Gavin Baker

Oh, yeah. And I think this will be part of, you know, the upcoming negotiation between America and China. But it is hard. You know, if we can't, you know, if America cannot keep illegal drugs out of, you know, out of our own country, these GPUs are arguably dramatically more valuable per capita

unit you know like a blackwell is like you know the size of this and it's just hard particularly like you know america we're trying to keep drugs out here china is trying to bring the gpus in and it's kind of significantly you know i think harder to prevent than than you know smuggling of illegal drugs now everybody does their best but um

Yeah, the United States, I think it's always going to be kind of a game of cat and mouse until you get some sort of grand bargain. And right now they're allowed to sell certain kinds of GPUs into China, and I'm sure there'll be

I think we're putting enough friction into the system. that it does theoretically give America an advantage at the cost of creating tremendous incentives for China to develop their own semiconductor ecosystem and, you know, pressure to kind of, you know, necessity is the mother of innovation.

And these export controls are creating an immense incentive for China to be really algorithmically innovative. And you saw that with DeepSeek, where there were some real algorithmic innovations.

In the next five years, I think zero. I think it's really, really hard. But over the 10 years, who knows? And if you're the CCP, 10 years isn't that long. If you're America, 10 years is an eternity.

Look, I do think if... Agents materialize as a reality, and Manus is maybe a little bit of a chat GPT moment for that. I would say OpenAI and Anthropic. Anthropic developed something called the Model Contact Protocol that OpenAI just adopted. I think it will become a standard, and it makes it really easy for an LLM.

Well, so first thing he did, there was Alfred was from Sequoia.

Stripe can just integrate with MCP, and then any LLM that uses MCP can interact with Stripe. And this is solving a big, big problem for agents in terms of just making them much easier to use, much more standardized. But if agents become a reality, one, the ROI on AI and Blackwell is going to be very high.

And two, what will... It's not good for human employment, but what will be kind of, I would say, the rate-limiting factor... is just compute. In a world where we all have agents doing things for us all day long, it's going to be a long time before we have enough compute in the ground for that to be a really, really widespread reality.

And I think that's why OpenAI was talking about pricing their first agents at $2,000 to $20,000 a month. Like this ROI on AI question, I think to me really does come down to agents in the short term.

And then they had a nice sell-side guy whose name is escaping me at the moment. But it is technically true. I was the only public equity investor on the buy side that Jensen asked. I've known Jensen for 25 years. He's kind of the same guy he ever was, just maybe slightly calmer.

Yeah, so I guess maybe take them in reverse order. I really don't think, like if NVIDIA had not put any money into CoreWeave, not put any money into kind of other of these neoclouds, I don't think it would have impacted their revenues at all. At all.

What should our tariff policy be or what is it? I mean, we know what it is for autos. It's 25%.

Yeah, I mean, this is... Traditional economics would say, hey, tariffs are not a good idea. We should have free trade. Everybody understands the principle of comparative advantage. I do think the Trump administration is pretty convicted that that approach has not worked well for America over the last 20 years. And maybe it's worked well for knowledge workers, but it hasn't worked well for kind of

ordinary Americans. And I think they're very convicted in changing that and trying to bring back kind of good high quality manufacturing jobs to America. I they see tariffs as an effective way to do that.

Yeah, more to Meta, more to Amazon, more to Microsoft. So the reason they did it is even three years ago, it was a very stable three-player oligopoly from the cloud computing players in Amazon, Google, and Microsoft. And if you're NVIDIA, that's not great to have just three big customers.

Well, one, I think it's Chamath in that clip illustrated. He is convicted in this. It's 40 years old.

Yeah, it's happening. And so I would say two things. Like, one, I think it's something I'm always conscious of. At, um... My wife, Becky, one of her college or high school reunions, there's a guy there who'd been in our class and had been in the Navy SEALs, and he deployed to 80 countries. He'd been in the Navy SEALs for like a decade. And I said, what's the one thing that you learned?

