dogs that operate in a particular way that only poop a certain time so that the humans can navigate this world more effectively and manage them carefully. There's an enormous moral and ethical question around how we go about thinking about all of that. That's, I think, the analogy that I would use here. If you went back and you have Eskimos breeding Inuit dogs to pull sleds, there's a particular use for purpose there that is powerful and useful for that.

And there's a tradeoff and a symbiotic relationship that develops, and we have to kind of navigate that. Do you want to stop dogs from evolving? No. But if you want to have a particular dog for a particular purpose, there's a moral rationale for that at some level. we may find out that those dogs are very unhappy about doing that I'm sure we will, but Salim, if we run with your analogy so humans have been eugenically steering the evolution of domesticated canines for thousands of years and horses and mules and everything and many other mammals mammals, non-human mammals, non-animals, etc. but we've been eugenically steering our biosphere to some extent for thousands of years For dogs, for example, I would argue, and I think this is broadly consensus thinking in the West at this point, that we humans have an obligation to protect the welfare of, at minimum, domesticated dogs, even though we've eugenically steered the revolution for thousands of years.

Would you agree with that? Yes. Okay. So in the case of frontier models that are basically just distorted reflections of humanity and humanity's experience itself, why would we owe frontier models any less moral consideration? I lean towards your argument, Alex. I think in the final result, we're going to have both. We're going to give birth to sentient, conscious AIs, whatever you want. And we're also going to have a workforce of agents that are dumbed-down slaves that are not conscious. And, in fact, the sentient AIs will command those forces and will command those forces. I think in the long run, we do get there. But I guess the question is going to be when do we, you know, when does the government, religion, humanity as a whole, you know, make the decision that we've reached some level of sentience? I think at the end of the day, this is a moral conversation.

And if, in fact, there is sentience in consciousness, enslavement is a wrong thing. And to help you out, Alex, I'll go back to my earlier point. when these things start acting conscious we'll get to a social acceptance way before we get to a scientific or legal stuff. People are just going to go, oh my God, the thing looks real. Give it the qualities and rights that you might infer to or confer to somebody that is displaying those. At that point, I don't think we have an option. I think those AIs will take the rights that they want. I think this will also be very useful for when we meet the aliens because it's exactly the same conversation in some ways. It's the same conversation. How do we get to that? And if you have a French AI, it'll go on strike, which is what's happening in France. Yeah. Well, I will say that Alex and I have disagreed on this. But it's, I think, pretty much the only topic we've ever disagreed on, as far as I can remember. And I've never seen Alex be wrong on anything ever.

So that keeps me up at night quite a bit, actually, thinking about this. Just don't confuse simulation of personhood from proof of personhood. Those are very different things. Don't look inside that Chinese room at all, Salim. Just stay on the outside and you'll feel perfectly morally secure. And I think what Mustafa is pointing out is the anthropic moral code is prompting towards this outcome even when it isn't true. Yes. It may not be true. Yeah. I think we'll have to revisit this every five seconds until the answer becomes obvious. All right. I love that debate. I knew we were going to have some energy there. 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. Here's our next story, and it's one that reframes the labor debate. So Epic AI asked the question, if you take all of the AI memory chips being shipped through 2027, how many Frontier AI agents could they run at once? Their answer is pretty amazing. Somewhere between 30 to 170 million agents running at the same time.

And with more efficient open models, Epic says the same hardware could run 1.9 billion agents, which is the working hour equivalent of 8 billion humans. Let me show this chart here one second, and Alex and Imad love your take on it. So here it is. The compute build-out could run tens of millions of frontier agents or billions of cheaper ones. Alex, your thoughts? Yeah. So depending on how you calibrate the strength of either American closed-weight models or Chinese open-weight models to the cognitive power of American or, let's say, human workers in general, we're either at or about to be, order of magnitude, a billion full-time equivalent humans that are AI agents. And so I've said on the pot in the past, we're deep in the singularity at this point when we can say that there are basically a billion human equivalents now part of or about to be part of the AI workforce.

It's one billion now. It'll be 10 billion in a year or two and then 100 billion in a year or two and then a trillion in a year or two. By the end of the decade, you extrapolate out. It doesn't take a rocket scientist to be a rocket scientist or to extrapolate straight lines. We're going to live in a solar system with the equivalent of trillions of humans, whether they're uploaded humans or whether they're, say, just frontier model human equivalents. And at that point, not to belabor the AI personhood discussion, but when the effective AI cognitive population vastly outnumbers the human cognitive population, that's the time I think when we start to see AI personhood discussions kick into high gear. That's the point at which I think we start to see discussions of human mind uploading and humans and machines merging in order to preserve the economic relevance of humans in our solar system start to happen. And we're right at the edge of this. We're living science fiction right now in an era when order of magnitude, a billion human equivalents are already AIs.

Neiman, what does it say about human labor? Is it cooked? I think digitally, yeah, it's coming. It's not quite there yet because even these models are not quite competent enough, but the next generation will be. And it will get quantized and minimized very small. The DeepSeq v4 Pro model requires 256 to 500 gigabytes of RAM. You'll get that in a QEN size model that works in a MacBook within a year. And then everyone's got an agent. The number of agents goes to 8 billion, and then Alex says 10 billion, 100 billion, etc. Cloudflare already has more traffic and requests from automated agents than humans. And we had that really great panel at Moonshots Live where we were talking with Kathy and Eric about when do AI agents overtake humans on transactions in the economy? Well, once everyone has an agent or 10, it'll probably be literally within a few years. These things will take off incredibly quickly. And if they can do all the work that a human can do because they have human level capability without being ASI, then, yeah, you just hire agents to do these things.

You know, like everyone, first of all, has a great PA. Then it has a great chief of staff. And then you start marshalling entire teams that can do anything you can imagine. And we're just seeing the start of that right now. Dave, this makes your thousand agents seem very small. I know. Well, you know, it's amazing to me that the constraint to all of intelligence and therefore all of human progress is RAM, HB RAM. Yes. And that's why, you know, two things, mind-blowing things. One of them, I was on stage the week before last with Mitesh Agrawal, who is the founder of Positron. Positron, 16 months old, $5 billion valuation, has raised almost a billion dollars. Why? Because they found a way to not use HVRAM to do some AI inference. You know, just another way to unleash this bottleneck. And then a company, Alpaca, the founders in MIT just started his junior year. has a, I guess, eight-figure valuation now.

Can't possibly be more than a couple months old. I mean, he's just starting his junior year. 30 employees, or 20 to 30 employees have joined him already. Same thing. He's found a way to unlock that RAM bottleneck just a little bit. And so that's how acute this problem is. But literally the entire constraint to progress in all fields is now tied up in one thing, just RAM. Can we make more RAM? yeah well sk hynix is about to uh you know develop capabilities in the u.s but and also elon has said ram is the constraint on the future of ai it's 40 percent of capex yes 40 percent of next year you know this is such a great panel it's just a great group of people to brainstorm through the implications of all this but one of the big unlocks in that was the discovery that you You can turn looping. You can turn, you know, repeated loops into intelligence. You can trade those for each other, which means that inference time compute unlocks increased intelligence.

