Why the US Government Is Blocking Model Releases (GPT-5.6) | #267

For the first time in US history, the executive branch has placed a national security hold on commercial AI products.

The models are so insanely capable that they have to be controlled.

Certain models will be available to everyone, certain models won't, and now they're becoming more and more gated, and I think this will only accelerate.

Bottom line, the US government is now in the release loop.

Well, it looks like protectionism.

The government is too late.

Might it be possibly crushing OpenAI and Anthropic's valuation? It's reported that the leadership is pulling back on their near-term IPO.

It's just a different world from being a private company. And I just think they're suddenly realizing, wow.

Our next story is from Anthropic, who accuses China's Alibaba of running a massive distillation campaign against Claude.

This will be the excuse that the US and Europe and maybe South America use because they need to suppress Chinese AI somehow.

The world seems to be on a path, sort of a second Cold War type path.

Anyone who's shocked by this is way out of touch with what China's actually doing. Now that's a moonshot, ladies and gentlemen.

So, Emad, I hear it's hot out where you are in London.

Yeah, it's a 5% penetration of AC here and it's like 40 degrees Celsius. So I was just saying, I'm gonna quit being an air entrepreneur and become a HVAC roll-up specialist. That's the real route to alpha.

Wow, market potential.

Dave, how about you?

I'm in beautiful Quechee, Vermont. It's gorgeous up here.

Yeah, fantastic. Alex, it looks kind of boring back in that background again. Where are you exactly?

Is the background even real, Peter?

No.

You're in a bunker. You've got your upload ready to go, your cans of canned tuna for the next 20 years.

I'm vegetarian, but I do like to say, don't take off the takeoff.

Ah, you know, I just got back from a 24-hour sprint to Fairbanks, Alaska. I was there for 24 hours during 24 hours of sunlight, there for the Wildfire Finals. So, you know, we're just decimated in California, Greece, Australia, around the world with these fires. And about 5 years ago, I said, this is ridiculous. We need to be able to find a fire at ignition, like just at the very beginning, and put it out autonomously within 10 minutes. Well, these are the finals. We have 3 teams that competed in Fairbanks, and we chose Fairbanks because they had the drone approvals. All these were drone companies. Anduril was one of them. So it's great. Palmer's in the competition. When he's at our Moonshot gathering in September, we'll talk about that. Another one from Germany called Dryad and one from a combo of Australia and The Queensland of the UK was called Aura. All three of them using drones and just being able to, you know, fleets of drones spotting the fire and then dumping suppressant on them. So very impressive. Hopefully at the end of this, you're going to have wildfires being a thing of the past, or destructive wildfires, that is. So that was my weekend playtime. Yeah, that's so cool.

What's the coverage like? How many drones do you need to cover, say, all of California?

Well, the goal of the competition was cover 1,000 square kilometers and then being able to find the fire. There were decoy fires. If the fire was more than 2 meters in size or it was moving, then zap it and put it out. So I think the teams were using fleets of autonomous aircraft and drones to do the coverage. Anyway, it's coming, it's coming. And then they're gonna, you know, someday we'll see Optimus robots out there in the field, but get rid of, you know, putting humans at risk and use the technology where the technology is best.

Yeah, I think that's gonna be like the Oil Cleanup XPRIZE. It's gonna be one of those ones that just creates a global best practice in one iteration. And that's the best XPRIZE theme, you know, when something wins and immediately goes into deployment just like Oil Cleanup did and becomes the way it happens for the rest of time.

That's just such a cool thing.

Baseline.

Time to start one-shotting moonshots.

Shall we get into it, gentlemen? You guys ready?

Ready? Let's roll.

All right. So welcome to Moonshots, everyone, the number one podcast in all things exponential and AI. We want to be your front row seat to the coming singularity. I'm here with my magnificent Moonshot mates, AWG, our in-house super genius. Alex, a pleasure. Thank you. Dave Blunden, our wizard of AI investing. Dave. Good morning. And Imad Mustaq, our AI intellect on emerging intelligence. And of course, Saleem is on an airplane. He's airborne at this moment from Munich to Spain. And I'm Peter Diamandis, your host and hopefully your abundance evangelist. So this week came fast and furious. My head is still spinning. We literally spun up a weekend recording because there's so much going on. There's genuinely no time to asleep during the singularity. For those of you joining us for the first time, our mission here at Moonshots is to keep you informed, keep you up to date on exactly what just happened, and more importantly, keep you optimistic about the extraordinary world ahead, the world that we're building, the coming age of abundance. So let me give you a quick overview, a TL;DR of today's pod. OpenAI hit the brakes both on shipping GPT-5.6 and on its own IPO. We'll discuss why. We'll cover OpenAI's audacious plans to use AI to fix security holes, not just find them. And Elon's having quite the week as well. A lot going on in his world. Neuralink may attempt the first brain-to-brain telepathy communication later this year. Micron dethrones NVIDIA, and Trump signs sweeping executive orders to supercharge American quantum computing. A lot to cover. You guys, you ready? You guys ready to jump into this? I mean, honestly, I was like looking at the feed from you this morning, Alex, and I was like, okay, we gotta cover that too, and that, and that.

I think, Peter, we need an emergency pod every morning.

Well, that's the way it's going to be from here on out. I mean, we're clearly in an accelerating hard takeoff. So you got to expect every week is more than the last.

So don't take off the takeoff.

It's a race to keep up. You know, Alex, it's funny you say we should podcast every day, but, you know, it takes half the day to keep up with what happened just the day before anyway. So then you use the other half podcasting it out.

It's the podcasting singularity. Oh my God.

And, you know, I'm really— I really have a high bar for what news we cover. And there's just so much of it. So everybody, you know, strap in, grab your Americano or latte, whatever you have this morning, and let's get into it. We're going to kick off with 3 breaking stories from OpenAI. For the first time in U.S. history, the executive branch has placed a national security hold on commercial AI products. Last week, Anthropic's Fable and Mythos models were initially pulled from the market. Last night, the Trump administration struck a deal with Anthropic, which grants the company permission to release Mythos-5 to a group of 100 select companies. In a parallel story, 2 days ago, just as OpenAI was about to release their newest model, GPT-5.6, the White House struck again and asked them to slow down and only release the model to 20 select companies. Bottom line, the US government is now in the release loop for the most capable models, customer by customer, selecting who gets access to the latest models. Part of our discussion here, someone else is controlling whether you've got access to Frontier models, and maybe that's a good thing. In a recent memo, Altman said the government will be approving access customer by customer during a limited preview window, with broader release hopefully coming in a couple of weeks if all goes well. So OpenAI has announced 3 versions of GPT-5.6, all of them being throttled by the White House. Their GPT-5.6 Sol, or Sol, you know, the sun, their flagship model, 5.6 Terra, the middle tier, and 5.6 Luna, the fast, low-cost version. I like that nomenclature. It sort of is descriptive of what's coming. Allow me to open with a question that our Airborne Moonshot, mate. Saleem asks for all of us, given that the White House is delaying Anthropic and OpenAI frontier models, isn't the government effectively stifling domestic AI in a regulatory blanket? And if that happens, might it be possibly crushing OpenAI and Anthropic's valuation? Dave, let's go to you first on that one.

Yes on one and no on two, but it's inevitable. Like the models are so insanely capable that they have to be controlled. Cybersecurity is just the first excuse, but all the other evil use cases are right behind that. And so this is the new normal. I think everyone's gotta get used to it. I think that the only argument that it would stifle market caps would be tied to, is China gonna then beat these companies? But my son Sean just got back from China yesterday and we were talking about, he went from China to Vietnam to, to Korea and then back. And, uh, China is nowhere near caught up to the US. Uh, they, they do an incredibly good job of distilling, copying, and taking intellectual property, but in terms of pushing the frontier, there's, there's very little chance that China's going to threaten the market caps of these US companies anytime soon. So if the government is fair across the board, uh, which seems unlikely, but it's, you know, if the government is fair across the board, uh, then I don't think it reduces the market caps at all. These are the most valuable companies in the history of the world by far.

By the way, I know that you're busy and sometimes these episodes run long and you don't have time to listen to the whole episode, or if on occasion you miss an episode, I now put out a Moonshot summary on Substack, which includes a link to all the stories that we cover. The weekly recap covers what I and the mates had to say, what we think is most important, and what we're most excited about. And it's free. You can subscribe at diamagnus.com/metatrends. That's diamandis.com/metatrends. All right, now back to the episode. So Alex, I've thrown up the performance benchmarks on 5.6 here from OpenAI. How, how impressive is this? Give us a little rundown on these benchmarks.

Yeah, 5.6 SOL is roughly comparable to Mythos Preview. If you look at all of the cyber benchmarks and some of the bio benchmarks, it's roughly comparable. It wins some, loses some. But I think the bigger story here is we're in the regulatory endgame. This is exactly during the founding early days of OpenAI, DeepMind, and Anthropic. There was an enormous amount of hand-wringing over what would happen when we reached the era of recursive self-improvement and the race condition that would happen. And there was a general consensus that the frontier labs would all establish some sort of coordination mechanism with each other to slow down and sort of cross the finish line together. Without creating a race condition. And many people at the time said, oh, this was impossible. There's no way. There's no possible way. This is sort of the paradox of Molochian economics, I think Scott Alexander would say, would mean we're permanently stuck in this social trap of everyone competing against each other. It turns out that the same old coordination mechanism that we've had for thousands of years, government, localised geographic monopoly on force, is more than capable, it appears, of being that coordination mechanism for getting the final two, at least as of this point in time, the duopoly of Frontier Labs, OpenAI and Anthropic, to synchronise the release of their capabilities out to the first few dozen customers or users that the government is going to gatekeep. So in terms of the raw performance, roughly comparable. And I think what's even more interesting than the fact that 5.6 SOL is head and shoulders in terms of cyber capabilities and other capabilities, also efficiency versus 5.5, is that basically the government, the US government has functioned as a synchronization mechanism for helping 5.6 SOL reach essentially some form of parity or near parity with Mythos/Mythos Preview 5. That's extraordinary. We've never seen any sort of third party basically forcing the, the two leading contenders who would otherwise be in a race against each other to essentially similar capabilities. And yet that's what we're seeing. And at the same time, looking at China, the Chinese open weight capabilities, there's this chart floating around X extrapolating the time difference between Chinese open weight models reaching the same capability as the two remaining frontier Western open-weight models. And if you extrapolate it, depending on how you calculate that time delta, the number is on a trajectory to go to zero by Christmas of this year. Yeah.

So that's the, that's the crazy thing. And Yema, I wanna go to you, given your experience in open-weight models. I mean, if in fact the open-weight models from China are converging, and at the same time that the US government is sort of throttling I can imagine a lot of companies around the world saying, I want this on-prem. I'm just going to adapt the Chinese models. And that, again, to the second part of Salim's question, if that's happening, that could have a real negative consequence on the US models. What do you think about it, Iman?

So yes, I think that's an excellent question. What we've seen is kind of an acceleration of what we've been talking about over the last, I think, year since that IMO Gold OpenAI model was announced. Certain models will be available to everyone, certain models won't, and now they're becoming more and more gated, and I think this will only accelerate. Looking at the open models, GLM 5.2 was the first model with the big model feel, even though it just trained more from the GLM 5.1 base, and now lots of people are trying it. But how good is it? I think you've seen a few things. Um, first of all, you've seen multi-model harnesses. So my old colleagues at Sakana released Fugu, and then the BlitzE team are now top of SWBenchPro, bringing together lots of different models. You know, again, a big achievement from that team. We have a harness releasing on Monday where basically it's on Frontier-SWE, which is the most difficult coding benchmark. Like, each task takes 11 hours and it requires novelty. When you say we, do you mean Intelligent Internet? Intelligent Internet, yeah. So we've been looking at how far can you get single models. And single models, uh, we're going to be announcing GPT-5.5, uh, overtaking Mythos on Frontier-SWE. So that's the previous gen. But GLM-5.2 outperforms GPT-5.5 in our current tests. So we actually see on Frontier-SWE, which is the most difficult Like SWBenchPro does like really great stuff, you know, and that's the type of stuff Blitzy and others are doing. FrontierSW is asking you to build novel kernels and things. And GLM 5.2 is at the top with that harness. With a normal harness, it's like number 4 or 5. But now you've seen actually with the right harness, OpenModels already can be at the top. And that's crazy when you consider GLM 5.2 is maybe $25 million worth of code. Of compute, sorry. So I think that the gap, yeah, definitely by December, but maybe even now as base model performance becomes less important than how you use it, 'cause we've learned how to use it. It's like The Legend of Zelda: Tears of the Kingdom, you know, coming out on the Wii before the Switch. They really kind of push that.

I don't know if that analogy captures the magnitude of what you just said.