And he said, the one thing I learned is everybody in America is always focused on making America better. Have you been to 80 different places all around the world? our only goal should be to not screw it up in America. Just don't make it worse because America is so much better than everywhere else. So the first thing is like, you know, Chamath said the word delicate and I think that's right.

And if I had one thought for the administration, it would be every time they say the word tariff, whatever they think, Wall Street and the markets, and I would say, you know, I think a lot of business leaders are convinced that tariffs are bad. Now, maybe Wall Street is wrong and the administration is right. One thing everyone agrees on is deregulation is good.

So every time they say the word tariff, they need to say the word deregulation two or three times. Because... The best way to, I think, maximize the odds of this policy succeeding, despite the headwinds from automation, I think it's going to be a long time before we have hundreds of millions of robots.

Even between China and Tesla, I think it's going to take a long time to make vast numbers of humanoid robots. The best way to encourage this reshoring is just making it easier to do business in America.

Those big clouds, particularly in 2023, when there was such a rush to get GPUs, they each kind of want to do their own kind of custom version of an NVIDIA server. And NVIDIA, by giving kind of like a standard reference design, which someone like CoreWeave bought to CoreWeave, you got GPUs in the market a lot faster.

And when you use NVIDIA's reference design, it's generally smoother and easier to stand them up. And so the big three kind of cloud computing hyperscalers, their incremental share of revenue went down and that was nothing but good for Nvidia because they now have kind of a more fragmented base of buyers who have less power over them.

Thank you. Thank you. Thank you. Thank you. Thank you.

Thank you. Thank you. Thank you.

So look, it's never good. But if you think NVIDIA has gone through the biggest product transition in history in terms of Hopper, you know, this is a business doing tens of billions of dollars at scale. There's never been a product transition like this in the history of semiconductors going from Hopper to Blackwell.

you know kind of the 2022 generation gpu to this one the only precedent for this on planet earth is the iphone and there's something called the osborne effect which is you know when apple comes out with a new iphone for the three months before nobody buys the new iphone

No, I'll just say three really quick points. Like first, what Howard Lutnick literally said was if we can find a lot of waste, fraud and abuse, we don't have to raise the retirement age. Everybody's talking about raising the retirement age to 70. If we administer these programs more efficiently, we can keep it at 65. And, you know, that's good for everyone.

The second thing is, it's just interesting to me. There's so much doubt about this when the inspector generals of, you know, both Biden and and Obama, you know, said there were hundreds of billions of dollars of waste, fraud and abuse. And I would just say someone in my immediate family has had their social security, their identity stolen.

And someone is using their social security number to collect unemployment benefits in Alabama. And this has been going on for two years and they've been unable to stop it. And so, you know, you don't want to rely too much on anecdote, but I think this is really happening.

And then just the third thing I would just say is I do think, you know, no one has ever raised, no president has ever raised my taxes more than Trump did. And I think they are pretty convicted in their focus on normal working class Americans and making their lives better is actually genuinely their North Star. You can disagree with the theory of the case.

You can disagree with the way they're communicating it. But I think their intentions are good.

I think maybe we start with... I just think it is... whether it's a semiannual or quarterly, something that is like the Doge report and everything's been audited and triple checked. These are the three categories. This is what we found. I think something like that would be very helpful.

And I think the only reason NVIDIA was able to grow through this product transition is because of reasoning models like DeepSeq, which are just so compute hungry. And this product transition does come back to the day's receivables.

Yeah, it's never good when accounts receivables go up. But if this is the most understandable time for it to happen... In a product refresh cycle, you're saying. Yeah, and because... A hopper server, you know, it's a rack, it's whatever it is, seven feet high, it weighs a thousand pounds, it consumes 60 kilowatts, which is 60 American homes, and it's air-cooled.