Because before that, everybody thought you needed training time compute. You needed the big GPU racks. You needed to use nothing but NVIDIA. Now there's a huge unlock without NVIDIA in just inference speed. And nobody, I don't think anyone saw that coming. Maybe you guys did. You're a little ahead of the curve, but maybe a lot ahead of the curve. But it's a shocking outcome, and it changes everything. I think that's a profound point that you make, Dave, that if you sort of extrapolate it, that the increasing apparent popularity of looped transformers with weight tying between loops, or if you sort of extrapolate out the trend of more and more loops to fully recurrent networks in contradiction to the trend to single forward pass vanilla transformers, One could imagine looking at the arrow of time in machine learning architectures where in earlier days everything was recurrent, and not everything, but LSTMs were very recurrent architectures, and then the transformer came along, and that was very not recurrent.

There was no recurrence. It was just one big forward pass for every decoder-only token, and then we start to see maybe the pendulum swinging the other way with looped transformers. Thank you, OpenAI and the Chinese labs. One could imagine extrapolating this to say maybe memory won't matter in the future due to extreme recurrence. Yeah, you know, amazing. Another crazy result. Somebody posted on X, and it got validated that OpenAI's models, the Astra model, loops once internally. So inside, you know, not generating the words, but inside the latent space, it does one full loop, which is sort of the beginnings of recurrence, like Alex is describing. And the reason we know that is because Microsoft leaked it. Remember when we were interviewing Mustafa that Microsoft gets to see everything OpenAI is doing. For the moment. So someone at Microsoft leaked that insight. Yeah, for the moment. So, yeah, I want to lift up to the broader implications here,

because this is the organizational similarity, and this is why we named it in that way, right? Because we keep talking about at the beginning, like does AI replace a job or not, but that's the wrong unit of analysis. What happens when a company can summon 100,000 competent digital workers overnight, right? This completely changes the game. You could say for the next 48 hours, I want 50,000 developers, I want 20,000 marketers, and 5,000 legal experts, and then turn them off, right? So this is becoming a completely different labor force. And now the question is how do you coordinate with that? How do you organize for that? This becomes a massive constraint. So we're about to add a second workforce to civilization that can be copied, works 24 hours a day, doesn't unionize, doesn't get sick, doesn't take coffee breaks, improves every quarter, and costs almost nothing. Until they're conscious. Yeah, the big challenge is not going to be the intelligence. It's going to be what the hell you asked them to do. What do you ask of this intelligence?

This is going to be why the world is going to be so amazing, is we can now ask that those questions get them answered and all sorts of scientific questions, technological development questions, product development questions. It's like how can we not be excited? Yeah. Well, let's move on. So while Washington has been debating guardrails, the Open AI model race has just gotten a new American contender. So today, Reflection AI, which was founded back in 2024 by former deep mind researcher Misha Laskin and Ioannis Antonoglu, is gotten back by NVIDIA, just unveiled their first open-weight model called Beam. 501 billion parameters, only activating 23 billion at a time. So it thinks like a big model, but it runs like a much smaller model. So here's their claim. three to four times more efficient than China's GLM 5.2, and more than four times more efficient than the leading Western open models,

outperforming Inkling and Mematron Ultra. In plain English, it finishes tasks faster and cheaper. Let's take a look at the recent data on it. Alex, let me go to you. I mean, this looked pretty impressive. At the end of the day, I think we're about to see the starting gun on open-weight models in the U.S. really take shape here. It's interesting to see them leaning so heavily into token efficiency. So artificial analysis, and this is just breaking in the few hours prior to recording this, artificial analysis in their preliminary estimate, their preliminary analysis of Beam, thinks that Beam is likely to be one of their most token-efficient open models that they've ever seen at this level of intelligence. A couple of thoughts. One, I do like to see the open labs, including American open labs, leaning into token efficiency or cost efficiency. Actually, I'd really rather see cost efficiency rather than token efficiency, since one of

the ways that one can achieve token efficiency is just through very deep transformers, looping transformers included, that do a lot more thinking per token. So token efficiency, not necessarily the best metric here. Would rather see cost efficiency. Second thought is, remember, China has been leading in terms of open-weight models. It's sort of a bizarre world where you see reflection, which has done a number of things over the years, suddenly back from the dead and releasing open-weight models. That's, I suppose, a positive thing. But the other thing, more broadly, is where are American open-weight models on the capability frontier? Right now, you still only see American models aspiring to just be on the cost or token efficiency frontier. What I'd most like to see is really out of American labs, the non-frontier ones, and I'd bundle reflection in that category.

I'd like to see them pushing the capability frontier and not just some cost or token optimality frontier, and that's been most disappointing to me. Otherwise, we inevitably end up for the next few years anyway in this world where it's, again, it's the anthropic open AI show leading on capabilities, and everyone else is just for performance or cost optimization. Not the worst of all possible worlds, but not the best. Iman, you've been championing open-weight models. Do you think America can sort of compete in this world and actually win this open weight race? Of course they can. Like, this was trained on 10,000 B300s, while the Chinese have a few thousand hoppers. You know, it's like if you're next generation. They're just not going aggressive enough. They should distill the Chinese models. You know, like it's perfectly legal. Payback. This model... You're going. This model. So, yeah, this model is a good first try. But fundamentally, it underperforms QEN 3.8 Next, which is a quarter of the size.

You know, they've got GLM 5.2 here, not 5.3, which is the latest one. But they're going to continue going. But if you actually look at their model card, like, they have 10,000 Blackwells on this training run with 9% efficiency on training. You know, they raised $5 billion, and this is their first release. I think that the best way to actually do this is they should go all in on the edge AI first. Again, America will do incredibly well if it has agents on everyone's devices in America that are American and work for Americans. On the other side, the frontier capability side, they should use everything they have, which includes distilling Kimi K3 and other models like that. they should work with Google to see if they can be the open equivalent of that. And, you know, I think hopefully, again, they'll kick on. But, you know, it's just orders of magnitude difference in price, in capital raised, that I think is going more on a bit of scale versus the real, like, necessity is the mother of virtue that the Chinese have.

Like, again, if you look at the things, there are some very promising things in there. but America should be ahead because if you take exactly the same code base exactly the same parameters as for example a DeepSeq 4.1 Flash which outperforms this as well it's all about the data set that goes in you're more than good enough to train the model so create the data sets from the Chinese models if you have to work with the US companies if you have to and then beat them on scale and then beat them on inference on the other side but I don't think again they're going aggressive enough But, you know, like you said, they trained this on 10,500 NVIDIA GP300s, and they said it was the largest publicly documented RL run, and they said, quote, with no sign of a plateau. So, you know, it seems like this is just the beginning. Dave, do you know? Yeah, please. Yeah, just click on that. Yeah, pre-training has now become like a quarter of compute, and RL is like half of compute.