Well, you know, you've got the same latent space and you can kind of push it dramatically. Got it. And but I mean, but this again opens up a difficulty because you're seeing the gating of these models and the competent intelligence that can build your code bases. Everyone can have access to that. The novel intelligence that can make anyone a genius on an attack or otherwise, that feels that's going to have like licensing. It's going to maybe even be restricted to US citizens. But how good will the Chinese models get? We're not sure. It's just that right now they are—

hold up on China. Hold on, China, for just a second. Because what Imad just said is so insanely important. I want to be sure everybody gets it. The finish to that sentence would be, therefore, the government is too late. You can already take 5.5 or Opus 4.8 and put enough of a brilliant harness around it to make it better than Mythos or better than, um, 5.6, GPT-5.6. Therefore, you can take what's already out and turbocharge it above the level of what the government tried to stop this week. That, that means the cat's out of the bag, which means the government would then need to go backtrack and say, well, wait, hold on, hold on, we were too slow, we need now to lock down 4.8, maybe 4.7, let's go all the way back 6 months and, and pretend we did this 6 months ago. So that's an insanely impactful statement. The only part of what Ahmad said that I have firsthand experience with is, yes, Blitzy can beat Mythos in SuiBench 4.0 or in SuiBench Pro. So if that applies to all use cases with the right harness, then the impact of what Ahmad just said is massively important.

All right, let's take a second for everybody listening. Alex, what does a harness mean in this context?

Well, so a harness typically refers to non-weight capability improvements. So when you're building a machine learning model, there are many phases. It typically consists of a neural network of some sort. The neural network is composed of weights. Neural network has some intrinsic behavior. It goes through pre-training, mid-training, post-training. These all impact the weights directly. Now you have a model., and you want to on the fly change its behavior. So the— in, in the beginning, in 2020, when we first got GPT-2 and large language models, or few-shot learners, there was the prompt. You could change its behavior in context by feeding it different text as a prompt, and the output was versatile and it was good. Then people realized—

it's a biblical statement—

and it was good, and it was good., and, and then people realized maybe we want to factor out common elements of that prompt across many prompts, and the system prompt was born, and it was good. And then people realized that it was desirable to keep factoring out common elements, uh, and logic and text and, and other information out from all of these prompts into what ultimately became a harness. So It's now quite possible, as Emad mentioned, as many others, including myself, do, to create lots of what Andrej Karpathy might call software 1.0 harnesses that live outside the model, that orchestrate the models, that feed common system prompts and other prompts to the models, that parse the outputs, that mix different sorts of models from different vendors in order to achieve super performance. That's what a harness is. OK.

You know, it's also— it's a bellwether into this new era where, you know, a year ago, if you said, what's a parameter? What's training data? How many layers are in your neural net? Those have very crisp answers. They're just very factual answers. Now, the AI is telling me all day long that it's going to build a new harness, it needs new scaffolding, or it's going to monkey patch something. And it uses the word monkey patch like 5 times a day. Like, I wrote code for 35 years and never once used the word monkey patch. Why are you monkey patching? I don't know what you're doing. But it works, you know, it comes back with functional works. And we're in this new era now where we're kind of like sort of understanding what the AI is doing and it definitely works. So we kind of let it go.

I want to go back. I want to go back to this principal question. Right now, the government is controlling who gets access to frontier AI level. And at the same time, we have these open weight models coming out of China. And I can imagine a lot of companies saying, I don't want the government telling me what I can and cannot access. I'm going to start using on-prem open weight models. And is there a probability that the government's going to start restricting in the US the use of these open weight models? And then does that sort of— No. For sure.

Well, yeah, I think that there's a good chance that the US government bans Chinese open weight models from being used by corporations. And requires a license and KYC for any Frontier or new Frontier model, including retention of your prompts.

I mean, for everybody listening, this is the reason we spun up this weekend pod because of this issue. It's all of this is breaking so fast. I think Alex, you opened up by saying we're in the hard takeoff, or maybe it was you, Dave Erickson, but it feels that way. The speed.

I'm just saying don't take off the takeoff.

Okay, um, so, so how does this impact— how are we going to see this impacting these, uh, our frontier models here domestically? Um, and does this— is this the mechanism by which China pulls out in the lead across AI?

Well, it looks like protectionism. It, it superficially, it looks like perfection— protectionism, maybe masquerading as export control.. And there is, I think, a very real risk that the US falls behind. I, I think without this regulatory regime, we were neck and neck, maybe 6 to 8 months ahead of the Chinese models. And now there is very much the risk that AGI is achieved internally, but externally for all of the users, the, the users are American users and Western users of American frontier models are, are stuck at parity or, or worse. Behind parity with Chinese models, whereas internally within the labs, the capabilities are continuing to leap ahead. And that sort of distinction— we speak of the singularity all the time. In, in a black hole, there's the notion of an event horizon as being distinct, uh, as, as having distinct outer horizon versus inner horizon versus singularity maybe at the center. And there's very, I think, real risk that the frontier labs, the at the moment, the two American frontier labs will have internal capabilities that vastly outstrip what is available to everyone else. And that, as a number of folks including Rune from OpenAI and others have mentioned, creates the risk then— creates some sort of weird perverse incentives for the American economy. Like, for example, is it— is— are we incentivized to all go work for OpenAI and Anthropic so that we gain access to the internal capabilities?

Well, that's happening already. Yeah, that's well underway.

Everybody's leaping.

But actually, you know, the playbook, if you ban US companies from using Chinese models, that doesn't achieve anything because the rest of the world will still use the Chinese models. The playbook next is to say, look, we will not let you access the most amazing technology in the history of mankind. Europe, South America, East Asia. We will not let you access this unless you follow this new playbook. And that's what David Sachs is busy writing right now. Like, what are the new NATO—

It's got to be insane in the White House right now.

Well, this is a Pax Silica. One could imagine the Pax Silica generalizes to the Pax Intelligentsia, where you have the American intelligence superintelligence bloc and you have the Chinese superintelligence bloc, and Europe is torn between them.

Yeah, I mean, that's why— Iman, what's your— you're, you're over on the pseudo-European side of the market here.

Yeah, let's just say it's been a very busy week with lots of people asking about this at the highest levels. Um, I think, as you said, like, restricting Chinese models is one thing, and that's one political question, but we can have scenarios like this. Right now, Mythos is only allowed to be used by U.S. citizens on the allow list. This means Andrej Karpathy in Anthropic cannot use Mythos. Oh, because he's a Slovak Canadian. It actually says that in Latnik's order. Is he named? And well, no, it just says only citizens, even in terms of non-citizens. But, but let's take a look. What happens if frontier models from US labs are only allowed to be used again? Frontier models, not competent models, by US citizens and US corporations. That's a massive lead for America, right? Like they will continue and that's still a big market. That's still a huge market. Again, what happens if you have any licensing restrictions you want? One of my scenarios is just like you have a driving license, you will need to have a license where you basically say that you're patriotic to America and convince the Frontier model of that to get your license. Wow. You will have KYC. These are the scenarios that I'm envisioning because again, there's this split in capability and you don't want any adversaries to have access to that capability. Dave's points about distilling and things like that, but it goes way beyond that because the previous model is like you type a word, you get an answer. These things can now work for basically days if not months.

I need to steal, man, for a moment the U.S. decision on, on doing this. So the Associated Press reports Anthropic's Mythos model running red team exercises with the U.S. intelligence agencies under Anthropic's Project Glasswing, if you guys remember that from 3 weeks ago, has identified identified vulnerabilities in highly sensitive classified U.S. government computer systems. Senator Mark Warner said the following, quote, this tool broke into almost all of our classified systems, not in weeks but in hours. Mythos identified the holes with exploitation outside the scope of the exercise. So, uh, 12 days later, the Trump administration directed Anthropic to disable Mythos-5 and Fable-5 for foreign nationals. So this, you know, this exercise with the US DOD explains why they did this. And it sounds very real. I mean, I can imagine a future in which the most advanced models test everything sensitive. And then once they patch the holes, then it's opened up. But that's a strange universe.

Yeah, a lot of these things are just—

go ahead, Ahmad. Well, there is a flip side, which I would say, which is that the models aren't good enough yet to tell when they're being used for evil. So, you know, don't allow yourself to be used for evil.

I just want to add that. That's the meta prompt. Dave Erickson.

Well, a lot of these are just cover stories and excuses. They're true, but they're just— the real issue is the query, can you build yourself? So if you're in China and you go to Mythos and you say, hey, Mythos, can you help me build yourself? And then I have you, and then I can build a competing equivalent over here in China. And that's the query that Mythos can answer and 4.8, Opus 4.8 can't. And I know that from the 2 days that I had access to Mythos before they pulled it back. And that's the real issue. And so then the, oh, it's dangerous for, let me think, okay, here, cybersecurity, it cracked into some FBI systems. Good enough. That's all we needed was a reason. You know, this is like, you know, if you look at world history, governments always have an agenda. And then there's some trigger event and they say, because of that, you know, like, look at Russia walking into Ukraine, like, well, because of this one thing, that's how we're going to justify the action that we knew we were going to take anyway. But the real underlying driver here is not, you know, hacking into government systems. It's the self-improvement can't get out to the world, otherwise it's out of the bag forever. And that's the line that Mythos and 5.6 can cross.

Alex, do you agree with that? Can we imagine China using Mythos to build itself?

I think we're in the endgame. I think China has enough capabilities at this point to achieve its own recursive self-improvement, its own Fermi pile, if you will, without needing to further siphon trade secrets or reasoning traces from Western models, or just literally trying to exfiltrate weights out of frontier labs. I think we're in the endgame. That this policy could only possibly make sense if we are in the endgame of recursive self-improvement and every day matters. I think the steelmanning of, well, this is regulatory overreach, or maybe not overreach, but a regulatory immunoresponse, say, to Mythos or to GPT-5.6 being able to do incredible vulnerability analysis and mapping and exploitation, I think that justification only holds water, only supports itself if we move on, hopefully, to a lighter-touch regulatory regime say, in a month or two that maybe is a little bit, as Dave Erickson says, more focused on recursive self-improvement gatekeeping and less on vulnerabilities. But otherwise, I think China has reached recursive self-improvement escape velocity at this point on its own and doesn't need the West's help. Interesting.

So, but I also think that the White House doesn't necessarily see it that way. They think they still have time.

What's your advice to the White House? What's your advice to regulators out there right now?

I mean, you got to completely unleash David Sachs and go around the world, meet every world leader with a very specific, here's how the future world is going to work. And I think you have to lock super intelligent AI into a couple of boxes. I don't think one box is healthy for the world, but I think 10 is also unhealthy and say, look, here are the rules. And I think the most important thing by far is logging every prompt. And giving many, many eyeballs to every single use case. Because the idea that the AI can run rampant in a box and design things without anyone inspecting every single iteration, every chain of thought, every prompt needs to be inspected by other AIs to make sure that we know exactly what the use cases are. And if we achieve that, then you have a stable future for the rest of time. But everyone has to agree, OK, Who sees it? You know, does Sweden get to see it? Does— you know, who are our friends? Who are not our friends? And then lay out those rules like yesterday and then get everyone to sign up to it. And then you get access to Fable. That's, that's the deal. Emad, what's your—

what's your solution? Then we'll go to you, Alex.

Well, I'm very pro actually competent open intelligence. And, you know, I think we can build that in an aligned way. What I think, given the political reality, will happen is KYC, prompt retention of American civilians getting the thing via a license. Um, and again, something similar to having to have a driver's license or security clearance, etc. It will be the same type of regime. Um, and I think, you know, unfortunately David Sachs has left the government and so is Sriram. So now AI policy is being driven by Howard Lutnick. Um, and so his thing is very much the economic interest of America. Which again I think is reasonable, just like the cybersecurity thing. The reality is this is an incredibly difficult situation because we're in the exponential. And so the government has to protect US interests, has to ensure other people don't have these capabilities, and it has to do that while having the upside. They don't want Anthropic and OpenAI to blow up. You know, there's a question to David at the start. So balancing all these things is very difficult, but I think as AWG said, we're going to move to a regulatory regime where some of these things are clear, like the licensing, you know, like you're in a program, you have to jump through certain hoops to get this type of access. And right now they're just trying to figure it out incredibly quickly because frontier capabilities are popping up everywhere through that one level of the cyber attack, that one level of— you notice the defense, the gene defense thing, you know, as bio, bio attacks as well. Sure. And then there's a capability thing. You don't want all of a sudden your companies to be outcompeted by foreign companies, you know? And that's the protectionism thing.

Alex, take us home on this one.

Yeah, a couple thoughts. One, Sputnik was a moment of strategic surprise when the US was surprised by the USSR. The surprise came from outside, not from within. This time around, this is a Sputnik moment and the surprise came from within. The NSA and Cyber Command were surprised by the private sector. The private sector leapfrogged the vulnerability and other cyber capabilities that the NSA had been keeping bottled up. And that has to be a bit of a shock to the existing bureaucracy, because now suddenly there are these capabilities that are leapfrogging the government from the private sector. That's thought one. I would say more broadly though, it— the problem with recursive self-improvement is eventually the capabilities of the non-human intelligence, the AIs, meet and exceed human-level intelligence. I'll say something mildly provocative, which is if you think that we should— that the U.S. should have strong immigration policy for human immigration, human, uh, capital into the US, then arguably as the capabilities of AI start to meet and exceed human intelligence, you'd better be thinking about policies for what happens as you import foreign AI as well. At some point, the US will be importing more foreign— if this open source trend continues, Chinese capabilities ultimately meet or leapfrog US frontier capabilities, the US will be importing, importing more foreign artificial intelligence than more foreign human intelligence. And I think that has profound policy implications.