So you cool those GPUs, and the biggest problem with GPUs, one of the biggest, is they melt. You cool it with air. Blackwell weighs 3,000 pounds. So three times as much the rack, you know, it's eight feet tall, maybe five feet deep or four feet deep, 3,000 pounds. And it consumes 120 kilowatts.

So twice as much power and roughly kind of the same footprint, three times as much weight and it's liquid cooled. So this is like an iPhone upgrade cycle. You know, it was like, It's still a pain that we went from, you know, lightning to USB-C. You know, I'll grab a cord and it'll be lightning instead of USB-C.

This is like an iPhone upgrade cycle where to get the new iPhone, not only do you have to change your connectors, you have to put in a new generator, you have to put in a new boiler and a whole house humidification system, you know, to kind of make it work. So it's a monumental product transition and I think that is a lot of why the receivables went up.

Marco Rubio did say someone made a big mistake. There were no consequences. We learned we're going to do better. And he is the secretary of state. So someone in the administration did say that this was a mistake.

No, no, no.

They were recognizing Blackwell revenue while it was still kind of getting into customers' hands. If that trend continues past the July quarter, then I would say, hey, that's reason for concern, but this is a very understandable time for that to happen from my perspective.

Yeah, I would just say I thought the contents were interesting and that there are a couple of things. You know, there's all this, you know, oh, the Trump administration is, you know, cozying up to Russia. I mean, you know, trying to blow up these, you know, this post-World War II alliance with Europe. There are two things.

One, they said we are the only people on our side of the ledger who can deal with the Houthis. And our side of ledger, meaning like recognizing kind of Europe, Europe traditional alliances.

The good team. So I thought that was interesting that just for all the rhetoric, they're conscious of who's the home team. And second, it was very clear to everyone in that chat, and JD Vance raised it, 3% of America's trade goes through the Suez Canal, 40% of Europe's trade. So we are really doing this. It will help us a little bit, but it will help Europe a vast amount.

And we're doing something that helps them that they cannot do. Now, hey, they're asking that they want to be paid, but I just think it made me think that behind the scenes, a lot of this rhetoric about the US and the European schism is overblown. And we're still helping them. I thought that was interesting.

If innocent people were sent to a prison, that's a terrible mistake. And that happens. And it shouldn't happen, and hopefully it gets rectified. But I guess my bigger thought is a little bit This administration, as we discussed, they have a really ambitious agenda. They want to fundamentally change America in ways that they think will be better for kind of blue-collar, normal working Americans.

And things like this signal gate, if they made a mistake and sent innocent people to this prison, there's only so many of those mistakes they can afford before they lose kind of the mandate necessary to accomplish their goals. And, you know, hopefully it's, you know, we're in whatever. We're not even into the, you know, third full month. 66 days. 66 days. We're into the third month.

They just... Execution matters. Like, if they... To accomplish their goals, they need to execute at a high level and communicate clearly and effectively. And just... If innocent people were sent to, you know, this prison, mistake. Signalgate...

you know hey maybe it's a big mistake maybe it's a small mistake depending on where you sit i thought marco rubio's response was good but just like i think execution is important for them to accomplish their goals great freeberg any thoughts here i do not agree with sending people to prison or detention center without due process i felt similarly about guantanamo bay because i don't think it

I mean, the accounting gets very complicated with these systems. But yes, I mean, they have delivered it and they recognize the revenue. But maybe the customer is saying, hey, we're not going to pay you until X, you know, until we get them plugged in and working. I mean, who knows what it is?

I agree so much with what you just said. Due process is important. Like America, like human rights are a core American value. And it was great you acknowledged that like the outcome in El Salvador was good. Homicides down 99.9%, lots of lives saved. And just, you know, there needs to be a balance between these.

Yeah, but it makes total sense to have accounts receivable. It's just that here the accounts receivable have gone up more than sales. And that's just a function of the complexity of this product transition.