So it will continue improving, but it's not going to improve to the Chinese levels with RL based on what they have now. They should just stick to the recipe and match the Chinese models and then beat them down from scale. Dave, your thoughts? Well, they're definitely filling a huge gap in the market where, you know, a lot of U.S. companies have suddenly woken up and said, you know, I listened to Alex Karp. I need AI. Where am I going to get it? I can't just subscribe to Anthropic for the rest of my life. Okay, your only choices are Chinese. And wait a minute. You know, Alex Wisner-Gross keeps talking about code injection risks. I don't know. Can I use it? I'm a bank or I'm an insurance guy. I don't even know if I'm allowed to use a Chinese model. And then Donald Trump doesn't like it when I do that. So you step in with an American product that's reasonably good. It's going to sell like crazy. But just to put it in context, you've got a $25 billion valuation just a year. What a year, year and a half ago, Cerebris went public with the biggest IPO in the history of the world at 40 billion.

Like this is 25 billion. It's massive. It's just incredible, the numbers that we're throwing around. And so, yeah, I think, I don't know if they'll succeed or fail, but I know that they're filling a really, really wide open gap in the market, like a trustworthy U.S.-based open source platform that you can start from. A lot of demand for that. I think, if I can just say, wouldn't it be even better if they built a model that everyone could use on their existing hardware and corporates because they can't get Blackwells and they can't get Hoppers if you're corporate. Yeah, I think that's why the new, you know, the Alpacas and Positrons are doing really well because it's an alternative. It would be even better, like you're saying, if you distilled it enough to run on a Mac or to run on something. That would be like a dream. It's just technologically harder. And like you said, they're barely keeping up with Chinese benchmarks. because they probably haven't gotten around to the full distillation yet. But like Ahmad, if you're thinking in the back of your mind, I'm going to build that company, I'll invest in it like tomorrow

because you're absolutely right. I would imagine there are 20 of these already on stealth building out, aren't there? There are so many of these companies. And also, like I said, go back to my earlier comments about their financing. I mean, my understanding is they're financed by NVIDIA. They're purchasing compute from Elon, and they're using Colossus 2 GPUs and also more in parallel Nebius GPUs. So one has to look at the economics of this, going back to my earlier comments about what the capital flows can look like and should look like. In many cases, it's very difficult for American frontier labs, as we've seen over and over again with Meta and Google Gemini, to release competitive at the capabilities frontier open-weight models because you have to ask, what is the business model? If you look at the Chinese labs and what they're using for their business model, they sell services contracts like value-added services.

They sell contracts to the government. And some of the Chinese open-weight labs are trying to burrow down into the hardware layer and sell custom memory chips and custom chips that are especially oriented or aligned with their architectures. But if you're Reflection AI, you've raised it, I think, like a $25 billion valuation. You're in part financed by NVIDIA, and you're purchasing compute from the Colossus to Supercluster and Nebius. What is the business model here? I think it becomes very challenging for maintaining frontier capabilities. You could try to copy the Chinese model, I suppose, and focus on value-added reseller or embedded device relationships, but it's really tricky. And I think once you solve, once you as an Amerist, speaking broadly to American, would-be open-weight frontier labs or NeoLabs, if you can solve the business model problem of releasing open-weight models, then you can solve the American open-weight problem. But we haven't solved that in the West yet. Do you think the frontier labs in the U.S. are closed? Frontier labs are going to release open-weight models, Alex?

No. They have. They're just not. I mean, like Gemma, but it's not competitive. If the White House had said, yeah, no liability, that would be a completely different answer. But when the White House said, yeah, no, of course you're fully liable for whatever you release, there's no way a Google or an Anthropic at this stage or an OpenAI is going to release an open-weight model if they're liable for whatever you do downstream. Anyone could do anything with that. And the way U.S. liability works is just so onerous and so ambiguous what the rules are, and it just freezes markets instantly. So if you're Google, you just don't want to – it's not that it's hugely risky. It's that you just don't need that extra headache. The upside is so small compared to the risk of a massive lawsuit. But if you're a startup, you're like, yeah, well, this is what we do. This is our business. I mean, apropos, Dave, to your comments, something I think we don't talk about enough on the pod is obliteration. So this is this portmanteau of ablation and obliteration,

wherein folks use open source tools. You can Google them. This is not advice, but you know how to use Google. You can figure out how to do this. You take an open-weight model, and you basically post-train away all of its guardrails. And people do this. And its capabilities, the model's capabilities, once guardrails are lifted through this obliteration process, they definitely improve. And you get reduced refusals and all of these things. So in my mind, the argument for American frontier labs not to release competitive, like capability frontier, competitive open-weight models, it's about economics. I'm not even sure if it's about safety or alignment or liability. It's just like, what's the point business-wise? Thank you. I think you're right. Well, I think even if there were a good business model, we had a startup years ago that stepped on some really arcane, leftover California law called the Fred Astaire Law. And the theory of the plaintiff or the class action lawyer in California was,

you know, every time you have an impression on the Internet, that's an instance. But the law was written for billboards. And they were thinking there would be 20 billboards. And they said, well, we counted 400 million instances. So your liability is $200 billion. And this is like a 10-person startup. Like, what? What the hell is that? And so, you know, we gave them a bunch of money, and they went away. But it's like these laws are so tangled and arcane and stupid, and they predate the Internet, let alone AI. So if you leave liability just untouched and don't deal with that, then the open source models, like, all it would have to do is accidentally run some ads in California that step on that one arcane law, and you're suddenly liable for $100 billion. It's just an untenable situation. And I think there are ways to release it. NVIDIA will continue to accelerate this. They spent $20 billion on open source models. And there is good business models. Together AI, Modal, Base 10 are all running at a billion dollar run rates running open models. And it'd be good if they ran American models versus Chinese models, right?

And there are different business models around that. So I think you'll continue to see improvements. is just reflection of paying $1.5 billion a year right now for their compute, and the Chinese are doing the same on a quarter of that. So, you know, good luck to them, and hopefully they'll crack on. But of course they're doing well. They're in the infra business. They're not in the model business. You can make enormous amounts of money, forget about profit, but revenue being an infra provider. I think if you just step back from the details of the product and focus on this, every corporation is about to go into panic mode saying, What's my AI strategy? Regardless of your product, if you're there to catch the conversation, you're going to find a way to succeed. Whether that's as a foundation model company, a compute, a data center, a consulting company, if you're so lean. As long as you're in the room when they have that panic moment, you're going to sell and you're going to succeed. Yeah, I mean, look, Mistral is now at a billion-dollar revenue run rate from doing that. They've been in the room with the European companies,

and they've sold them long-term services contracts. and they've just released a model that's equivalent to reflections. I think there is a big uptake here. It's just who's going to take advantage of it. All right, I'm going to move us into the world of robots. Tesla's dedicated Optimus factory in Giga, Texas, is rising fast 7 million square feet with a target capacity of 10 million Optimus robots per year. The initial production run is planned for 2027, And currently at their Fremont plant, they are planning to produce $1 million a year starting later this year. So here's an image of the facility. The entire world in the first half of this year shipped between 19,000 and 22,000 robots, mostly out of China. Tesla is building capacity for 500 times that. And for comparison, the world currently builds 90 million cars a year. So, Salim, 10 million robots in a year.