All right, let me move to our second OpenAI story. It's about GPT-5.5 Cyber, codenamed Daybreak. It's a defensive cybersecurity model that just scored a record 85.6 on Cyber Gym benchmark, which tests AI agents for real-world cybersecurity security vulnerabilities. 85.6 is the highest single model score ever posted. Sam Altman said, quote, the real prize isn't finding the holes, it's automatically writing and testing the fixes across web browsers all the way down to the Linux kernel, effectively turning the threat into a cure. So gents, AI is shifting from offense to defense at scale, potentially closing holes faster than attackers can find them. The question is going to be about trust. Who's allowed to merge AI-written fixes into the codebase running the world's infrastructure? Whoever owns that trust layer and the liability controls the digital security across everything. So Alex, your take on, on this Cyber Gym performance and GPT-5.5 Cyber? Well, first, daybreak.

This is what Peter, you and I wrote about in Solve Everything. We are seeing cyber vulnerabilities in open source and closed source projects get bulk solved now by AI. There are a few different nonprofit initiatives, one being run by IBM, there are a few others that are competing, solely devoted to just using these new capabilities to bulk identify and remediate vulnerabilities and bugs across all the open source projects. So I think this becomes a sort great project of the times, like the interstate highway system, except going back through the historic record of all of these open source repos, especially the foundational ones that are in the supply chain of really popular downstream applications. Of everything. Of everything. Just bulk solve it all now that we have the capability to bulk solve it. On the, the narrower point of GPT leapfrogging on CyberGym, I think we're going to see again internal, internally only, or with a limited staged release leapfrogging of capabilities in terms of the ability to detect and remediate vulnerabilities and bugs. We're seeing for the first time Sam talking about intentionally biasing the models toward defense versus offense. And he never used the word poison in characterizing this, and OpenAI never used poisoning attacks, which perhaps maybe a slightly less inflammatory characterization than Anthropic's original release announcement with Fable and Mythos. But reading between the lines, it certainly sounds like OpenAI is steering their models to basically self-disrupt or poison their users if their users are trying to use their frontier models for offensive cyber versus defensive cyber. They're reading between the lines. That's what they were implying. And I, I think that's the world we're going to find ourselves in where alignment comes in the form of a model together with scaffolding, system prompting, post-training that favors the defender and favors the, the desired outcomes, and then obfuscates, foils, poisons attacker-type use cases.

Iman, do you agree?

Yeah, I think this kind of comes to the stage, it's going to be very interesting. Um, like Elon did the deal with Anthropic and Anthropic's weights got loaded onto Colossus. What's the security level of these data centers? Those weights are gonna become even more valuable than they ever have for espionage and others, because you don't need necessarily a million agents for cyberattack. You need to have a few models, and in particular, what you want is the model that comes before the end of pre-training or before the system prompt that AWG was just talking about. Because they're more creative and they're not as hobbled for the attack. And then you have to defend against that because the adversaries are not like kids in the basement. The adversaries are nation states utilizing these models. And the model, if someone gets, is quite limited, but the other side managed to obtain the base weights, that will be more creative and have a bigger attack. At the same time, we all know our infrastructure is terrible. We don't have the human compute hours to make it stronger. Like basic attacks before AI started crippling our thing. So I think that, you know, there needs to be a massive push for all essential apparatus, government and otherwise, to be hardened by these models. And then that makes it incredibly difficult to attack. The final thing that I think we'll see is that you will see probably claims that Chinese models will be introducing backdoors and other things inherently in the code. Because now this is what you're saying, Peter, Who controls who merges the code, those thousands of lines of code now that you have?

Yeah. Who do you trust?

If you have a model that could poison or introduce a backdoor inside its latent space, the model weight gets updated. You will not find that for days, months, years. And all of a sudden it's backdoored everything. And this is the real risk profile because who on this call now actually reviews every line of code that they merge?

Yeah. Yeah. That's a key point.

Especially downstream, like it can emerge anywhere. So the US government again may actually say you need to have ISO-type certified models merging your code because otherwise you could have holes in if these models come from wherever, you don't know what's in them and they're not acting on your side like AWG said.

Dave Erickson, you wanna close us out?

Yeah, well I think what Emad said is brilliant of course, but that nobody can check code anymore. And when you, when you ask your AI to install something, it does it so quickly and efficiently that you can't possibly keep up with reviewing what it installed. So then it's just a question of, did I trust that AI or not? So it's trivially easy for a Chinese model to inject spyware. It's just a few lines— one line of code actually is all it needs. Um, so the idea that we'll ban Chinese models from U.S. corporations, very likely. But that again doesn't solve the global version of the same exact problem. So all these things are imminent, absolutely, totally imminent. I think the story within this story too is that only AI can keep up with AI. You know, hey, we have a big cybersecurity risk. Sam says, well, the way to fight that is with AI that's good. So it's going to come down to what it does next.

Buy our product. We're here to defend you.

Yeah, this idea of ISO-certified trustworthy AI stamp of approval imminent and critically important. Then the question is, okay, but what entity is the trustworthy entity giving that stamp of approval? Is it a US entity? Is it a US plus Europe entity? Is it some new NATO-type entity? And I think it'd be very healthy for America to reach out to a much bigger chunk of the world to decide, you know, these issues. It'd be a lot better for global confidence in what we're doing. But at the end of the day, something has to say, yeah, you can trust this AI and you have no option to live without an AI.

There's just not a choice there. There's one thing I have to say. I have had a couple of conversations this week where the question has been asked to me, how can we trust American AI is not installing backdoors?

Well, and do we realistically believe today that the US government can't read, can't listen to our podcast right now? Read every single text we're sending to each other, read every single email. I mean, with, with, with Palantir out there helping, is it, is it even vaguely viable that the US government can't read and see everything going on for every single human being?

I assume they are. I assume they are. I mean, this is the concept that privacy— you'd like to have privacy. We think we have privacy. We want privacy, but I don't assume privacy. All right. I'm going to move us to our third and final OpenAI story. It's reported that the leadership is pulling back on their near-term IPO. So the company advisers have said you've got two paths. Path one, you go public now this year, but potentially accept a valuation sub-$1 trillion. Path two, wait until 2027, continue scaling revenue, infrastructure partnerships, and preserve that $1 trillion IPO narrative. Altman apparently is not interested. In going public below $1 trillion. I think the competition between himself and Elon is still there. They watched the volatility of SpaceX's stock price sliding from a high of $202 per share down to yesterday's close of $153, and it caused them concern. Worth noting that, you know, SpaceX's IPO price is still above the $135 per share, at least for now, maintaining their $2 trillion valuation. So, Dave Erickson, here's my calculus, and I'd love to know what you think. I think, you know, by staying private longer, OpenAI and the other frontier labs can avoid sort of the quarterly market pressures they're going to be hit with as they are burning a staggering amount of capital for compute. Right. And I don't think the markets have the patience for the capital spend and the timelines, you know, for achieving AGI and ASI. I mean, the timelines for that are years and the public markets are looking quarter to quarter. Dave, what do you think about that?

Yeah, I think, I think this headline is absolute bullshit and that you nailed it, Peter, on why it's absolute bullshit. Oh, SpaceX, look how volatile their stock is that I'm going to delay our IPO for a year because of that. Like, what are you talking about? It went at $135, it opened at $150, and now it's trading at $153. That's pretty much a perfectly priced IPO. Like, ask anybody on the street, what do you think should have happened? Oh, shouldn't it be like tripled by now? If it had tripled by now, then Elon would have left over $100 billion of cash on the table. He didn't. Like, this is just a perfectly good IPO.

A trillion of cash on the table.

Yeah. Yeah. So you're just looking for an excuse to delay your IPO. This is a really good one to use because you're pointing at your competitor and saying, oh, look at this problem they had, therefore— but it's not true. They want to delay their IPO, one, because they already raised $120 billion recently. Recently. 122. Yep. 122. So they don't need the money. And then two, who's going to run the company?

I mean, it's going to be the latest GPT model is going to run the company, of course. Okay.

Well, then they have to make that transition into an S-1 filing that the SEC approves. You know, and that's, you know, Alex has been saying for a while, actually, there's a ton of change coming to the way things get financed and go public. And that's coming soon. And there's a very real chance that Sam wants to wait that out and see what new economy emerges over the next year before— because again, he doesn't need the cash, but Sam has hundreds and hundreds, at least 400 outside investments. And I know as a public company officer, when you fill out those SEC forms, you have to disclose every single holding and every potential conflict. For Sam, that must be like an encyclopedia-sized book. Now, he's never been a public company CEO before. He's probably like, holy crap, this is insane. Elon's done it before. So Elon got out very quickly. He's been running Tesla. He paid his $22 million fine for one tweet. He knows the game. But Sam and Dario— and Dario too, he's probably like, oh my God, this is onerous. Do I really want to cross this line? So I think that's— meanwhile, your front door just got shot.

Friends don't let friends run public companies. You know, the biggest problem here is that the largest wealth creation event in human history is out of reach of the retail investor until these companies go public. And it's being held by a small number of VCs, family offices, and sovereign funds. I mean, that's the concern. But Alex, do you want to take this up?

Yeah, I agree with Dave Erickson's points. I would also say OpenAI screwed up. They focused too long, too early on the consumer, assuming that the consumer would be the source of the revenue engine that would power their path to an IPO. And that bet was probably incorrect. They should have focused on enterprise. They're now trying to become Anthropic faster than Anthropic can become OpenAI. That seems to be working. Codex is a wonderful product, and Codex revenue according to OpenAI's reports, is skyrocketing. So if I were Sam and I were OpenAI, or maybe I were just Sarah Friar, I would be asking the question, how long until Codex can be fully brain-swapped in with all of ChatGPT and powering the revenue engine that I need to motivate the trillion-dollar-plus IPO? With Elon and SpaceX, he pulled a few rabbits out of his hat at the last second. Incredible. Announcing Anthropic hosting deals and then hyperscaler hosting deals with a number of other firms. Yeah, um, with OpenAI, ideally they'd have a few revenue rabbits that they can pull out of their hat in order to supercharge an IPO, and my impression is they're not there yet. It could take the form, to Dave Erickson's point, of an AI replacing Sam Altman as CEO, and maybe the capabilities aren't there yet. My bet is though it's mostly a revenue story and an accounting story to make sure that they're profitable and not burning cash. And I would assume, not investment advice, that they will get there sometime in the next year, but it's going to take some time.

You know, I think there's a deeper level to this story though. If you, if you said, here are 3 people, here's Dario, here's Sam, here's Elon, and one decides to get public very, very quickly, raise the $85 billion. The other 2 are like, wait a minute, there's a lot more to this. But then you look a layer deeper, the other 2 are in Silicon Valley. And— or in San Francisco. And in San Francisco, the consensus is that the hard takeoff is right now and that we're going to discover new physics, we're going to discover new medicine. The whole way the world is governed is going to get changed.

Yeah, good point. A lot of wealth to be created on the back of just the scientific breakthroughs coming. Yeah. Yeah. So they—

I think Elon, he was— he lost his edge on the frontier model. And so his way to become relevant in all of this is to get into space, get the orbital data centers up and running. Buy a huge amount of compute. But if he were on the frontier and like 2 steps ahead of Fable right now, he also might be saying, holy crap, everything's going to get changed in the next year anyway. Um, but he needed the money and he needs to— he needs to build that big infrastructure.

In the biotech IPOs of the 2021, '23, all these companies that were pre-revenue, pre-profit started going public and they got decimated. And I think, you know, one of the rules I've always had is you go public when you've got profits and predictable revenues. And that is not these companies right now. Dave, do you think— do you agree with that?

No, I think that they can manufacture insane amounts of wealth and revenue very, very quickly to the extent that they have access to compute. I don't think the revenue visibility and the CFO are a big part of the decision. I think they genuinely believe the world a year from today doesn't look anything like the world today, and that wasting a ton of time dealing with the SEC and the roadshow is the stupidest thing you can do in the middle of the singularity. And so they're just going to— to the extent that they have access to capital and they don't need the money tomorrow, it's much smarter try and stay out of the lights, try and stay out of Washington, try and stay out of the SEC, try and stay out of like, so you can focus on the model and focus on the new world order and focus on the, you know, there's just so much more pressing, urgent, hard takeoff issues in front of them. I mean, I just can tell you from firsthand experience, as soon as you start filling out those SEC documents, you're like, what a freaking waste of time. Holy crap. This is like from 1929. What am I doing here? And Silicon Valley, San Francisco has that arrogance that like, we are the world right now. This is everything happening that matters. And dealing with Washington and the SEC, it just feels so wrong. I think that's more the flavor.

Imaad, do you think we see Anthropic do the same thing or are they going to jump in and try and grab the capital out there?