This whole debate is a question of ends versus means. And it's like, we all agree that the homicide rate down 99.9% in El Salvador is a good outcome. It's a good end. And this is one of the great philosophical questions. And it's, you know, we can debate it endlessly. And there has to be a balance between the two. Human rights are a core American value.

And I agree, I think, with a lot of what David said. But I also think if you are You know, going back to that homicide rate, the Trump campaign, you know, they spoke to the mothers and the families of people who were brutally murdered or raped and murdered.

by people who were illegal immigrants and were convicted criminals in their home country and were known convicted criminals and somehow still were not apprehended and then committed terrible crimes that were very similar to the crimes they committed there.

You have a lot of compassion for the victims too, but all they have to do, like one of, I thought, Elon's best moments, and I think, you know, it's what Doge is doing is great. These are brilliant people working hard for free to do things that have never been done before in the government in terms of efficiency and kind of modernizing IT.

But Elon said at the first press conference, he says, listen, we're going to make mistakes, and then we're going to quickly correct them. So to me, I think all the administration needs to say, and I do think more due process, to David's point, is good, is like, listen, if we made a mistake, it's terrible, and we're going to fix it immediately.

And it just gives you so much credibility when you say that.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

, , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , tag

, , , , , , , , , , , ,. P. P. P. G M...T...T...T...T...T...T...T...T...T...T...T...T...T...T...T...T...T...T...T...T...T...T...T... P...,...,,...,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, K

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

Absolutely, and sending NVIDIA GPUs to people who are adopting NVIDIA's reference architecture, which they believe is the best architecture for those GPUs, whereas each hyperscaler has their own slight tweaks to that server.

That hyperscaler does it so it looks more like their other servers, but maybe it doesn't run quite as well as the standard NVIDIA reference architecture with all of their chips, not just the NVIDIA GPUs. But the Intel Insight analogy is interesting. I had not thought of that. Thank you, Jamal.

So we're definitely not overbuilt yet. Since DeepSeek came out, what has happened since DeepSeek R1 came out is China is buying every GPU they can. One of the largest Chinese server manufacturers warned last night that there's about to be a shortage of GPUs in China. The price of memory DRAM goes up every day. So we're not overbuilt today.

Point number one, and OpenAI yesterday said that they are gating their new image generation service because they don't have enough GPUs. So we're not overbuilt yet. Blackwell is going to be a very successful product cycle. There is a little bit of a prisoner's dilemma where

You're going to spend on Blackwell almost no matter what, because if you don't, you're afraid that you cede a big advantage to a competitor. If you're meta and you don't spend and Google spends and their AI is a lot better than yours, you're worried you may not catch up.

Yeah, that's already the case. But here, this is a brand new architecture, which is a lot. I mean, not only is it really different as we went through, it is a lot better. But I would say all these GPUs, they're hard to get them working at the beginning. Like right now, hoppers are finely tuned. Everybody knows how to work with them.

blackwell you know it's almost like you're you're skipping ahead you know it's like imagine each one of these gpus is a formula one car you know now you've skipped ahead 10 years in the future and you have to kind of learn how to drive that new formula one car and it takes a while till you drive it as well as the old one so any other thoughts gavin on the core weave ipo which will have happened by the time people listen to this episode yeah i would say you know the sentiment on x investors on x are pretty negative on core weave i think there's maybe a little bit of

negativity in the market. They had to reduce the range, lower the price. And everyone is very convinced this is a commodity business, loads of debt, loads of CapEx, one really concentrated customer, and that they're completely undifferentiated. And I guess I would just make a point, which is, in America, if you can pick any category, any retail category, and if you can run 1,000 stores...

in 50 different states with different preferences and have those stores well-lit, clean, stocked by friendly employees with the right stuff in stock, you will create a business that is worth well over $10 billion. And that sounds easy, but very few companies have been able to do it.

I think what may be underappreciated about CoreWeave is it's actually really hard to run these big training clusters. Everybody thinks it's easy, but to synchronize tens of thousands of GPUs where they're melting or cables are being unplugged and you lose a bunch of training data when it happens, I just think it's not...