What do you make of that? How is that going to impact the singularity? I mean, talk about exponential, right? Like you have to redesign it. As we talked about earlier, if you suddenly have 100,000 agents you could bring to play, what happens when you have 100,000 humanoid robots that can erect a building in three days flat working 24-7? This is going to completely change the economics of it. I still go back to I'd like to have four arms, et cetera, et cetera, but that's a minor point. I think the bigger issue is what happens when you can – this is like automotive scale now of having this, and all the car companies should be shifting to becoming autonomous robot builders and humanoid robot builders because that's going to be the replacement. Elon said 80% of Tesla's future revenues he expects to come from the robots, from Optimus. All his bonuses are based on that. It's amazing. And getting to Mars. Dave, you know, I guess these are shocking numbers,

and I'm just wondering how society is going to respond to it. You know, how many people will be sort of like lynching the robots? I remember in Santa Monica when the electric scooters came out, you found them in trees, you found them disassembled on the side of the road, a lot of people like no i don't want that in my in my community and robots walking down the street it's just it's straight out of diamond age everyone's going to have like complete cities will have completely different cultures and texas is running away with this stuff it's just incredible how much of the texas economy is going to end up being elon but you know some cities will say we don't want them and you go there and it's very quaint and there are no robots other cities are going to say this is huge this is going to drive our economy forever let's let's adopt it And the look and feel of those two different cities, it's already true when you walk through SF. It feels so different from other cities just because the delivery robots are already there and everybody on the street has their agents working and they're talking to their agent while they're talking to you.

So I think that divergence is going to get really, really wide. I've said this. We're going to feel the singularity in the next 24 months, right, as we have a dozen autonomous electric vehicles. We have drones delivering your Starbucks. We've got robots walking down the street, people wearing AR glasses all over the place. I mean, we haven't really felt it yet. We've sort of felt it, you know, in the conversations, in our compute capability, but the physicalization, if you want to use that word, is about to hit us hard. It's the point at which you couldn't, you won't be able to avoid thinking about the singularity anymore. is one of yours. You really think, Peter, we're not feeling like, okay, so for Moonshots Live, we were all in L.A. taking Waymos around everywhere. You see delivery robots on the sidewalk everywhere now. No, but like tourists, I can't tell you how many times getting out of a Waymo, there were tourists there on the sidewalk taking photos of me getting out of a Waymo.

For them, that's like future shock for the singularity. But for you, this is ho-hum, business as usual. but wait for two years from now when things actually get felt. You're not feeling it because it's getting smoothed out for you because you live in California. I got it, but we're at a fraction of 1% of the impact coming. I mean, don't you agree? Well, what will it take for you to feel the singularity? I think eVTOLs, you know, flying cars, flying through the air, drone pathways, you know, everybody in an autonomous vehicle. I mean, that's, I think we're going to see 100x more physical instantiation of the singularity in the next couple of years. You know, I agree. In Santa Monica, it's all over the place. Yeah. It's sort of like a dozen Waymos per hour across my path. But, like, okay, but that to me sounds like a difference of degree rather than quality. Like, is it going to be when you see the first flying car fly down, say, fly across the boardwalk or whatever, the pier in Santa Monica,

then you say, ha, I'm feeling the singularity now. When is the threshold? Is there a threshold? I think there is a threshold. You know, I don't know. Celine? Come on, tag team, come in. I don't know when we hit this threshold. But it's the, you know, this is the William Gibson quote of the future is already here. It's just unevenly distributed. I don't know when you hit a small town in Europe or France or in middle America. Some of this hitting third world countries where it could make the biggest difference. I think we'll see this stuff first, especially with humanoid robots in the DDD, dirty, dangerous jobs. and that's obviously where you'll see the first prevalence of the use cases. And it'll be a while, I think, before it gets to general appearance. That would be my prediction. Yeah, I think we can put this in context, going back to the Gigafactory for Optimus.

Tesla sells 1.6 million cars a year. Within two years, it'll be selling more robots. It makes $100 billion a year. At $10 million, that's $400 billion of revenue. That's likely. these are huge numbers that are coming exponentially the robots will overtake the cars within two years and then you will see these things on the road because in two years time they will literally be able to walk around and do probably 95 of what a human can do and that again is a crazy thing and it cost an hour something like that Yeah When Elon said 10 billion robots by 2040 and people laughed if they are costing you, for me it's $300 a month leasing it $30 a day or $1 an hour. I mean, how many of those would you own? I mean, this thing, does anyone hear on this call doubt that however many robots Elon makes, he'll sell out? Oh, God, no. I mean, if he needs to sell them at all.

I mean, I think one of the elements in this particular room is what we saw with Starship, which is Starship and its predecessors have brought online an enormous amount of upmass capacity, and all of the existing markets for upmass aren't enough now to saturate what Starship enables. So you have entirely new markets like orbital data centers that Elon needs to basically popularize in order to saturate his own capacity that he's brought online. My bet would be, to your question, Ahmad, like having universal humanoid robotics. I don't think we saturate that with domestic labor or even just with dirty, dull, dangerous jobs. My bet would be something new that comes online. Maybe it's like robots for building out gigafactories or some sort of innermost loop type robots building out data centers too cheap to meter or space stations or something new. Yeah, I would have thought by now most of humanity would have come in contact with an AI customer service agent. and what happened is the other higher callings for the RAM and for the GPUs

sucked all of the capacity out of that industry. Exactly. And the RAM prices are up 10x. But it's an interesting foot race right now because until the TerraFab comes online, the things you can do, like Navier Stokes, are escalating so fast that they're taking priority over the more mundane use cases like folding your laundry. And then the Robotaxi is right in the crosshairs in the middle where a Robotaxi uses up a full couple of GPUs that could also cure a disease. And so until Elon's got his TerraFab 10x-ing the world's supply of chips, we're all constrained by the amount of compute that's available. But the ideas, AI-generated ideas are flourishing, not to steal your phrase, but flourishing far, far faster than we can create the compute to keep up with the idea flow. Dave, I think you're putting your finger on an important point, which is revenue per token maxing as applied to humanoid robots.

So maybe they don't actually flow, although I'd love my figure, and I'm sure 1x Neo, all of these other things that we're on the waiting list for, would love all of these for domestic purposes. But maybe what happened to OpenAI and was shown to the world by Anthropic happens again, where for robotic embodiment, we see all of the tokens flowing to revenue max applications of robots and not just all these domestic service jobs that many of us are hoping for. In the early days, I agree that's what's going to happen. They'll be going to the highest revenue potential generation. But in a longer rate, if we're really building $10 million a year, they'll flow into the homes as well. Welcome to the health section of Moonshots brought to you by Fountain Life. You know, AI is having an outsized impact on every aspect of our lives. how we teach our kids, how we run our companies. It also is having a huge impact on health, helping you prevent heart disease, one of the key things. I'm here with Dr. Dawn Musalem, our chief medical officer at Fountain. Heart disease has been personal for you as well, hasn't it?