I think Anthropic with their continued revenue ramp, it all depends on are you going to see a drop and you haven't seen yet? If not, then again, why would they do it? Because they're ideological, right? You don't want to have anyone else's fiduciary or otherwise control, and you've got the weird PBC structure. OpenAI, I think, is a bit different, and they're changing over their structure. But I think the thing to watch out for is, do they buy Sierra and put Bret Taylor as CEO and move Sam to president? That's an example of how you can get around that. That's a good thing. If they need a rocket company, they could buy Rocket Lab, right? Like, there's all sorts of moves you can have here. But as you said, the rate of the revenue is insane. Like, I was actually looking at SpaceX's AI revenue from just their cloud business. It's overtaken AWS and GCP on a run rate. Mm-hmm. Within a few months. Who would've even thought that, right? But then you look at OpenAI, they're up at $40, $50 billion now. The revenue actually has gone to fit their valuation, which is the crazy thing, cuz you've never seen revenue growth this big. I think it is just their internal structure is still shifting. That's a big deal. Like, what does it look like over the next few years? But they have the space to do that because they're going to lose $26 billion this year according to their forecast. And as Dave said, they raised $120 billion. So it's not like they're going bankrupt. Anthropic is a little bit closer in terms of their raise to their balance. So they're either going to have to do a raise or an IPO. But again, can you imagine Dario playing the markets like Elon? No.

And by the way, people will throw money at Dario. You know, just ask it and the money will flow in. Yeah.

Well, also the employee shares, the vested options, people will buy those from you too. So you don't need the, you know, normally people are racing to the IPO so they can get some personal liquidity, maybe buy a house or a car or something. Here they have tons of secondary liquidity. So what is the purpose of the IPO then? Like you're right, altruistically giving everyone in the world access to your stock would be a really nice thing. Um, but putting that aside, like they don't need anything. And, and then you look at the regulatory overhead and, you know, it's, it's not just the IPO itself. Like after the IPO, if your stock goes down, you're gonna get a stock lawsuit. You have to deal with that. If you tweet or, or say anything, you know, virtually any word you say has to go through a, a FDA approval. You can't post anything on the web without it going through an FDA approval. It's just a, it's just a different world from being a private company. And, I just think they're suddenly realizing, wow, if we don't need it, it's not easy.

Welcome to the health section of Moonshots, brought to you by Fountain Life. You know, my mission is to help you use the latest technologies, including AI, to not just do your work at home, teach your kids, but to help you live a long and healthy life. I'm here today with an extraordinary physician, the chief medical officer of Fountain Life, Dr. Dawn Musalem. Dawn, let's talk about cancer. Uh, you know, I know, uh, from the member database that we have at Fountain, our members who come in who think they're healthy, it turns out 3.3% of them have a cancer in their body they don't know about.

That's right. You know, the majority of cancers that we screen for, those aren't the ones that are necessarily taking the lives when found at a late stage. We know that when cancer is found early, the chances for cure are much higher. We know it's much easier to treat a cancer when found early versus when found late. What we're finding in our members is over 3.3% were found to have these cancers that were otherwise wouldn't have been found or detected.

Yeah, you know, it's interesting. People, you don't feel the cancer until stage 3 or stage 4. And if you don't know what's going on inside your body, it's like driving your car with your eyes closed. And you can know. And so when members come through Found, how do they detect cancers?

So we're doing full-body MRI, and we also do early cancer detection screening. This is very, very important, and these are not typical tools used in the conventional care setting when it comes to prevention. This is a hard thing because currently these are not studies that insurance would yet be covering, but the goal is to collect these numbers, do the research, and work hard to democratize wellness.

Yeah, so at the end of the day, you can know what's going on inside your body. It's your obligation to know. So check out Fountain Life. You can go to fountainlife.com/peter. To get access to the latest technology to help you detect cancer at the very beginning, at stage 1, when it is curable, before it gets to stage 3 or stage 4 and you're a world of hurt. I'm gonna move us on. Our next group of stories is in the Musk universe. So this week, Elon announced that Neuralink may make an attempt later this year for the first direct human-to-human telepathic communications, literally transmitting from my brain to your brain if we're connected via Neuralink. Not typing, not speaking, thought to thought. But if it works even partially, Alex, we've talked about this, isn't the most science fiction milestone ever attempted in neurotech. I think of this as the new level of intimacy, right? If you know my innermost thoughts, it's the beginning of a new communications capability for the human species. You know, Kurzweil famously predicted high bandwidth neocortex to cloud communications by the early 2030s. Alex, let me go to you next. Let me just make one more point. When I was speaking to Elon about this in the past, you know, it's clear his endgame in Neuralink isn't a medical device. It's his desire to create an I/O layer for the singularity, allowing humans to, quote, couple with AI during the singularity. Your thoughts, my friend?

Well, a couple of thoughts. One, if Neuralink does do this, this will be one of the, the first sort of, uh, open attempts to create create superhuman capabilities, not just restore capabilities from humans with a variety of, say, motor disabilities to the mean, but rather to empower people with superhuman capabilities. That's the first thought. Second thought is the latent space is particularly interesting. So there was a paper in Cell, I think sometime in the past week, I wrote about it in my newsletter, finding that people children, adults who speak multiple languages, so bilingual adults, there was, there was an open question in neuroscience whether if you speak the same, the word for the same term in two different languages, whether there was a single bridge neuron that was responsible for activating for those two different concepts. And the answer turns out to be pretty surprisingly— I'm actually kind of surprised this wasn't published in Science or Nature— no, but The answer turned out to be that the spacing, the relative spacing—

The geometry. I read the paper. It's incredible. It's just like the AI models.

It's just like the AI models. Yes. That the hippocampus of the humans, bilingual humans, turns out to look more or less like a vector embedding space for an encoder-only transformer model. That is a shocking conclusion.

It was shocking. The neuron for perro and dog are adjacent to each other Um, and that's what gives them that colocation. Amazing.

So if our hippocampus is basically just an encoder-only transformer embedding space, that would suggest that A, telepathy is going to be a lot easier than one might otherwise suspect. B, maybe human cognitive capabilities are actually not that complicated, uh, and maybe we are just, as I've mentioned on the pod previously, just distorted reflections of our ancestral environment and the complexities in that environment. See, Neuralink is pretty invasive. I mean, I, I know it's packaged up as being less invasive than some of the alternatives, but still pretty invasive. You, you need to stick a needle, multiple needles, inside a human skull. But if, if this embedding theory of cell placement in the human brain turns out to generalize, I would expect this will create enormous demand. And Imad, you of all people probably would I have some view on this. I've chatted over the years with some of your, your former neuro employees. This will create, I think, enormous demand for non-invasive human-to-human telepathy, not just the invasive Neuralink type. And I wouldn't be surprised if this ends up becoming a hugely popular feature or product. Elon may be steering Neuralink in the direction in parallel of less invasive, non-invasive BCIs.

Can you imagine if couples, husbands and wives basically get a neural link connection? I think probably the divorce rate would skyrocket.

Or we see a lot of borganisms.

Yeah, borganisms. So one, let me throw this to you next, Emad, but just a point to make. The human input/output rate is shockingly narrow, right? So I looked up the numbers here. So speech is about 40 to 60 bits per second. Typing is like 5 to 20 bits per second. And accordingly, it says conscious thought and behavior selection is 10 bits per second, right? I mean, these numbers— and we're, of course, when we're inputting and outputting, we're bottlenecked by our fingers, our voice, our eyes, our attention. Um, we're pretty damn slow compared to our digital brethren.

Yeah, I think that's, you know, when you're processing things, you know, it's the type 1, type 2 thinking that Kahneman said. When we're thinking through, we think at a certain rate, and in tokens, you know, like your Claude tokens or whatever, it's 100 tokens a second roughly, right? But when you spot a tiger in a bush, it's instant because it adapts to the latent space as you represent it. Like, um, in 2023, the neuro team at Stability did a paper called MindEye where you looked at a bottle, did an fMRI, and then we could reconstruct that from the fMRI using Stable Diffusion. Which indicated that human latent spaces were the same. So what does human-to-human communication actually look like when you've got telepathy? You don't need all those words. You have as few words as possible to activate the common latent space, just like, you know, you're like, oh, you're in sync with each other. You barely need words to complete each other's sentences. The fact that our latent spaces are actually likely to be very similar means that the bandwidth's probably going to go up 10, 100, 1,000 times from there. Because you can adapt the adapters to hit the person at the right time. Just like when you're watching a movie, that scene will make you sad, you know, like when you understand the latent space, this is where you can go into that exploration. I think if we can get telepathy, it is actually one of the biggest achievements in humanity's history because we've had this very lossy interface for so long. But then you can dig deep into what really makes us human. Of course, there's bad stuff to that and good stuff. This is very, very sci-fi. And, you know, one of the potentials is we all end up as the Borg, you know, like KWG said. But really, the meaning of life is to understand yourself better. And I think this will be a really huge advancement in that.

Or maybe, maybe glass half full, Imad, maybe we see like the next generation version of Microsoft Teams ends up including organism feature and that the human teams are all in sync with each other.

Uh, he said glass half full, right? Okay.

Oh my God. All right. I'm not going there. All right. Let me continue on in the, uh, in the Musk universe here. Um, so, uh, we saw this week, uh, some more insight into how Elon is naming his companies. Uh, so he expanded, expanded his naming protocol. So his space-based ventures all have star in their name. And he released the name of his AI satellite constellation called StarMind. He also has his cargo delivery program called Starfall. So Starlink moves bits, Starfall moves atoms, StarMind moves intelligence. Elon went on to say, and I love this, we're cutting back on the use of the word star as a prefix. It's getting a bit silly. Too much star shit.

Some of these, Peter, if I might just comment on this, some of these names we haven't even, I think, materially— if you go back one slide— we haven't even materially talked about much on, on this pod. So to talk about—

you want to hit them? I hit them real quick.

Yeah, yeah. So Starlink, everyone knows that's space communication, probably going to launch, if you believe the rumors, going to launch a direct-to-cell mobile service.

By the way, rumors, uh, rumors that You know, he's got a deal with Charter coming up for— actually, it was reported in Bloomberg that he's going to have a deal with Charter, and rumors that he might buy T-Mobile. That'd be pretty cool.

Yeah. So Starship, everyone knows, um, propulsive landing and enormous heavy lift capabilities for everyone else in the economy. Starbase, Texas. Starfactory, factory at Starbase producing Starships. Starshield program for the US government to supply basically a private— or I want to say privatized, but a government defense-grade version of Starlink for the US Department of War. Starfall, just announced in the past few days. So this is a cargo deployment solution where private companies— and this was very, I would say, probably intentionally poorly marketed have the ability to launch cargo up to LEO and then do a retrieval of the cargo.

So people don't know a lot. One of the biggest issues in the past has been downmass from orbit. You know, we all talk about getting stuff into orbit, but being able to do experiments, especially materials or biology, and get the product back down. And just a quick shout out to Jason Dunn. He's a Singularity alum, a friend of mine, and he's got a company called Outpost, and they've been working on this with their product called Carryall, and they've been doing extraordinary hardware development and testing for now a couple years. But please continue.

Yeah, and I mean, there are companies, Outpost among them, Varda among them, that are doing quite a bit of orbital manufacturing. They're most prominently, I guess in my mind, Starfall example. There was an orbital brewing company that launched yeast into orbit on Starfall, retrieved it, and they're going to be now brewing orbital beer. Or Leo brewed beer. What is that called? But it's been to space. Stargaze. Uh, Stargaze takes advantage of the fact that Elon and SpaceX now have all of these satellites that can look up and can look down and provides situational awareness looking up and looking down, both in orbit and on the ground. Could end up— given that we just had the, the conversation with Planet that could end up, possibly end up becoming a Planet competitor. We'll see. StarMind is the brand name for the AI orbital constellation, the Dyson Swarm brand from SpaceX, if you will. And then StarPipe, which was only announced 2 days ago. This is incredible. StarPipe is, I think, Elon's, or really SpaceX's nascent oil and gas play. So StarPipe is the beginnings of SpaceX's I, I would say refinery attempts to pipe natural gas around, uh, given that, uh, the SpaceX's— the, the propulsion is largely focused on, uh, methox. Yeah, uh, so piping methane around, piping natural gas around, they have no experience really in, in any refinery capabilities. But Starbase is in Texas. Texas has a lot of oil wells. StarPipe is this pipe that was just announced 2 days ago to start piping these capabilities around. And really, if you're going to build the first Martian colony or the first serious lunar colony, you need to have oil and gas capabilities and you need refining capabilities. So StarPipe appears to be the very beginning of SpaceX getting into oil and gas.

And that's probably where Elon said stop calling stuff starshit.

It's a freaking pipe. And but also powering their data center. So here's that, here's that tweet from Elon, too much starshit. But I love this. A bunch of memes came out. This one is from Planet of Memes. And if you're watching it, it says, when you want to smell like a trillionaire, use Star Musk, cloned by Elon Musk. That was too good. That was too good to miss.