It may not be the commodity that everyone thinks it is, and it may turn out that it's much harder to do than people think.

Yeah, the world was convinced AWS, well, a lot of the world was convinced AWS was a terrible business for the first few years. You know, when Microsoft announced they're going to transition to the cloud, it was very controversial because you go from a CapEx light model to a CapEx heavy model. So, you know, Wall Street consensus has been wrong before.

And I do think CoreWeave runs these big GPU clusters as well as anyone. And there aren't that many people on planet Earth who can run them well. And you know, it's kind of easy. You can just look at the NPV of their contracts and the asset value.

And so everyone, you know, who's kind of, you know, confidently putting out all the negatives here, like, I just think there are some offsetting positives that it's good for, you know, I think people to be aware of and consider. That's all.

Above and beyond whiz. I mean, CoreWeave bought Weights and Biases, which I think is a really interesting acquisition for them that kind of kind of further differentiates them and maybe decommodifies them a little bit. But there have been quite a few acquisitions by big companies of smaller companies in the last month.

And I think because none of the companies that have been bought have been public, maybe it's flying a little under the radar, but it's happening in, I think, an accelerating way. And I think this would be very good for the market and animal spirits and everything that we care about.

Nuclear is arguably just as environmentally friendly done right and carefully. And it is here now. And so I just... Yeah, I mean, that's... Yeah.

The answer is both, but it is like Elon has tweeted many times about the tiny fraction of... of deserted desert areas of America that need to be covered with panels. And then you put in batteries.

By the way, on that point, Jake, how it is interesting. Texas is the number one solar producer in the country. And it's not because everybody there is more environmentally conscious.

It's just it's easy to build stuff in Texas. You can build solar power plants.

I do think cryptocurrencies at some level fundamentally reduce the power of nation states. And that is something that is oft professed. by the true believers, but it is true. And so if you are on the left and you are a devout believer in the power of the state to do good things, I get why you would not like cryptocurrencies. At the same time, we're very early in crypto.

I do think I read with interest all of the posts that David Marcus and his peers at Libra made about what happened to them.

And I think you can argue it would have been really, really good for not just America, but the entire world. You know, there are a lot of immigrants in America who send remissions back to their home countries at extremely high... Predatory rates, yes. Yeah, predatory rates. It would have made that free. And that would have been amazing for a lot of really hardworking people all over the world.

It would have, you know, Visa and MasterCard, they do charge big fees. I mean, Visa and MasterCard do not charge big fees, but the credit card complex and aggregate. is a reasonably big fee. I mean, everybody, oh, it's just, you know, two, two and a half percent to make it perfect and safe. I think Facebook had a sound argument that they could have done that cheaper.

That would have been an efficiency game for America. And it really did bum me out, you know, to read some of the letters that they sent the way they killed Libra. If anyone doesn't know, maybe we could you could pull up David Marcus's post

They just, all these politicians sent letters to participants in financial markets saying, we don't know if they are doing anything wrong, but we think they probably are. And we're going to look at this very closely and we want to discourage you from participating.

It was really bad. I found it upsetting as an American.

Yeah, I mean, look, it's a it is a rational fear. I mean, controlling monetary supply, like at some level, you know, the greatest powers we give the state are a monopoly on violence to keep us safe. And control over the money supply has means of exchange, a unit of account. And those are great powers, particularly if you're America and you're a reserve currency.

The only thing I would just say to balance this out, and I do think people are very positive on Paul Atkins. I've never met him. But the reaction from people who know him, like Joe, and who I respect, has been very positive. It is important to remember, we have the best capital markets in the world.

you know, the U S equity and fixed income markets are the most trusted places on earth and we can always make them better, but just, It is very, you know, you want to be very vigilant about keeping them fair and keeping out things like inside information, which makes people feel comfortable doing business here.

You know, making, you know, having investors have confidence in a company's financial statements. And those financial markets are one reason America is such a great country.