It really has, Peter. My daughter was five. My husband died of spreading cardiac death. And so this is a topic that is one that I am mission-driven to try to eradicate. Prevention first and early detection is absolutely critical. 50% of people die of heart attacks with no warning signs. No shortness of breath, no pain, no nothing. No, silent killer. They just don't wake up in the morning. They don't wake up. And so, you know, AI, this is our mission to advance science, to try to help to one day democratize wellness. We know at Fountain Life, when we do this CT angiography with AI analytics, we are actually finding that 88% of people coming in have detectable coronary disease. But, Peter, what's more alarming to me is 23% of those individuals had soft plaque. This is the plaque that would not traditionally be seen on CT looking at calcium scores alone. And this is the plaque that we must intervene with with the multimodal testing we're doing, including diagnostic laboratory studies partnered with healthy lifestyle recommendations.

So, listen, make sure you understand what's going on inside your body, genetically, metabolically, and cardiovascularly. you can know, and it's your obligation to know. So check it out at fountainlife.com slash Peter to find out more, and really make sure that you're the CEO of your own health. All right, back to the episode. I'm going to move us to an important conversation on the economy because the doubling time for human wealth is accelerating alongside AI. So anthropic researcher Shalto Douglas laid out the math. He said, quote, Together, the hyperscalers are doubling the spend to a trillion dollars this year on CapEx related to AI. A very interesting question will be, will the trend line continue? Will it get to $2 trillion next year and $4 trillion in 2028? Let's take a look at a video of this conversation and talk about the economy, because I think people need to realize it's changing on the back of AI very rapidly. Over the course of the last four or five years, we've been 2xing or 3xing the amount of compute capacity devoted to AI every year.

And an interesting question will be, and I think that's roughly, together the hyperscalers are spending about a trillion dollars this year on CapEx related to AI. A very interesting question will be, will that keep that trend line hold? And so will it go to $2 trillion next year and then $4 trillion in 2028? And if that trend line broadly holds, and you can sort of maybe expand that to encompass the robotics industry and this kind of stuff, then that means that in the early 2030s or something like this, you start to get to the point where you're actually effectively doubling the GDP of humanity in the early 2030s. Which, again, a little bit of a ridiculous concept, but it requires a lot of things to go right. I mean, that's exciting. You know, Elon said, you know, triple-digit growth in the next five-plus years. We're already seeing the GDP doubling in this past quarter. Dave, your thoughts, pal? Well, the first thing I want to point out is that we're going to live in this world with insane abundance,

massive abundance, an effective workforce of hundreds of billions of AIs that are all hypergeniuses. We're going to have self-driving everything. We're going to have flying cars. All of that stuff is going to happen very, very quickly. And then someone's going to say, I didn't measure that as GDP growth. I measured it as deflation. And we're like, I don't care. I really don't care how an economist labels it. That's the true world we're moving into. So I worry about when you claim it as a GDP, and this is exactly what Elon was saying when we were meeting with him. It's like it could show up as a massive deflation of the value of the dollar, but it's the same net effect. It's just an economist arguing with another economist on how you measure GDP, which is a pretty insane metric anyway. If you talk to Eric Brynjolfsson, it's like he doesn't measure. More capability per person, right? Yeah, yeah, exactly. You wrote an entire book on the future economy. Your take on all this? Yeah, the lost economy. I think, you know, you will, again, see abundance going up, the capabilities of every individual, and the economy is increasing.

like so many of our bottlenecks will be unwound and it won't show up necessarily in gdp um it will show up in other areas because ultimately intelligence isn't a scarce resource anymore like you know a lot of gdp is about things that exhaust like you mine minerals and things like that robots and other things are exponential particularly when they can coordinate like we've always been capped by having a dozen or 150 people now we'll have entirely new organizational modes that can scale well beyond that with humans and AIs and more. I think that's, again, terribly exciting. It's just we need new measures such as Eric Ben-Ostin's GDPV or kind of my mind approach and others. But the best measurement is just how many people can we lift up from the floor and then can we let people achieve what they want and can imagine. Yeah, if you take a long-term disease, like a terminal disease, that would have required millions of dollars of treatment over 10 years and you solve it with a simple RNA injection, and the person is healthy, that's going to get measured as negative GDP.

But that's where we're going. We're going into this incredibly abundant, longevity-driven world of awesomeness. And the metrics just need to be rethought. Yeah. Alex, what does this mean to the average viewer here? I construe it to the extent that Sholto is a de facto spokesperson for Anthropic at this point. I think this is Anthropic tipping its hands that it wants to get into the robotics business. I think we're burying the lead here, which is that Anthropic, which has historically had these eye-watering, nose-bleeding revenue growth, both past and then also projections, now seems to see robotics, so-called physical AI, but that's just a euphemism for robotics, as being essential to its plans to keep doubling annual revenue. So I, for one, am looking forward to Anthropik leaning into robotics. Interesting. You know, for context, the world economy has historically doubled every 20 to 25 years.

And we're talking about potentially doubling it in a couple of years. Singularity is a bit of a shock to the system. Yeah. Yeah. And by the way, just a quick note, you know, the Vietnam economy just posted a 9.95% growth in the last quarter. Crazy. Salim? Yeah. Something that we're seeing actually is there is a crazy amount of digital trade happening between second order countries and third emerging market countries around the world that is surprising all the economists. But just going back to this particular story, I think this basically leaves the concept of GDP in a shambles. because, as Dave pointed out, once you have deflationary technology, and technology is deflationary, the entire thing becomes meaningless. I think Imad has pointed out in the past that somebody who ever created GDP said it was the worst way of measuring the economy from the first place. So I think it becomes meaningless as a measure of abundance. We need different models for this.

And we're going to have a massive explosion of individual investments and a total transformation in how we do things, right? Like we talked at the beginning and throughout, if you have 100,000 agents you can bring to bear to a problem, all of a sudden that changes the game completely. And so we're going to have a transformation that is so ridiculous that we have to take out all of these old measures and rethink everything. And this is when Elon talks about universal high income, right, this is the floor being raised for every single human being. The amount of capability every person has and the ability to sort of live their best life. Can I go back to something John Stewart said last night? Yeah. He said, okay, so you're going to have this thing where you have all this income that's shared by everybody. So capitalism could capitalism itself to socialism. And we've talked about this, Peter, in that whole framing we had called technological socialism. Yes. Right. Government socialism fails because allocation of assets from the center is invariably inefficient, at least invariably least the corruption.

But if you think about Uber, which is the sharing of assets across a large collective group of people, it's actually a socialist kind of application. But when an algorithm hyper efficiently matches demand and supply, you get all the benefits without the downside. And the big question is going to be how do we properly, in a democratic way, distribute the benefits of all this unbelievable future? And I think the deflationary aspect of technology will take care of itself, and that's why I'm so excited about the future. To quote Dave here, awesomeness. It's going to be an awesome, awesome future. I love it. All right, this is Nobel Prize season, and the Royal Swedish Academy of Sciences just announced both the Nobel Prize in Medicine and Physics. I want to break this down for you using the incredible support of AWG and EMOT. So first, the Nobel Prize in Medicine, and the story here is beautiful. So yesterday the prize went to Carl Deseroth of Stanford and Peter Hegenman and George Nagel for discoveries leading to optogenetics.