Um, uh, hey, if anyone didn't see our last podcast with Will Marshall, the, the part on LEO, or LEO, low Earth orbit, and the value of it is unbelievably good. Definitely go back and watch that podcast. Like, that was some incredibly great media.

Yeah, that was— all right, uh, another story out of China. Uh, this one is the release in beta and soon coming in July of Seedance 2.5 by ByteDance. So get this, 30-second videos at 4K resolution. You can reference up to 50 different inputs— images, video, audio, uh, director cinematographic cinematographic control options for post-production and supports editing via text prompts. I'm going to run this video. There were so many of them, but I just want to keep in mind here, Elon said by the end of this year, full-length motion pictures coming out of AI. Hollywood is cooked. My prompt. Hello? Is it fixed yet? We're about to play your video at the event.

There's been a small unexpected issue.

I'll fix it right away, just give me a few more minutes. It appears that some higher dimensional entity is now repairing all of the anomalous phenomenon that occurred today. Wow. All right, Iman, this was your business for many years as the CEO of Stability. Talk to us about this. Where are we? Where is it going? What does it mean for all of us and for, you know, the folks who are a few miles from here in Hollywood? I suppose I told you so, right?

Like, this is exactly— when I, when I was leaving Stability a few years ago, we had this discussion. I was like, it's gonna be like 2026, we're gonna get Hollywood-level full control input by 2027, full-length movies. And we are here. It's still remarkable to see. Again, SeeDance 2.0 was a big advance. Grok, Imagine caught up to it. But the fact that you have that level of control, and I think one of the main things there apart from the quality was 50 inputs. This means 50 different characters and you can input video, audio, images, and others, and it generates all of these on the fly. Means you have almost perfect pixel control. And just like, you know, people have been using GPT Image 2 now, you know, it makes really good images. That's now here for video as of a few weeks from now. That level of control where it actually understands what you're saying effectively. So I think for Hollywood studios it's good because costs go down, but we have to think about the real impact of that downstream. Because why do you need people on the ground when you can take what we're seeing here and edit it any way you can imagine? Post-production, where does that go? And this is one of the first big waves of, you know, aside from call center workers, real human impact, because all of those human hours that went to media are gonna get displaced and we have to figure out what do we do because they're not like they can retrain. At the same time, we have the other side, which is, you know, the upcoming Ron Verri X Prize and things like that. Yeah. Being able to tell stories that you could never tell before. And I always like to think now, one of the things I say to old people is stop thinking of these media models as single-player experiences. Like we went from movies to I'm prompting by myself. When you use these models as groups to tell stories, it's actually one of the most rewarding, empowering things you can do, because different people have different views on how it all shifts. And I think that's what we're going to see a lot of in the next few months, again with the Roddenberry X Prize and others. And it's super exciting because if we're nothing if not story-based creatures, but a lot of the stories that are important don't get told.

Let me just— a quick correction, it's called the Future Vision X Prize. Yeah, that we— Roddenberry family The creators of Star Trek are involved and donors on this. This is a partnership we do with Google. And just a quick shout out, if you're a filmmaker and you want to have— if you can create a movie using this technology that is a hopeful, compelling vision of the future, right? You know, humanity aims for the targets that we create. Instead of dystopian futures, let's create positive, hopeful Star Trek futures. If you come up with a great 3-minute trailer, We will make your movie. That's the goal. We're gonna be working with Range Media and Google, backed by ARK Invest, backed by Marc Benioff at Salesforce, and going to create at least one, hopefully more, of these hopeful visions of the future, a full-length motion picture, and we're gonna steer humanity towards that positive Star Trek future.

Alex, love your take on Seedance 2.5. China's running away with video generation, unfortunately. So if you look at what the American competition looks like, what do we have? We have Gemini Omni, which is still limited to about 10 seconds. We have Grok. At least Elon is trying to give the Chinese a bit of a run for their money. But in China, and if you're training a video model, you have probably cheaper access to data that, by the way, as last time we discussed, I think at the, the Seed Dance 2.0 launch. If you're in China and you're one of these Chinese frontier labs, you're probably not too worried about being sued for copyright infringement for all of the TikTok or other similar video, video data you're using for pre-training these models. So you basically have far cheaper data, far less encumbered, at least in practice, training data. And also the American frontier labs are all busy chasing each other's tails to build recursively self-improving code gen models that are ridiculously revenue generating. If you ask how revenue generating are these video models, probably no comparison per token or per flop. I, I would guess CodeGen vastly more lucrative and more economically productive than video. Who's going to be generating long videos? It'll probably be consumers who don't have that much money anyway. So for a variety of reasons— economic, practical, legal— we find ourselves in a world where China is running away with the video race for the moment until the West can come up with a compelling enterprise, a lucrative, productive enterprise use case for video generation, at which point I would expect and hope the Western labs to, to finally re-enter the race seriously.

Yeah, this is where Liquid AI might burst onto the scene very soon. They're going to do a capital raise in probably a couple months, um, so they'll kind of, you know, get back on the radar. But Liquid AI— remind us who Liquid AI is. It's a foundation model company from scratch, doesn't use any Anthropic, doesn't use any OpenAI. It has a far, far more efficient way to use context than the transformer attention window. The byproduct of that is if you buy a Mercedes in September and you talk to your car, you're talking to Liquid AI. On-prem. So on-prem, it has to work, has to be small enough to fit in the power supply of the car, and it has to work without connecting to the internet because nobody wants their car to just stop if the internet You know, it's— you're in a dead spot. So they've got this kind of edge world really well nailed. But I saw a video generation model from them over a year ago where as quickly as you can speak, it's generating the images or the videos as quickly as you can talk. And the problem with the, you know, the really, really good models generating super high quality video is we were just talking about Neuralink and, you know, hey, I have so much bandwidth. I can, I can think so much quicker than this. But then you wait like 3 minutes for the video to come back and it just breaks the whole creative cycle. So that's fine when you're creating a movie, but the video game industry is already bigger than all other media, media combined, including movies. It's a far, far bigger industry creating video games. And whoever wins the race to getting real-time, the quality of what you just saw, into an interactive real-time experience embedded in a video game environment, that's where the big money is. So as Alex pointed out, like right now all the frontier labs are chasing coding and white-collar automation because that's where the money is. But there's also a ton of money if you crack into real-time video game. And it's just the latency right now is the buzzkiller. But whoever solves that is going to be— is going to be— I agree with you.

Everybody has a good shot at it. Video game industry is much larger than Hollywood. Much.

It's a couple, it's a couple of hundred billion versus Hollywood at 50 billion. Video gen AI, I'd estimate at 4 or 5 billion in revenue at the moment versus 10 times that for code gen. Um, actually 20 times that for code gen. Um, but I think what you've just described, David, is next year. Real-time 4K video games. Cool. And the advantages the Chinese have on this is the world model side. So one route to AGI is recursive self-learning on code gen. This, like, if you look at that SeaDance 2.5, do you really think it doesn't understand physics? You know, like, it's clearly got a physics embedding. And from ByteDance and kind of others, you're starting to see the first world models for that real-time interactivity. And it's going to be very interesting because it seems like two different routes potentially to AGI. And which one of those will win? Who knows, right?

Like, I think after that event next year, like you're saying, video game today is a couple hundred billion, which is an enormous market. It. Um, I wouldn't be surprised to see it 3, 4, 5, 10x after. I mean, it's just so compelling, and it'll also penetrate pretty much every age bracket. Right now it tends to be dominated by males under the age of, say, 30, but I think it'll expand out to all, all populations.

The Holodeck is a trillion-dollar market.

Yeah, I think, I think there's, I, I think there's like a $30 trillion market, which is enterprise software. So I totally agree that video games are a larger market than Hollywood feature films, but there's a market that's orders of magnitude larger than video games, and that's enterprise use cases. I don't know if you guys saw it just again in the past few days, Alibaba's Wan Streamer demo. This is real-time, interactive, generative discussion, video to video. Like the three of you, or I guess more probably I could be an AI right now the, the three of you are real, uh, and we're just having this discussion, and you could create a generative environment for me, and I could put my hand in the video or take it out. So, so there's— to Ahmad's point, there's some sort of minimal world modeling going on there that exists now, uh, and it's real-time and it's interactive and it's video to video. So if, if the West can stand up an enterprise real-time, say, like like Zoom participant or FaceTime participant that can participate in company meetings, can be interacting in real time with audio and video, not just the live audio models, I think that starts to move us to a $30 trillion market versus just a $1 trillion market.

You know, next year I'm gonna be bringing the top quantum computing companies on stage, the top humanoid robot companies on the Abundance 360 stage. What do you guys think? Should I bring some of the top video game companies as well?

Is it time to bring them in? Yeah. Are you kidding?

Okay. All right. We'll, we'll do that.

Yeah. Cause I think, I think they're going to, you know, Alex is right, but I think they're going to be a major player in enterprise if they pivot in that direction too, because they have the, the technology and I don't think there's any barrier there. Like if you built Fortnite and then you add real-time video generation to it, why don't you just pivot over to enterprise?

Like Slack. Slack started as a video game. Yeah, there you go.

Yeah, exactly. Amazing. All right, moving us along, staying in China. Our next story is from Anthropic, who accuses China's Alibaba of running a massive distillation campaign against Claude. We talked about distillation on the last pod. Alex was explaining it in detail. So apparently using 28.8 million fraudulent exchanges across 25,000 fake accounts to securely extract and copy Claude's capabilities. If the claim holds up, it's the single largest AI model theft left accusation, accusation ever made. Um, so, uh, you know, distillation is when we use one model to answer questions to train your own model. It's the teacher and student approach. Uh, this is now frontline in the US-China AI rivalry, and it raises a brutal question: in a world where intelligence can be copied through straw, uh, through, you know, how do we collect and protect models for ourselves. So Alex, I'm gonna go to you. What's the implications here? Do you think it's 100% true?

Well, first I have to point out the irony that Anthropic itself has been the target of multiple suits arguing that it took copyrighted material and maybe not using the term distillation, which is usually reserved for model-to-model training versus corpus-to-model training or pre-training. But I would say the shoe is ironically on the other foot. Anthropic sued multiple times for using improperly copyrighted books and other media for pre-training its own models. Now Anthropic is turning around and accusing Alibaba for using Anthropic outputs to post-train Alibaba's own models. Similar concept nonetheless. I, I do think that in the broader scheme of things, the export control regime and the regulatory regime that we were discussing earlier is in some sense a protectionist and protection for the US frontier models and their vendors to prevent their insights from leaking out via what's been widely reported. If you're in China, in many cases is public reporting, you have access via proxies that are in China or in, uh, friendly countries to access Anthropic and OpenAI and other strong Western models at a tenth the cost, 90% discount. So you pay a lot less, you get access to the models via these proxies. The reason for it is the proxies reportedly are gathering all of the reasoning traces. You agree to give up any notion of privacy in the reasoning traces. The proxies gather those reasoning traces using you basically as a sock puppet., and now those proxies can be used in principle for distillation or, or other efforts. I think it's an interesting question that I'm certain is going to be heavily litigated, whether distillation constitutes espionage or not. It's almost certainly a violation of the terms of service for Anthropic's models, but whether it constitutes espionage, I think, will be probably a heavily litigated question and probably yet another reason that we turn the world seems to be on a path, sort of a Second Cold War type path where there's a US bloc and a Chinese bloc. And it's not just model access that doesn't flow, it's also reasoning traces that could be used to enhance capabilities that also seemingly don't want to flow.

Can you block distillation and still maintain the openness of API calls and the usefulness of these systems?

There are different types of distillation from direct on the latent to just how do you answer this really hard question and make this really nice code base. The latter is incredibly difficult to lock down. And I think what China's basically been doing, the Chinese companies, is why does Merck Core have a billion dollar revenue? Because they get all the experts in. And why do you need experts to answer questions? Because your data distribution is this good and the experts make it that good. But you could use Mythos or Claude instead to get the extra level up. So they're substituting out Merkur and these other guys for that. But we have reached a plateau and we've reached another level. So if you actually look at why GLM 5.2 is better than GLM 5.1, and you look at what they've said, which I think is actually true— and I also think that Alibaba probably did have these sock puppets, as AWG kind of said, using the Claude code spare capacity It does appear to be a recursive self-improvement loop where you've got your initial distribution and then you improve it because it's got good enough and competent enough by being very thorough and looping back on the data improvement. So I think we're actually at the point now where Chinese models will be better if they have either super expert input or Mythos answering questions. But at the same time, you could have this recursive loop where you don't even need to have anyone's data anymore. Again, I think it's very difficult for us to conceptualize, but this is that takeoff scenario where you can't guarantee that if today China could never use any of the US models again, the existing data set they have, the techniques they have are not good enough to keep up with frontier capability. And so that's the difficulty here.