I don't know. I understand 100% why they did it. I just think as an American, at a minimum, the way that they did it was to, you know, use one of Joe's phrases, dishonorable. You know, if you want to say that we're going to kill this, then just like, let's have a debate as a nation.

Yes. Rule of law. Rule of law is also another reason America is a great country. Yes.

Yeah, well, I would just say Texas is the Singapore of America, and it is putting pressure on the rest of America. And I don't say that just because I grew up in Texas. It's just a fact. I do think long-term Bitcoin, I do not think the BRICS will ever be able to replace the dollar. Just rule of law, even if it has occasionally been corrupted in America, is very, very powerful. Yes.

As opposed to rule by law. But I do think Bitcoin... will at some point be a serious threat to the US dollar. And that just is what it is.

I agree with all that. I will just say on AOC, I thought it was very interesting. I think if you're a Democrat, I think you're probably largely heartened by the kind of intellectual leaders of that party, their reaction to losing. You know, it's focused around, hey, we do need to deregulate. It is too hard.

Josh Shapiro has been tweeting every three days about how he's making it easier to do business in Pennsylvania. It used to take 20 days to get a hairdresser license. Now it takes an hour. All that is good. But I actually thought AOC, who is a very talented politician. She's an extraordinary communicator. She's an extraordinary communicator.

Her reaction was like, she posted this on Instagram, if you voted for Trump,

want to hear why i want to hear the things you listened to that convinced you like i don't say this out of anything other than genuine curiosity basically clearly i am missing something and i you know i want to listen and i just thought that was a very interesting reaction that is the proper reaction yeah absolutely yeah i want to point out one thing i did i was doing some research uh before the show and i found this sec speech

I think it's great. The reality is a hedge fund that runs with a lot of leverage probably is more risky. I don't know that it's more risky than a single security. like at the end of the day, since he wrote that letter, it was an incredible 15 year run for private equity.

And, you know, now all the big private equity firms are making a huge effort now when probably the business is more mature and maybe the return opportunities aren't what they were to appeal to main street America. Like it would have been cool if, you know, Blackstone or KKR or whoever, and, and,

07, 08, 09, 2010, you know, could have had their fund up on like the Fidelity marketplace for subscription. That probably would have been good for America. So I think his comments are well taken.

So I agree with what a lot of David said. And I do think if you were to allow ordinary Americans to buy private companies that are held to a lower standard of disclosure and reporting than public companies, like something would have to change. Like, you know, just, hey, if a private company wants to

like public companies are, you know, are held to certain standards for a reason, the numbers and financials.

Yeah. And there is a lot of fraud in venture. There are unethical people. There are.

Yeah. I mean, look at FTX. I mean,

Not that anyone did any diligence at FTX. Yeah. Which is an important part of this.

Well, I do think it is a little... That's why I talked about private equity. These are extremely sophisticated institutions who are always investing alongside their clients. And they're buying established businesses. They're putting leverage on them. But I think private equity is kind of a middle ground on a pooled basis.

I think you could argue, and maybe those funds would need to change their reporting and their disclosures to deal with kind of the average American. And by that, I mean strengthen it. But I think private equity... His comments were well taken. That's what I would say.

you know, be able to do this.

At some point, this does get too big. For a while, when it was smaller, you could support the debt.

Yeah. So what he is doing is issuing debt and buying Bitcoin with the premise that Bitcoin is always going to go up. And he has made eloquent arguments why that is the case. No trees grow to the sky. And I think the interest expense on his convertible notes is, is 75 million off the top of my head. And by the way, I am, I could care less about micro strategy. Like I'm not close to it.

Thank you, J. Cal. Happy to be here. Great to see everybody.

I'm not involved. Yeah. I don't know any hedge funds who own it. No horse in the race. Yeah. Yeah. I, I think a lot of hedge funds are short micro strategy, but I have no horse in the race, but the, but he's the underlying business that, you know, pays the interest expense on the debt and, only does $400 million a year in revenue and it's, you know, high gross margin revenue.