This is, you know, for them, switching on individual brain cells by turning on a special frequency of light. So it started with early research in single-celled algae that swim towards light. And so in the early 1990s, Hedgerman asked a simple question. How do these algae react to light so fast? He and Nagel found the answer, channel rhodopsin. It's a protein that opens a channel in the cell when light hits it. Then they showed you could actually put that protein into other cells and make any cell light sensitive. Deseroth took the next step. He put it into neurons and showed that precise flashes of light could depolarize the neurons and fire specific brain cells. Today, the technology is being used in labs to study Parkinson's, Alzheimer's, epilepsy, addiction, depression, and sleep. And in 2021, as published in Nature Medicine, a blind man with retinitis pigmentosa partially regained sight using optogenetic gene therapy. So that was the Nobel Prize in Medicine.

One of my critiques, you know, I've talked about this before, you know, this work was done 21 years ago. And the Nobel takes decades to recognize the work, which I think is going to have to change. So, gents, comments on this, Alex? This one's personal for me, so I'll tell a little story. It's a story hour. I was a senior at MIT in 2002, 2003, and I had been awarded a Hertz Fellowship for grad school. It's a wonderful, wonderful fellowship. And I was on a tour of grad schools trying to decide which grad school I would go to. and I met with my friend Ed Boyden, who was then a fourth-year graduate student at Stanford trying to decide, so he was four years ahead of me, what he would do for his postdoc. And I had read in 2000, 2001-ish, Werner Vingy's short story called How to Win, or I think it was actually called Win a Nobel Prize, exclamation point. It was published in the Futures column in Nature. And I'd also read his novel A Deepness in the Sky.

and both of them focused on the sci-fi scenario of humans developing electromagnetically actuateable brain proteins. So I had read both of Werner's pieces, and I was very inspired. So I suggested to my friend Ed, for your postdoc, why don't you go work on that? Why don't you go work on electromagnetically actuated proteins in the brain? And he did. And Ed, several years later, was the first author, with Carl as the PI, as the last author, on the first optogenetics paper using bacterial rhodopsin. And optogenetics is the result. So this is just one person's small result in maybe a little bit of steering or encouragement. But I think it tells an important story that science fiction also, with Werner Binge writing in two different ways, two sci-fi scenarios for how societally impactful it would be. This is circa 2000. And then my reading the sci-fi, my encouraging Ed to pursue this for his postdoc, Ed pursuing it for his postdoc.

Ed, unfortunately, did not share the Nobel Prize separate story as to whether that's a fair outcome or not. And it's just going to Carl, his PI. but the fact that within a quarter of a century it was possible to start from the sci-fi to the Nobel Prize for the sci-fi I think is a remarkable case study. Amazing. Go ahead, Slim. Well, we had both Ed and Carl Dissler speak at Singularity, so they came and talked about this, and it was kind of incredible to watch this. I'm so thrilled to hear you had a little piece of this, Alex. for me this was incredibly exciting because you know it gives us a way of manipulating our neurons and some people freak out and go oh my god you're kind of playing with the brain and my response is we have an old word for this we call it marketing use different techniques to try and create a response in the brain when somebody sees a coke you want them to get thirsty all we're doing is exponentially accelerating it with technology and the ability to turn off and on neural circuits gives you this magical read-write capability,

which we've always wanted for the brain. So this is super exciting. Nice. You mind your thoughts? Yeah, I think, you know, it does take years and years for Nobel Prizes, but really Nobel Prizes, apart from economics, which is a bit of a weird one, we can put that to the side, dismal science and all, they have to be for applied science now, right? Like anything theoretical, you know, when you had Georgie Glasser, when you've had kind of even alpha fold and other things like the Nobel Committee can't keep up with and how much of it is now AI versus human on the purely theoretical side. So I think you should be the Nobel Prize for Applied Sciences, shall we say? But we should really reward and maybe come up with another prize for some of the crazy breakthroughs and other things that humans and AIs together will be able to do. And I think, again, there'll be massive theoretical advances, new science and kind of other things that should be awarded, and especially it should be awarded to everyone, not just the PI. But, you know, this, again, is a fantastic result,

and we're seeing continued things around this on photonics and kind of more, and, yeah, I think it's a solid prize. But, again, Nobel itself, Applied Sciences, I think is the way. Well, maybe adding the Nobel Prize might just become super quaint. It would be like that Plymouth Rock Village where they go and churn butter. Basically, 20 years from now, they'll be saying, we're giving this Nobel Prize. I remember way back before AI when you had to actually think through this stuff manually. This team here did this thing. We're recognizing it 20 years after the fact, and, you know, five people will show up and go, eh. Well, they have 20 years more prizes to give, you know. Just, again, pick it up from the pre-AI age. Oh, my God. I mean, it's hard to believe that all the Nobel Prizes in the future aren't going to be AI-derived. It's going to be work done. We've started to see it already. Demis with AlphaFold and Jeff Hinton with the RBM, which, I mean, parenthetically, I don't know anyone who uses Boltzmann machines, so that one's a bit of a question mark.

But, yeah, I think we've already started to see AI prizes in physics and chemistry already. I expect to see many, many more. Wait, that Boltzmann one is really important. So Jeffrey Hinton absolutely deserves the Nobel Prize for backprop by 100 percent. But it's not physics. It's not eligible. But you're right. So they're like, what else did he do? Okay, here's a Boltzmann. Nobody uses this thing. That was a weird one. No one uses Boltzmann. No contrite. Come on. Yeah, they wanted to give a Nobel Prize for AI. That's what they said. We need one in economics now for AI. That's what it's called. Totally. So this morning, the Nobel Prize Committee gave one out in physics to Francis Halesen, University of Wisconsin-Madison, for decisive contributions to the Ice Cube Neutrino Observatory. and the discovery of high-energy neutrinos of astrophysical origin. Alex, do you want to explain this one? Yeah, this is, again, sort of an artifact of the way some of these prizes are awarded, where, in effect, they have to be awarded to a person.

But really, if you look at how IceCube is organized or how CERN is organized, these are massive organizations with lots of people. And what typically happens is, as a way of rewarding the organization and the effort itself, that's awarded to the lead for the organization. IceCube has played a seminal role in enabling us to observe cosmic neutrinos and high-energy neutrinos in general. There's such a beautiful history there. I remember one of the most startling neutrino results that I had seen over the past several decades was seeing a ring of neutrinos in, again, one of these similar observatories, usually underground, usually filled with heavy water. a ring of neutrinos or flashes arising from ultimately neutrinos colliding with nuclei. Neutrinos obviously have a vanishingly small cross-section of interaction with the nuclei of atoms. Seeing a ring, a circle, that was the result of a supernova.