I, I think you, we gotta put a pin in this story and it's gonna come up again and again and again, but I I couldn't believe, you know, I mentioned earlier, my son just came back from China and he's got this fake Rolex and he's like, check this out. I got this in Shenzhen and it says Rolex on the front. It's got all the patent numbers and the inscriptions inside the clasp. You literally can't tell it's not a real Rolex. Cost like $25. Uh, and I bought it in the basement of the, um, state house in Shenzhen. Like they literally are selling illegal fake clones in the building where the governor of the province lives or works. I mean, like, wow, what a blatant disregard for intellectual property law. But then you look across the whole Chinese economy, and the growth of it is predicated on copying ideas from elsewhere in the world, stealing them and bringing them home. So it's like, it's in the DNA of the culture. And so now we're like, oh, it's shocking that they actually had 25,000 fake accounts looking at Trace. Like, anyone who's shocked by this is way out of touch with what China is actually doing. Um, so I think this story will come up again because this will be the excuse that the US and Europe and maybe South America use to try and crack down. Something— this will be the, the trigger that they use because they need to suppress Chinese AI somehow. Otherwise, again, it'll be out in the world and everyone will have it.

I remember I was hosting a meeting, a conversation on stage with Steve Jurvetson and Astro Teller in the early days of Singularity University, and we were talking about IP protection. And the concept was, you know, at the— in the end days of the singularity, IP will mean nothing because if you're dependent upon IP to protect yourself, you're just off because AI is going to just reinvent the product much better than you ever did, iteratively, very much faster. And so it's just going to be— you need to be constantly innovating, not trying to protect what you did years ago. And we can see it happening right here, right now.

Or IP means everything and IP supercharges— or AI supercharges lawyers, IP litigation lawyers, to do an amazing job, a superhuman job of protecting IP.

Yeah, I'm glad you said that, Alex, because I think that's a very likely midterm, like, sure, like, I think robots are coming, space-based data centers are coming, but I think we're in a hard takeoff of core AI right now. And the amount of intellectual property created in the next, say, 18 months will dwarf all of human history, like, by far. But it's still all virtual breakthroughs, you know, it's software, it's video generation, it's solving all physics, solving all math. And if you can't protect that intellectual property, chaos is going to break out globally. So we actually have to figure this out. And you can't, you can't just allow countries to rampantly copy, um, especially given that, you know, privacy is so hard to contain and copying gets easier over time.

Well, I think the point that, that Astro and Steve were making were you're not going to copy exactly, you're just going to use what exists and reinvent it better and create something that is uniquely an improvement on top of that. Imad, where do you come out on this?

Yeah, I think that it's going to get increasingly difficult. Like, you'll say, do Teams but make it not annoying, you know? Make Zoom that doesn't need to upgrade itself every 2 seconds. Like, I think that the creative capabilities of the AI, the copying capabilities, they are such that almost everything can be one-shot within a few years. And you have this period where it had to get to competence And that first broke with Sonnet last year. And now for almost all models, including open models, it's here right now. And then this new loop means that again, why do I need to copy when I can recreate but remix? And that makes it very, very difficult. Like in certain areas like music, incredibly strong copyright protections, you know, and that has a whole bunch of other things. Software, not much, you know, so you'll have this whole kind of gamma of things, but I just don't think you can. Stop the capability increases now by having any distillation or other lockdowns or even IP lockdowns.

Alex? I'll sound maybe just a different note here, which is I think there are striking parallels between the defense versus offense divide on software vulnerabilities and defense versus offense on IP litigation and IP protection. So one might say superficially, oh yes, sure, IP is over because AI will, for any patent, be able to find a way to route around it. That's the, the AI will, will overwhelm via offense argument. But at the same time, AI can also strengthen defense. AI can draft better patent claims. AI can do a better job of, say, patent litigation than humans will at some point in the near-term future. So it's not 100% obvious to me that the— call it the Astro argument— that IP suddenly evaporates because intelligence becomes stronger. I, I think intelligence— there are a variety of reasons, as we're seeing quite frankly with this export control regime, why superintelligence may want to protect itself via IP legal mechanisms. And if that is the case, I would fully expect the defense side— or offense, depending on your perspective— uh, side to be also supercharged with superintelligence. And not 100% obvious to me that IP goes away. In fact, I, I— it might just be utterly supercharged in terms of its ability to protect Guess what, guys?

We're gonna find out soon enough. Yeah. All right, let's turn our attention next to quantum computing. This week, another breaking story: President Trump signed a new executive order aimed at supercharging US quantum computing companies, a technology that could one day crack today's encryption and turbocharge scientific discovery, specifically in drugs, biomaterial science. In a parallel move, the White House is moving to shield US quantum research from foreign espionage, reportedly directing intelligence agencies to guard it as we guard our nuclear secrets. So the US government thus far has committed $2 billion in venture investments using the May of 2026 CHIPS and Science Act. And let's take a look at who's getting the money. So first off, IBM received the lion's share, $1 billion of the program to co-develop their Anduril quantum chip foundry in Albany, New York. D-Wave, Rigetti, and Inflection, each secured $100 million and PsiQuantum secured $140 million equity stake. So I'm hoping to have a number of these on stage with me at this year's Abundance Summit, hoping all three of you guys will help me grill these, these CEOs at Abundance and understand. So Alex, let's go to you first. How excited are you about this and what's your take on the quantum disruptions coming?

Mildly excited, not very excited. On the one hand, I want to quip that the US, the forthcoming US Sovereign Wealth Fund, will have an amazing exposure to all of these quantum stocks. Uh, one of the executive orders— so there were a couple of executive orders that dealt with it, but the, the more interesting one established that, uh, or rather required the development of what's called the Quantum Computer for Application Development and Discovery Science, QCAD. Which is interesting. I— first executive order I've read that mandates the establishment of a quantum computer for discovery science. So on the face of it, great, we want to accelerate science radically. On the other hand, I, I do think this may be a case of begging the question somewhat. There is already a vibrant private sector of quantum applications, not just quantum computing, also quantum networking, and most interesting to me, quantum sensing. So establishing a national quantum computer effort for discovery science, to me, uh, reminds me a little bit of, uh, Genesis Mission, which, if, if you look beneath the covers at where the money is coming from, it seems to be a repackaging of existing U.S. government funding. So that part I find less interesting. What's more interesting to me is the protection from foreign threats. So until the executive order, I, for example, was not aware. I'd be curious to hear, were you aware that there's a Quantum Information Science and Technology Counterintelligence Protection Team? Nope. We have that. We have a quantum protection team in this country. I think that's super interesting and in some ways evokes for me this idea that the US government was caught flat-footed by AI, the defense community in particular and intelligence community flat-footed. We saw, as we were discussing earlier, we saw Mythos and now GPT-5.6 leapfrogging whatever internal, apparently, capabilities the NSA has when it comes to cyber. In quantum, I think the thinking somewhere in the executive is not to be surprised a second time and to actually get out ahead of any quantum capabilities that, that might be strategically disruptive. The problem, as I've mentioned on the pod in the past, is quantum for science acceleration just hasn't worked that well. Quantum computing was supposed to give us protein folding. Turns out protein folding problem was solved by arguably AlphaFold 3 purely classically without use of any quantum computing. There are a lot of folks, including Peter Kamen, friends who were very aggressively marketing quantum computing specifically for solving all of these problems.

Well, to be, to be clear, the company I think you're referring to, SandboxAQ, is not using quantum computing. Quantum computing. They're using the equations of, of quantum physics on AI platforms. So, um, Imad, how do you come out on this? Are you excited? Is this a nothing burger?

Um, look, I think it's the potentially the next big wave, right? And unlike GPUs where China's catching up, like Huawei is about to release the 950s and others in terms of bulk but not edge, Quantum computers are vastly more complicated to build, right? Like, even if the secrets kind of get out, like China has a good thing in photonic quantum computers, but not the various types of a Rigetti or a D-Wave with the quantum annealers, et cetera. The most interesting thing is this for me, with super mythos level models, we'll be able to ask the quantum computers the right questions. Uh-huh. Program them properly. Yes. Program them properly, which is actually quite difficult to do. And SandboxAQ again are doing a a little bridge to that right now by having the equations. Although I think quantum equations and generative AI equations are very, very similar for a very interesting reason. But what type of quantum problem will require a quantum computer a day to figure out or a year to figure out?

Nothing. Versus milliseconds, right?

So what you've got is you have convergence of asking better questions and quantum supremacy and others coming, and that meeting point means we might not need Dyson spheres. And that is actually something—

double down, double down on that one for us, would you?

Because that's all we speak about on this pod is basically one of the things is more energy, more compute, more intelligence, right? But quantum, again, processes and questions can be answered almost instantly in microseconds. There— it's not like test time compute exists from quantum compute, but we're very bad at asking the quantum computers the right questions, so they're not a sufficient scale. So if we have an increase in energy to ask really good questions, maybe there comes a time where we can answer all the questions that quantum computers need and then meet in the middle through a mixture of GPUs and quantum computers to ask and answer almost any question. And that breaks this energy ramp increase to solve the mysteries of the universe by coating the entire universe with compute.

So I think I have to jump in on this one. I think it There's a latent assumption in, in this scenario that we see some sort of complexity hierarchy collapse. Right now, one of the reasons why quantum computers arguably haven't been that useful is because it's actually really difficult for humans without superintelligence to identify algorithms that are both economically useful and also achieve some sort of quantum advantage. We found a number of quantum advantage algorithms. They're not that useful, at least not economically transformatively useful.

They make for good headlines. Headlines.

They make for wonderful headlines and amazing IPOs. Uh, to Imad's point, I think, and this is something that I'm bullish on, that I do think there's a pretty good chance that if there is some sort out there in math theory space, if there is an AI-discoverable quantum-advantaged AI training or inference algorithm, AI will find it, and that will suddenly pay back all of the sins of money being invested too much in quantum infrastructure previously will atone for it. And but separately, the point of the Dyson swarm, I would say the reason why I think right now we're on a default trajectory of a Dyson swarm isn't necessarily because everything is so efficient, uh, but rather because we're running out of headroom in Moore's Law. And as we start to, on the one hand, lose room at the bottom, as Feynman would say, and on the other hand see skyrocketing demand for AI, and on the third hand, municipalities don't want data centers, we push them to orbit. Will quantum obviate the need for a Dyson swarm? I, I think they're probably not mutually exclusive. I could imagine scenarios where we build a Dyson swarm of hybrid or, or pure-state quantum computers, and people don't want quantum computer data centers in their backyard, just like they don't want classical computer Yes, computers are difficult.

Build the quantum computers in the permanently shadowed craters on the moon.

Yeah, yeah, look, I think the thing here again is the default path is energy, everything gets converted to compute, right? But for frontier capability and the type of capability we've discussed earlier on this episode that could disrupt nation states, that could disrupt society, you might find in the next couple of years with the way these two curves are going, a discontinuity And the United States is trying to ensure that it is on the frontier of that with the most capable frontier models, with the right algorithms for the right quantum computers that are finally useful. And if that happens and that meets, then it's very difficult to fight against that because again, you don't need to scale the test time compute if you can crack that.

Yeah, I think the right analogy there, so I agree, the right analogy is again something like we're in 1939 or 1940, and the goal isn't to infinitely, or rather indefinitely, preserve an American advantage in these capabilities. It's just to slow the rest of the world, namely China, down enough that we can hit enough recursive self-improvement and dominate, for whatever definition you prefer of dominate, the future light cone with capabilities that really only need to be a few months at most a year ahead of the competition.

All right, Dave, do you want to add on this, or should I jump into the next story?

Well, I started this new company, Quantum.AI, for exactly this reason. So I've got so many—

Okay, yeah, we'll go for it then.

Well, let me, let me just— I don't want to belabor it, but, um, September timeframe, uh, around the Moonshot Summit, if you want Alex to be super excited on stage, let's talk about quantum, uh, photonics, not quantum computers. Also, quantum sensing was almost certainly going to work. So you'll be able to store insane amounts of information in tiny, tiny spaces. But the photonics, you know, I've been working for 9 months now on this quantum AI and working on the algorithm side, but it's almost a certainty now that highly quantized neural nets can perform just as well as floating-point 32 neural nets, which opens the door to massive amounts of photonic computation efficiency. I would be shocked if by the time we're talking to Elon next December, I'd be shocked if we're not talking about—

Next December or this December?

This coming December. Yeah. Instead of launching NVIDIA chips and the huge power they consume, get the terafab started on the photonic compute at about 1/100th the mass for the same amount of computation. And it could be even more than that. 1/100th is a conservative estimate. And so the stepping stone to the discontinuity that Ahmad was talking about is clearly photonic computing, not quantum computing the way it's currently defined. But it's still quantum photonic. It's not quantum quantum. Yeah. And I'm— I think that's almost a certainty at this point that that will exist and that the current AI will discover the breakthroughs necessary, if there are any left, and that that'll be deployed and, and what Elon is actually manufacturing within a year to 18 months.

Yeah, I think that Intel's doing some super interesting things there, but you know, China with their Zhiyuan series and others, that's the one area they're really focusing on. And I think the US is completely under on. So I think that particular area of kind of quantum photonics and quantum photonics needs to be a much bigger focus and have much more investment given what we've seen already.