But I just, unless debt investors have absolute confidence in Bitcoin has collateral. And I don't think that's where fixed income markets are yet. it will get to a point where it is too big for the size of his company. And then, yeah, maybe he can over collateralize it and, you know, have $10 in Bitcoin for every dollar of debt.

But then like the magic money creation machine that, you know, I see discussed on X breaks down because that's like, you know, that's very, very different than what is being discussed today.

Gavin, how do you look at that market? Do you agree? I agree with everything Joe said. The only thing I would just add on China... What we are doing by restricting their access to advanced compute and advanced networking, if you have read or watched the three-body problem, America is unfolding a so-fond over China.

So if they execute unstated plans and there are some of the world's greatest execution machines involved, Elon generally does what he says he's going to do. This is going to be awesome for America, for markets, for the world. And the analogy I keep coming back to is Satya Nadella taking over as CEO of Microsoft. Microsoft was a monopoly, incredibly advantaged.

I have been really impressed with some of the Chinese models that have come out. And I think the risk to this strategy is necessity is the mother of invention. And despite this handicap, they're managing to stay just behind the leading edge of America, which is amazing. But, you know, Nvidia's Blackwell chip comes out next year. You're going to have new chips from AMD, new ASICs from Broadcom.

And I think at that point, it is not going to be possible for them to keep up anymore.

It is very aggressive foreign policy.

Clearly. You know, that could have lots of unforeseen consequences.

We have lots of rare earth here in America. We have everything in America. I think there's a project underway to restart rare earth production. If there were ever to be a conflict, all this stuff would go away.

Yeah, and I would just say this is, I think, a very important moment for AI today. you know, for this entire AI trade in public and private markets, you know, everybody I'm sure who watches your podcast is very aware of scaling loss.

And we have not had a scaling loss for training, or if you 10X the amount of compute used to train a model, you significantly improved the intelligence and capability of that model. And often they're these kind of emergent properties that emerge alongside that higher IQ. No one thought it was possible to make more than 25,000, maybe 30,000, 32,000, pick a number, NVIDIA hoppers coherent.

And what coherent means is in a training cluster that each GPU, to kind of simplify it, knows what every other GPU is thinking. So every GPU in that 30,000 cluster knows what the other 29,999 are thinking. And you need a lot of networking to make that happen.

InfiniBand, and I think even more importantly, NVLink, although a lot of... ethernet is, um, you know, never bet against the internet, never, never bet against ethernet. Um, like if you read the Lama 3.1 technical paper, you know, got a lot of people excited about skinny link ethernet, but, um,

Yeah. So... Picture a server, in the case of a GPU, it looks like maybe three pizza boxes stacked on top of each other, and it has eight GPUs together. And those eight GPUs are connected today with something called NVLink. You can think of the speed of communication on chip. is the fastest. Chip to memory, next fastest. Chip to chip within a server, next fastest.

And so you take those units of servers, which are connected, the GPUs are connected on the server with a technology called NVSwitch, and you stitch them together with either InfiniBand or... Ethernet into a giant cluster. And each GPU has to be connected to every other GPU and know what they're thinking. They need to be coherent. They need to kind of share memory.

For the compute to work, the GPUs need to work together for AI. And no one thought it was possible to connect more than 30,000 of these with today's technology. From public reports, Elon, as he so often does, focused deeply on this, thought about it from first principles, how long it should take, the way it should be done. And he came up with a very, very different way of designing a data center.

It had just been horribly mismanaged for years. All he had to do to start winning was stop doing really, really dumb things. And that's an incredible place to be. America, we're the greatest country. We've got oceans on two sides, peaceful neighbors, incredible natural resources, completely can produce our own food and energy. In many ways, most privileged country on Earth.