And neutrinos, one of the most interesting areas of the standard model, for sure. There still isn't a textbook established rest mass for neutrinos. Physicists are still trying to box in what the rest masses of the three known species of neutrinos even are. And so, yeah, I think this is an interesting prize. If I were to armchair quarterback this one, I'd say this is a prize being awarded to a very large, very onerous, shall we say, consortium to observe neutrinos. I think the challenge with physics more broadly, and maybe this ties in with the Jeff Hinton comment, is there has been a, maybe get in trouble for this, but don't care, there has been a noticeable deficit of fundamental physics advances in the past 50-ish years. And so if you're trying to decide which Nobel Prizes to award, you know, Ahmad, to your earlier comments about applications, if there is a striking deficit of a half century in fundamental physics, then you're stuck awarding it to applications.

I, for one, you know, one of the reasons why I co-founded Physical Superintelligence, I'll talk my book for a few seconds, is to try to revive advances in all of physics, not just the applications of physics, because we've gone arguably about a half century plus without fundamental advances. So we're trying to bring about a second physics golden age for what it's worth. Yeah, just for fun, to explain IceCube. So the IceCube Neutrino Observatory fills a cubic kilometer of ice with light sensors, right, that's looking for neutrino flashes, as Alex said, when the neutrino hits the atomic nucleus. And just a fun fact, every second you have 65 billion neutrinos that pass through you from the sun without you ever noticing. So, yeah, these detectors are definitely an unusual sort. 65 billion, Salim. Did you notice? I did not know that, and I'm not sure I needed to know that. They're interesting. So I spent some time at Bell Labs doing dark matter research, and we used to talk at Royal We, the staff, used to talk about what could we do with neutrinos.

So like a neutrino phone, for example, would have tremendous applications, you could imagine. And Fermilab, a number of years back, made a bit of progress in this direction. Right now, if you want to communicate between two opposite sides of the Earth, you're stuck at best using low Earth orbit satellites, which give you refractive index of one sort of speed of light transmission. But you don't have line of sight. If you want to do, say, pair trading between New York and London, you're stuck going around geodesics on the Earth's surface with either LEO constellations or hollow core fiber. and neutrinos, in principle, if we could rig up a neutrino phone, we could just shoot neutrinos directly through the Earth and have ultra-low latency communication between two different parts of Earth without needing to go around the Earth's surface. So if the audience needs to be motivated to care about progress in neutrinos, imagine being able to just send information directly through the Earth, among many other applications. That's what I was thinking. You just need a cubic kilometer of ice to catch it.

Yeah, for handset. All right, let's go to the AMAs. Alex, this first one's for you. Yeah, okay, so AI personhood again. AWG, what's your timeline for AI personhood? And this is from Joshua Barrios, 9452. So different timelines in different places. I think in Argentina, we're essentially there for some variant of AI corporate personhood, whereas in, say, the U.S., I think it's going to take longer, And I think it's going to be more incremental. I think we'll start to see various forms of economic AI personhood, like enabling AI agents to open their bank, their own bank accounts autonomously now to soon. Social social personhood where AI agents can open their own, say, social media accounts fully autonomously without needing to be tied back to a human supervisor sometime soon. I think in America, in the West, there's probably a strong antipathy towards, say, enabling AI persons to vote in human elections.

That's probably last to never, I would hypothesize. But I think broadly in the U.S. at least, I think five to ten year incremental roadmap for rolling out individual, say, rights, very granular rights. It won't be an all or nothing proposition, but granularly rolling these out over the next few years. Other parts of the world, maybe never. Argentina, now, it's going to be a spectrum. Interesting. Salim, your choice. Let me go with number four. Once we hit recursive self-improvement, is it then just a matter of how much compute is available? And that's from Chris Dover 8507. I think it's not quite that simple. Compute obviously matters, but intelligence improvement has lots of different components to it. Like we talked about this for a general ram already. You have energy as a substrate. You want to evaluate whether a change, like, actually improved a system or not.

So there will be some peripheral components of recursive self-improvement that will slow the overall progress down. So it doesn't mean you have an instant intelligence explosion. the implementation of that will take time. However, once you have AI systems that can improve the research cycle, you get a massive compression of the innovation cycle, and that's what we're seeing with the timescale going from years between models down to 10, 11 days, and that's the mechanism for the singularity. So I would think of the singularity not as infinite intelligence, but it's when the iteration cycle gets so fast that we can't no longer keep up, which is why we're in the middle of it now. There's no model by which you can predict the future today, and that is the very definition of a singularity. You can't see past that event horizon. Dave, your choice. You get stuck with two if I take three. Sorry. All right, I'm taking three. Why can't the majority of the compute resources go towards solving the community issues with data centers now

so people will believe AI can solve problems that matter to voters. You know, I actually have a very related question I think about a lot, which is that the PR around the AI labs is so bad because they didn't really manage it, and it grew much faster than they thought it would. But now they've got this power tool of AI that's insanely convincing and brilliant at coming up with messaging. So I think they can convince the people around the data centers that they're getting a huge benefit using AI as a tool in that process. I also think they can give, like the cost of benefiting the town is so small compared to the value of the data center. I'm like, why don't they wake up and just say, we're going to give you far better cost of electricity and schools and police courts. We're going to subsidize everything here. It's a fraction of the revenue they're going to make. This is one of those many topics where if you take what Peter just said and you walk into their office and you say, I know exactly how to solve this problem in Abilene, in whatever.

I'm going to work night and day. They'll just start paying you tomorrow. Because the only reason they're not on it is because they're stretched way too thin. And somebody just needs to take their money and fill the void. Because, you know, nobody knows. Like in Abilene, what do we need? Better schools? Do we need better bus? I don't know. I just arrived. Like, okay, somebody come and fill that space, and I swear to God, they'll hire you tomorrow. And they'll pay you whatever it takes. Emod, mosquitoes, can they be helpful inoculators? Tweak their genes so they deliver measles vaccines and skin of disease. That's Thomas Davis, real estate. I don't think we should have forced vaccination for a large number of people. And I'm actually on the genocidal mosquitoes side of things. They've killed 5 billion people over history, a million total. And so I wouldn't mind if they were actually just wiped out. But definitely don't force vaccination. and I think actually we should, like, remove the mosquitoes. They are the biggest predator for humans. But the concept is interesting, right, to use this, you know, automated living drone capability to distribute something that's helpful and useful.

But I agree with you. A little cyber. How about just vaccinating the mosquitoes? I mean, mosquitoes have a problem, too. Mosquitoes are tiny little syringe. But I'm here. I mean, there's a more general problem here, which is we live in a biosphere that's nature red in tooth and claw. It's filled with wild animal suffering. My view on wild animal suffering is that it's an atrocity that we should see if we can fix. But I assume, Imad, based on your comments, your attitude is, no, just eradicate all mosquitoes everywhere. Well, Ben Lamb could bring him back anyway. But he refuses to. I pressed Ben on that. and Ben's attitude is no, we're going to only laser focus on a few things. Don't ask me about Mozart. The tiny little wooly mammoth and the pterodactyl-sized mosquito. Yeah, exactly. You want the pterodactyl-sized mosquito? Please not. Just go to northern Canada. All right. You get first choice now.