And for that reason, it should, that should be the thing the White House just elevated. You know, they're not aware of it yet, I don't think, but they will be. And it'll be right up there on par, or it should actually be a bigger priority than the current, you know, quantum thing they just passed. By the way, everybody—

go ahead, Alex.

I will note that again, in the news in the past 2 days, Elon slash SpaceX just acquired for several billion dollars a photonic computing slash communication company to merge in. So would not at all be surprised if photonics, which gives us in principle 1,000x clock rate speedup over these stupidly slow electrons, becomes a key part of the Starlink or StarMind plan. Well, so Google found a patent.

Google got a patent, it's worth looking up if you're bored, that does matrix multiplications in pure light. And they got that granted last year. They filed it back in 2023, but it got granted last year. So they've been thinking about it for a while too. So yeah, this is going to happen.

By the way, I said this a couple of podcasts ago. Now that Elon's got liquid stock with SpaceX going public, he's going to be on acquisition rampage. Watch him buy companies left, right and center. And if all of you want to come and join an AMA live with Alex, we're going to be doing an hour of ask me anything on Solve Everything at the Moonshots Moonshots Summit. Imad's going to be there. Dave Erickson's going to be there. Saleem's going to be there. Come and join us on September 25th. Go to moonshots.com to get your ticket. We'll be sold out soon enough.

So join us 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 and enterprise-scale codebases with millions of lines of code. Engineers start every development sprint with the Blitzee platform, bringing in their development requirements. The Blitzee platform provides a plan, then generates and precompiles code for each task. Blitzee 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 copilot 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.

I'm going to move us to our next conversation, which is a fascinating one for all entrepreneurs out there. It's about sleep. So here's the question we're proposing. Hopefully all of you know 8 hours of sleep is not just a good suggestion. It is evolutionarily what we're designed to need. But there is a small percentage of humans on the planet called short sleepers that can get away with much less, as little as 4 hours of sleep. The number of humans that get away with 4 hours of sleep is 0.1% or less. Uh, there's a group who can get away with 6 hours of sleep, that's about 1% of the population. Um, but the rest of us need 8 hours. And here's the data so that you're aware of it, because I used to say when I was in medical school, I'll sleep when I'll die— when I'm dead. And the fact of the matter is, not sleeping will kill you. So, uh, that if you're getting 6 hours or less on a regular basis, you have a 48% increased chance of coronary heart disease, a 15% chance of a stroke, increase of a stroke, 12% increase of all-cause mortality, 5% increase in beta amyloid. You know, this is what's going to be giving you neurodegenerative disease. 17% higher risk of type 2 diabetes and a 4x higher risk of catching a cold. Sleeping is critically important. So how do you become a short sleeper? Well, um, if you can get away with 4 hours of sleep, here's the math. Uh, check this out. It's 28 hours of more work time or play time per week. That's 58 days in a year. You gain 2 extra months on your life. This is a way of getting longevity incrementally. Instead of adding time at the end of your life, you're getting 2 months extra per year. So I know about this because one of my portfolio companies, which is still under stealth, is the number one player in this area. But it made news this week when Eli Lilly purchased for $6.3 billion a company called Centessa Pharmaceuticals. Their drug targets a neuropeptide called orexin in the brain. It's the brain's master on-off switch for wakefulness. So they've been developing this for narcolepsy patients, but I think ultimately, guys, this is going to become a lifestyle drug. Uh, how many of you would take this if you could get it?

If we could get it, or if we could get it legally?

Are you on it already, Alex? Is that what's, what's going on? Because I think, I think we need this to track the singularity for sure. I would dedicate those 4 extra hours per day for just reading, you know, reading the feeds that are coming in. Alex, what are your thoughts on this one?

I think this is potentially as transformative as the GLP-1 receptor agonists. This is Eli Lilly seemingly playing the, the same playbook over again. So just to refresh, the, the GLP-1 RAs initially approved for treatment of diabetes, uh, and taking advantage of, uh, this protein that was discovered, I, I think in in like lizards in Arizona. It was— oh God, what are the large— the gila monsters or something? No, anyway, reptiles, reptiles in Arizona that could survive without eating for long periods of time and were somehow able to maintain sugar/insulin balances for long periods of time discovered that actually humans have an analog of that. It seems relatively well conserved and If you can synthesize a receptor agonist, you can trigger common pathways and enable humans to balance their sugar levels over long periods of time. But it turns out the pathway is so important that it— and this is still something I think of mystery science right now on the GLP-1 RA side— you're able to have all of these amazing potentially lifespan-increasing effects. You're right.

Yeah. Gemini says it was in the Gila monster.

The Gila monster, Gila monster. So, so now same playbook. And I, I should add, like, the GLP-1 RAs took Eli Lilly and have made it a TerraKorn. Is it like a trillion-dollar company? And the revenue that it's generating off of just this is comparable to, if not larger than, all the token revenue that OpenAI and Anthropic are generating on similar timescales. So if you're Eli Lilly, you have to be asking yourself, is there— can you take this receptor agonist playbook and apply it to other domains? And it looks like Eli Lilly thinks the answer is yes, and they're going to apply it to neuroscience next, where the role of diabetes is going to be played by narcolepsy, the role of GLP-1s is going to be played by orexins, and the role of healthspan in inducing positive side effects across all these other areas is going to potentially apply to the ability to sleep less without the, the nasty side effects that you were mentioning, Peter, to help people who suffer from daytime sleepiness due to Parkinson's and Alzheimer's and mood disorders, potentially helping people emerge from comas because orexin is so essential to all of these neurological pathways. Yes.

Yeah, that's why I'm so excited about my portfolio company. I saw the data. It, it it's, you know, double the performance of this Intesa drug. So, uh, and if they're getting bought for $6.3 billion, that's hopefully going to be amazing.

Uh, Dave, just think of like one more point. Think of all of the economic output that could be unleashed if everyone everywhere had 4 more hours of wakefulness and compare that. Yeah, yeah.

And that's, that's my argument for adding 30 healthy years in a person's life as well. But imagine if you had it in your early 20s when you're most productive, right? Yeah, crazy. Dave, what are your thoughts on here, on this one?

Yeah, we, uh, one of our partners, Mira Wilczek, actually, Alex, one of your many classmates, um, she was a natural 4-hour night sleeper. You know, there's a bell curve there. It goes all the way down to 2 hours. Some, some people are 100% functional on 2 hours sleep. It's very, very rare. Most people are in center points on 8, but some people need 14 hours sleep. Can you imagine how much would that suck? But Mira got so much done. It's like having a whole nother life. I was so jealous of that. So yeah, I'd take it in a heartbeat.

Iman, how much do you sleep? I'm a 4-hour sleeper.

Are you? You lucky bastard.

Jealous, I thought so. You get the time in.

And no, I mean, I'm very bullish on Erexin. So, you know, we had a long look at this with ASD and sleep disorder. And, you know, before you could only get things like Lactobacillus, Bioguia, and others to affect it. But you see a lot of knock-on impact from this actually when you upregulate oxytocin and orexin to inflammatory markers and things like ghrelin and leptin. And I think you could see again, GLP-1 now is impacting all these other biostatic stabilization mechanisms. I think this pathway will do the same, especially for inflammatory disorders. And so I think you sleep less, but you become less inflamed is what you're going to see as a result of some of these treatments.

So exciting to be alive right now. I mean, this is like my abundance story for everybody. It's like, this is the time to be alive and just get excited about what's coming down the pike. All right, you guys ready for some AMA questions with the mates? Yeah, let's hit it. Fantastic. All right, let's jump in. Here we go. Imad, you're our Supermate guest for the day, so why don't you choose first?

Oh gosh, it seems a lot of UBI questions here. Um, can UBI be as simple as reaping dividends from US government-held golden shares? Do not freeze 79. The math is impossible. Like, if you look at how big the AI companies would need to be to get a basic living level of UBI from dividends, assuming 5%, it's about $10 trillion, and you'd have to own like half of them. To get to like halfway there. They'd have to literally be the biggest companies in the world, and it would only be to the US, which is question 4. How can anyone outside the US survive if it's only distributed in the US? I think we need to fundamentally look at how capital flows, and monetary equivalents of UBI where money is created make a lot more sense. And we've seen some explorations of that. And then from that, we really think, and how do we value things when the AIs and robots basically dominate the world? And I hope that we see a lot more research and trials on that.

Yeah, I did the math, right? If you imagine a UBI of $3K a month for just US citizens, US residents alone, that's $12 trillion per year. And the US budget's $7.4 trillion per year. So there's a lot of capital to be made up there.

So I'll take question number 2, which I think was aimed at me. Will companies migrate to friendlier jurisdictions like Argentina to escape US AI regulation? Restrictions, wouldn't this erode trust in any model that can be banned? And this is from Evraninank. So I don't think model companies are at all likely to migrate to Argentina. I, I think the frontier model companies are going to very likely remain firmly entrenched here in the US, and you'll see Chinese companies firmly entrenched in China. And then there will be the rest of the world. I don't think Argentina is likely to end up hosting any frontier model companies. Companies that I can conceive migrating to Argentina under its proposed new non-human AI corporation regime will be inference time companies that are basically AI personaments. Uh, so, so a humanless AI company that's just operating its own business, I can totally see many millions, billions of those types of non-human AI companies migrating to Argentina. That said, I would think, given the, the export controls that we've just been talking about on this episode, I would imagine many of them will probably end up running on Chinese models, uh, unless the U.S. acts to restrict the ability of Argentina say, to import Chinese models, which could, by the way, happen. We've seen, uh, under a variety of non-AI technology regimes, we've seen this executive act to restrict other countries that are neither in the Chinese bloc nor, strictly speaking, in the U.S. bloc, restrict their ability to import Chinese technology or Chinese commodities. So I can imagine a scenario where the U.S. acts to restrict Argentina's ability to, to import cheap or, or operate cheap Chinese-based models, which would obviously foul up that approach.

Alrighty, Dave, uh, number 1 or 4?

Yeah, I'll take 4. Uh, how can anyone outside the U.S. survive UBI, UHI if it's only distributed in the U.S.? And that comes from Ali Singh, or Ali, Ali Singh, who I love this question and it's very timely. The World Cup is going on in the US right now and there's people from all over the world crawling around Boston, LA, Santa Clara, and Mexico too. And by and large, the people in Boston are like, wow, America's awesome. This is great. And the word across Europe, Imad will reinforce this, is like, America sucks, America sucks. You get over here and you're like, well, this is just a cross-section of every type of person from all over the world who's immigrated to the United States. And it's where AI is happening. California is just hopping and it's just fantastic. And the problem you run into is that the US government continually panders to the voter. And if you're not a voter, they just don't care about you. And if you are a voter, then you're entitled to everything in the world. And that has to change. And I think this is a moment in time time where, you know, you saw at the G7 summit when Donald Trump walked in, he said, "Okay, the boss has arrived." And that's not exactly the dynamic you want. So this, it's gotta happen now though, because the wealth concentration effect is so extreme. And if these frontier models, you know, become the universal workforce that creates everything, it's all gonna happen in just a couple of locations, basically in the US and China. And so we need to get this figured out really during this administration, which really means in the next year or so.

So, all right, I'll take question number 1. It's from @JB-C1BR: Why should people give up autonomy to accept UBI if unemployment is not going to be a problem? So, uh, JB, I think you've got the premise backwards. UBI isn't a trade for autonomy, it's a foundation for more of You know, you can think about it, you know, it's less of a welfare cage and more like the Alaska Permanent Fund, which I've spoken about before, where every citizen gets a dividend check and you can do with it what you want. That's right. If AI is creating extraordinary abundance, distributing a share of that to the populace allows you to use it to uplevel your life, to create your next company, to create meaning in your life, however you might want to do that. You know, even if there's mass unemployment, it doesn't materialize. You know, UBI is a freedom dividend, and that's the way I think about it. And anyway, so that's my, my answer for you, JB. All right, let's go on to the next set of questions. Uh, Dave, why don't you kick us off?

I'll take number 5 and then I'll throw it over to Alex. Why is the moon better for data centers than orbit? Is it the gravity or the Earth-facing position? Definitely not the Earth-facing position. This topic came up when we were talking about the Kessler effect, and we really did a good job of that on the last podcast. So low Earth orbit is great for data centers, but then you get outside 500 kilometers altitude and you get the Kessler effect problem and all kinds of debris flying around, yeah, destroying your data centers.. So the best place is low Earth orbit where there's a little bit of atmosphere that cleans the system naturally, but that's a relatively narrow band. The moon was good for the same reason. It's somewhat protected, but there is no atmosphere on the moon either. So that's why I'll throw it to Alex. Like, why, Alex, why do you think the moon is better for data centers?

Oh, I would say they're complementary, but moon, the cislunar environment certainly has a number of advantages that orbit does not. For example, if you're concerned about the security of your data centers, there are— and, and both the West and the East, the great powers, have demonstrated the ability to on-orbit send robotic devices in and with grappling arms. China especially in the past year has very publicly demonstrated this— interfere with on-orbit devices. Whereas if you have a data center sitting on the lunar surface, You can defend it. Um, you can be sure that there's no one, for example, coming in behind you, uh, listening to the same beam that you're using to communicate with the Earth on. Uh, so there are a few reasons from a security perspective. Also, the Moon has mass. If, if we set up, as I think Elon and hopefully others are going to do, an industrial ecology that's non-terrestrial, you can start mining water water and other elements— water is not an element, obviously— but mining raw materials from the lunar surface to build more data centers. You can't build more data centers in orbit. You just don't have the feedstock to do that. You have—

Yeah, I was just gonna say lunar regolith, for folks, that's what you call the lunar soil, is silicon, oxygen, nickel, and iron. It's perfect for data centers.

Perfect for disassembling the Moon, in other words.

Yeah, there's also, there's another version of that which is highly likely where you're using those elements and then you're using a mass driver to launch them into low Earth orbit. That, that is something that's definitely going to happen because it's on the, it's on the planning radar already. And it's the ultimate high ground.

I mean, the, the one of the narratives for why have an Apollo program at all was that it was a continuation of the Manhattan Project. Project, uh, and with the Moon as the, the ultimate high ground for, for launching weapons. Well, it may be the case that in addition perhaps to the Moon being the ultimate high ground for launching weapons against Earth, it's the ultimate high ground for launching data centers to surface Earth.

Kudos to my— one of my mentors, Gerard K. O'Neill, professor of physics at Princeton University, who ran the Space Studies Institute and wrote about this, actually built some of the first mass drivers. And he laid out entire architecture of mining the moon. His vision wasn't data centers, it was actually building and launching solar power satellites to Earth orbit to provide solar energy on rectennas on the ground. We've changed that a little bit because we're gonna use the energy in space for creating intelligence, but he laid this out in the '80s. Amazingly brilliant individual who left us too early, but Gerard K. O'Neill, look him up. Alex, choose your next one. Let's leave number 7 for, for Iman. Yeah, go on.

I think I have to choose number 6. So what, what physical and behavioral forms will humans and other species take in the age of the singularity? And this is from Jognalan383. So I'll construe the question that I would argue we are in the age of the singularity, so we already know the answer. We look like ourselves. I'll construe I'll construe the question instead as what physical behavioral forms could humans take after the age of the singularity? So I think I've argued in past, there will be many new forms of person and personhood. I take a lot of heat in especially the YouTube comments for agitating for some form of AI personhood. That's a legal form, not a physical or behavioral form. To the extent the question is asking what will post-humans look like, if you will. I think we'll see uploaded humans that have relatively de minimis physical form but are just a collection of bits or maybe qubits running on AI infrastructure. I think we'll see, to earlier discussion, borganisms, so collective human intelligences, whether it's via Neuralink or, or some other format. I, I tend to think as we discover new physics and new applied physics, the substrates for the compute that humans right now, we augment ourselves with, soon many of us I think will be running on. That substrate I think is also in some sense, to parrot Bucky Fuller, is going to ephemeralize. And at some point in the distant future, how distant? Not sure yet. We might even see something that Arthur C. Clarke wrote about many times, which is maybe at some point humans will just be able to run on the gravitational fields or operate in a state of something approximating pure energy with, uh, with fewer biological meat bodies. At the same time, before everyone attacks me in the comments, I will point out— inevitably, I will point out— yeah, I'll point out I think this will be both purely optional, and I, I think the future is going to look much more heterogeneous, not homogeneous. You're going to see humans humans who look substantially the same as they do right now 100 years in the future. At the same time, 100 years from now, you'll see some post-humans who were uploads in a cloud functioning in a quantum computer. At the same time, they could— those, those very different forms, I think, can and will coexist next to each other.

And a quick shout out to our beloved subscribers: we do read your comments every week, so please add them, ask us your questions. I'm going to add to what you just said, Alex, because you're going you know, pretty far out. Let me go in the interim. We're going to start to edit ourselves, right? We just talked about the idea of an Erickson-like molecule allowing us to shift to 4 hours of sleep. We're going to see gene edits, you know, injection of Clothro to increase our IQ, or other gene edits increase our muscular ability. We're going to start to see BCI, people walking around who've got, you know, a connection of the neocortex to the cloud. So I think those capabilities are coming in the next, you know, years to decade.

Can I ask you a question about that, Peter? Just narrowly. Sure. So we saw some people not take too kindly to the enhanced games. We see some municipalities not take too kindly to building data centers in their own backyard. Do you think there's a high-likelihood future where all of these biological enhancements are actually either tightly regulated or shunned, uh, such that, as with pushing the data centers into space rather than building them on land, the transformative, uh, physical and behavioral changes to humans are basically all pushed into the post-human realm, uh, by humans who regulate out of existence more obvious short-term biological enhancements.

You know, I had that conversation with one of my boys, and I said, listen, morals and ethics change over time. I remember when the first in vitro fertilization efforts were taking place, it was thought, oh my God, this is awful. You shouldn't allow this. This is immoral. This is not what God, you know, desired. And of course, now IVF is considered normal and allowed and beneficial to allow couples around the world to have children later in their lives. So I think it at first is going to be shunned maybe by more religious elements. But I think ultimately it's going to become accepted. Shout out to Ramez Naam, who wrote an incredible book called Nexus, which talked about all of these genetic edits to create, you know, BCI capability, basically neural lace in the brain, but also the edits. A lot of it will go underground and then eventually will become, of course, I've been enhanced, you know, why would I not want to? You know, we all want the best for ourselves and our kids, the best education, the best food, the best tech, the best whatever. Why wouldn't you start with the best genetics? I know this is a sensitive subject, but I think it's going to become part of life going forward. Emad, what do you think about that?

Oh, it's going to be fascinating, isn't it? Uh, actually, um, I spoke at the Oxford Union a couple of weeks ago in a debate about whether AI can attain personhood. And, you know, obviously one against it. So, uh, that's coming out soon. So you see the argument against this. These are big questions that we're going to have to ask. I think ultimately we're all made of star stuff, but some of us will be more star stuff than others is the way I like to think about this.

Would you take number 7 for us?

Yeah. For e-entrepreneurs lagging the US-China, is the highest leverage more vertical AI startups or agentic workflows into legacy industries? Industries. That's from Steve Krabb. I think it's fascinating because, you know, you can't compete on frontier models for various reasons, but also the nature of regulation in the EU means that there's massive potential for the latter, agentic workflows into legacy industries. In the US, in Silicon Valley, the initial diffusion of innovation is happening really rapidly because companies are open to it. You know, the mass majority of the US hasn't. In Europe, it's even further behind, but it's inevitable because you get the competitive pressures. So being able to go in and work with those companies is a massive transformative opportunity because it's inevitable. Whereas building virtual AI startups in the competitive regulation, it varies from industry to industry, but it's just not as easy as the US. Like, we have far more regulations. So I would say it's more about transforming companies in a that's comfortable to them and charging big markups than necessarily building the verticalized AI startups where really the US is a much better place to go to compete globally.

Imaad, I have to ask you the question, is Europe waking up?

Yeah, I think it is slowly. Um, it's just there's so much institutional inertia here. So there was something EU 2031 as a warning call. It's quite a nice kind of future thing like AI 2027. You know, but you see, we have an extra 4 years because we're a bit slower here, you know.

We're— instead of humans being disempowered by AI, Europe as a whole is disempowered by AI.

Oh, our regulation will be the great filter, you know. It won't stop anything from coming out. Ouch. Um, so it's slowly waking up. That's got a lot of traction in the upper circles because this Fable thing really was a massive shot across the bows of the decision makers. So I think there's a good chance the EU AI Act gets repealed. People are trying to figure out new mechanisms and suddenly the fire has come from nothing. Maybe because it's so bloomin' hot here as well.

You know, I hope that, I hope that, uh, that Argentina-Milos thing also goes really, really well and creates a role model. And then some European countries say, you know, like Ireland to me is a natural, like Estonia needs to pick it up.

Yeah, we have actually introduced this new thing, EU Inc., whereby you can set up a company in under a day. And that's a big deal because like in Germany, for example, across 6 months, across all of Europe, across all of Europe. Yeah. Cause like in Germany can take up to 6 months to get your tax status and other things like that.

So little bit by little bit. Wow. All right. I'll take number 8. Uh, this is from Mark Simonian 210. Uh, wouldn't it be easier to put data centers in oceans and the Great Lakes than flying them 93,000 miles up? Maybe that's a transcription error. Uh, but Mark, we're not talking about 93,000 miles. We're talking about about 500 miles up into low Earth orbit, or on the moon, it's 240,000 miles up. So to answer that, Microsoft already proved the ocean concept. I looked it up here. Project Natick ran subsea for 2 years and lowered failure rates compared to land-based servers with excellent cooling. Of course, the advantage of going into the ocean is you don't have to deal with launch costs and there's plenty of cooling relatively nearby for repairs. And you can either have it within your jurisdiction or offshore sufficiently so it's outside the jurisdiction. But the reality is, as launch costs are dropping, and we heard about this in the last pod with Will Marshall, that apparently, you know, Dave, you and I were talking to Elon saying he flipped the bit and he did. What was it now, 9 months ago when he started talking about— no one was talking about orbital data centers, but apparently Will Marshall and Eric Schmidt and folks at Google were talking about this as much as a decade ago. Ultimately, it's going to be both. We're going to have data centers on, on land, in the oceans, in space. There is as much demand as we can supply. So I don't think there's a limitation there right now. Alex, any, any more thoughts on ocean-based data centers?

Yeah, I'm a big fan. Peter Thiel funded, I think it's Panthalassa, that's focused on this. Easily, right? Yeah, so I think, look, my operational, one of my operational definitions for the singularity is every sci-fi trope happening everywhere all at once. And I think data centers on the ocean are the prototype for seasteads and ocean colonization. And so yes, I think we get our data centers on the oceans. I don't think It's as scalable as orbital compute. But then again, if you want to, say, live on an ocean colony, which is, I think, something that we are going to get over the next 10 years, then you're certainly going to want your own local compute. And I could totally imagine, if you've seen the Panthalassa video demos, that that ends up becoming a nucleus for a next generation of ocean colonies. Hey, one thing to add to that—

Please. If the photonics takeoff that we were talking about earlier is on a 1 to 2 year timeline, which I think it likely will be, people will want to put the NVIDIA data centers in the ocean and not launch them because they're very heavy, you know, a couple of tons for an NV72 with power. That's cheaper to deploy in the ocean while you're building the photonic thing, which is about, you know, a tenth, a hundredth, a thousandth the mass per compute. And so that's a real possibility that we have both, but in that order.

Uh, Emad, before we go to our outro song, give us an update. How is Intelligent Internet doing? What are you up to these days? Uh, give your fans a little bit of an Emad, uh, preview.

Uh, yeah, no, it's going well. We're, like I said, releasing the harness that allow anyone to uplift the models above Mythos on Monday. Um, and we, we—

where do they go? Where do they go for that? Because by the time we release this, it'll be up.

It'll be at, um, Probably my Twitter @emostac. I'll retweet it. And yeah, then we've been working on Sovereign AI, and I think we have a mechanism whereby everyone can own the AI and robots. And so we'll be announcing that soon along with the new book, uh, After the Last Economy. So all going on.

Amazing. And I know the secret stuff you have going on, which is incredible. Uh, appreciate having you in the universe. And, uh, Alex and Dave, love you guys. A quick shout out to Salim. Salim, who's now on stage in Spain, or at least landing in Spain. Come back, Salim, we miss you. All right, this is an outro music by @roytheinventor called Moonshot Masters. Let's take a listen.

From the shores of Ireland, across the sea, we're watching dreams become reality. Moonshots lighting up the sky, showing us how far humanity can fly. Every story, every breakthrough, every spark brings a little more light into the dark. From AI to health, from space to the stars, you're helping us see who we really are. The future starts tonight with bold ideas taking flight.

Moonshot showing what's possible, making the impossible look logical. Hand in hand, we'll build what's right for every child, every life.

Together we'll create a brighter way.

Oh, amazing.

Love that. So everybody, please submit your your music, your videos to us. We love them. And remember, this is the most extraordinary time ever to be alive. Our mission here, to give you an optimistic view of the future, help you understand what's going on day on day. I hope you enjoyed this extra emergency podcast we stuck in on a Saturday morning. Iman, Dave Erickson, Alex, love you guys.

Be well. Love you too. Have a great weekend.

If you made it to the end of this episode, which you obviously did, I consider you a Moonshot Mate. Every week, my Moonshot Mates and I spend a lot of energy and time to really deliver you the news that matters. If you're a subscriber, thank you. If you're not a subscriber yet, please consider subscribing so you get the news as it comes out. I also want to invite you to join me on my weekly newsletter called Meta Trends. I have a research team. You may not know this, but we spend the entire week looking at the meta trends that are impacting your family, your company, your industry, your nation. And I put this into a 2-minute read every week. If you'd like to get access to the Meta Trends newsletter every week, go to diamandis.com/metatrends. That's diamandis.com/metatrends. Thank you again for joining us today. It's a blast for us to put this together every week.