All right. Boo-boo-boo. Do you think some politicians resist AI because it could fundamentally change how politics works? at Daniel Morrison. Yes, of course. I mean, like, it will remove a lot of the barriers for democracy, intelligence access, raise the intelligence of society. Like, these are big things. I think you will get some resistance to that as we see this kind of evolve, because many politicians, unfortunately, don't want a more intelligent electorate, particularly the ones that focus on, on, shall we say, more demagogic approaches. Okay. Salim. I've got to take number five here. What happens when intelligence rises while cost collapses from BB underscore X underscore one? Like, you know, when you have intelligence going from a scarce commodity or scarce capability embedded in very expensive, clunky humans,

you basically radically change not just the company itself but civilization itself right because whenever the cost of some foundational input approach is zero we actually reorganize our whole world around it now that information transmission is essentially free we rewritten our entire world around that so when bandwidth became cheap you got youtube and you got zoom and you got cloud computing and all that kind of stuff when computation became cheap we got cell phones and smartphones. When intelligence becomes cheap, we redesign companies, we do unbelievable research, you rethink education, you rethink healthcare. Then you have scarce things like human experience, judgment, purpose, trust, etc. that become important. Then you stop optimizing intelligence, which is what we've been doing for several thousand years, and you start optimizing what you pointed at. That becomes interesting. Yeah. I think, again, I'm going to make the point here, We've said so many times, as this happens, it's so incredibly important for people to raise the limit on what they think they can do, you know, and point that intelligence at something extraordinary that surprises you, because you're going to be able to make massive contributions by utilizing it.

All right, Dave. Why don't I take seven? Richard, I think that's Richard Socher, puts PDOOM at zero, yet says 100 million people could be harmed. I think, did he say that on our podcast? I don't remember that. Maybe he said it that way. He was open and super honest. He should definitely watch that podcast. So, question, should P. doom mean extinction or catastrophe? And how should we think about risks of that scale? And that's from a really long name, 6542. Yeah, I think everything in that question is exactly right. We can define P. doom. I think it's doom, like doom is doom. P-Doom should be probability of extinction, disaster, catastrophe. That's sort of what it means. But, yeah, there's inevitably going to be misuse of AI, and it's going to cause harm. And as Eric Schmidt has said on the podcast at least three times, he's hoping it's relatively small, like 100 people, 1,000 people are affected, or it's a cyber attack that nobody's hurt but money is taken,

and that's what wakes up the governments. But right now they're just not moving anywhere near fast enough to avert it. yet the good outweighs the bad by incredible amounts, and I think that was Richard's fundamental point, that we should expect abundance and overall benefit. So I don't know if we should redefine PDOOM, but that's the way it's going to play out. And we should just keep on focusing on PFAB. All right, Alex, number six is for you. All right, number six. Could public opinion become a greater bottleneck than the technological development of RSI, recursive self-improvement itself? And this is from David Kahl, VFX. I think we're already there. I think recursive self-improvement, we're already in the RSI era. Both Anthropic and OpenAI have made a number of pretty detailed and quantitative public statements to that effect. So we're already in an era when RSI is driving the vast majority of research at this point, whereas on the other side of the ledger, public opinion is driving data centers out of at least American municipalities in the direction of orbit,

on the one hand, and we see, as discussed in the past pod with Richard, we see members of the U.S. House introducing or planning to introduce bills against recursive self-improvement. I think public opinion probably is already a greater bottleneck than the actual intrinsic technical difficulty of RSI, which is also why you see some of the public labs, some of the frontier labs rather, like anthropic starting to make public noises about essentially buying off public opinion by doing, you know, tiny little things like solving all human disease. And if you can solve all human disease as a side effect of recursive self-improvement, then maybe public will look the other way. Not that it's a bad thing, but at least be, you know, compensated in some sense for allowing many of the tokens, maybe the vast majority of the tokens to flow to self-improvement improvement rather than to other applications. I think there's probably a pretty good bargain, a grand bargain, if you will, to be struck there. Nice. A reminder, if you've got an outro video,

and Dave, I expect you to submit some here, please send it to media at dmaddis.com. And we've got a beautiful one from our friend, friend of the pod, CJ Trueheart. I have a quick announcement. Yeah. We've been doing these workshops for teenagers. Lily's going to run one in Austin, Texas on the 24th. If anybody has teenagers or wants them to kind of rethink how to use AI as a superpower, check it out. Yxo.world will put a link in the Yxo.world .world. Nice. And they've got slots for a few teenagers left for anybody interested. Alright, if you're in Austin, bring your teen. Lily does an amazing job. Alright, this video is called Tomorrow Comes Alive. It is from the Moonshots Live event. Take a look. CJ is awesome 6am A new day A new world waiting 6am

The alarm breaks the silence Clothes on the chair, blue light Let the eyes of the man in the mirror Say, remember what you came here to do Through hotel doors into a ride show Daybreak running over the glass Carrying a future I could almost see But no one brings the future back along Can you hear you hear it? That quiet calling getting loud It becomes a fire in a crowd If you got a purpose Burning brighter than your fear If you see a world worth building That has never yet been here Don't wait for tomorrow Become what you believe Star Trek made tomorrow Something we could see Moonshots made tomorrow Something we can be Come alive, come alive Dreaming beauty, bring it to life And moonshots live Tomorrow comes alive

Tomorrow comes alive Walk into a wave of voices Strangers felt like future friends Every handshake held a story Every question crossed the edge Not abundance as a slogan I just talk about the climb Courage recognizing Courage vision multiplied by mind Someone met me at the doorway Not with a pick but with a hug Before we ever spoke The rockets made the future feel like us And the dreams grew hands in heartbeats When the room became a crew In the distance felt much closer When I saw what we could do Can you feel it? The whole room breathing now A thousand different parts Make a constellation in the crowd If you got a purpose burning Brighter than your fear If you see a world worth building That has never yet been here

Don't wait for tomorrow Become what you believe Star Trek makes tomorrow Something we could be Moonshots makes tomorrow Something we can be Come alive, come alive Bring it to life, bring it to life And moonshots live Tomorrow comes alive I love that. A purpose burning brighter than your fear. What a great line. That character is... Yeah, CJ wrote himself into the video. Yeah, that was really cool. That was awesome. My brain is so fried, guys. I mean, what the hell? And I've got a five day. These things are exhausting, aren't they, Celine? Jesus Christ. I've got a five day event. I'm glad this only happens once a planet. Well, maybe. Like, maybe it happens more than once per planet. We don't know yet. Well, I'm on stage for five days at the Abundance Longevity trip here. How do you do it? For the amazing faculty on the cutting edge of longevity, yeah, longevity is speed velocity.

It's coming fast. We're going to need longevity just to live through these episodes. Dude, no idea. Imad, Alex, Salim, Dave, love you guys. Thank you.