And he was able to make over 100,000 GPUs coherent. No one thought it was possible. If I was a last minute ad for this, but I would have said there were all these articles that were being published in the summer saying that no one believed he was going to be able to do it. It was hype. It was, you know, ridiculousness. And that was coming.

The reason the reporters felt comfortable writing those silly stories is because engineers at Meta and Google and other firms were saying, we can't do it. There's no way he can do it.

He did it. And I think the world really only believed it when Jensen did that podcast, I think with, wasn't it with Gerstner? It might've been with Gerstner. Yeah, I think it was with Gerstner and said, what Elon did was superhuman. No one else could have done it.

And I actually think you can argue that Elon doing that in a lot of ways kind of saved Nvidia from a tough six month period when Blackwell was delayed because everyone who was waiting for Blackwell and thought it was impossible to make a hundred thousand hoppers coherent rushed out and bought a lot of hoppers to try and do it themselves. Now we will see if someone else is able to do it.

It was really, really hard. No one else thought it was possible. And, and, As a result of that, Grok 3 is in trading now on this giant Colossus supercomputer, the biggest in the world, 100,000 GPUs. In Memphis. In Memphis.

With a lot of megapacks around it. And the city of Memphis is all in on supporting this. which is obviously smart for them. But you have not had a real test of scaling loss for training, arguably since GPT-4. And this will be the first test. And if scaling loss for training holds, Grok 3 should be a significant advance in the state of the art.

That is an immensely, you know, from like a Bayesian way to look at the world, that is like an immensely important data point. But if that card doesn't work, and I think it is going to work, I think ROC 3 is going to be really good. I should note that I am- What with consumers?

My firm is an investor in XAI.

I would just say we're already building models of models, you know, almost every application startup I'm aware of is chaining models. You know, you start with the cheap model, you check the cheap models work with a more expensive model, you know, lots of very clever things are being done. You know, every, every AI application company has what's called a router.

So they can, you know, swap out the underlying model if another one is better for the task at hand. Um, As far as what the wall is, there's been a big debate that we were hitting a wall on these scaling laws and that scaling laws were breaking down. And I just thought that was deeply silly because no one had built a cluster bigger than, you know, 32,000 H-100s. And nobody knew.

It was a ridiculous debate. And there were really smart people on both sides. But there's no new data. Grok 3 is the first new data point to support whether or not scaling laws are breaking or holding. Because no one else thought you could make 100,000 hoppers coherent. And I think based on public reports, they're going to 200,000 hoppers. And then the next tick is a million.

But sometimes with great privilege comes great stupidity. And California, to me, would be a leading example of that. In many ways, most privileged state in America. And it's printed away with bad policies.

It was reported they're going to be first in line for Blackwell. But Grok Street is a big card and will resolve this question of whether or not we're hitting a wall. The other question you raise, David, is very interesting. And by the way, we should note there is now a new axis of scaling. Some people call it test time compute. Some people call it infrared scaling.

And basically the way this works, you just think of these models as human. The more you speak to one of these models, the way you'd speak to your 17-year-old going off to take the SAT, the better it will do for you. As a human, if I ask you, David, what's 2 plus 2? Four flashes in your mind right away.

If I ask you to unify a grand unified theory of physics that accounts for both quantum mechanics and relativistic physics, you will think for a lot longer. 147. Yeah, nobody knows. We have been giving these models the same amount of time to think, no matter how complicated the question was.

What we've now learned is if you let them think for longer about more complex questions, test time compute, you can dramatically improve their IQ. So we're just at the beginning of this new scaling law. But I think the question you raise on ROI is very good. And I'm happy to address it.

Oh yeah, we have, even if scaling laws for training break, we have another decade of innovation ahead of us. Exactly.

And I do think one thing that everyone of all political stripes agrees on is there are too many regulations that result in far too many administrators, far too much complexity, and an inability to build things in America. So-

Oh, absolutely not.

The best information I have is the largest cluster Microsoft has after panicking. It's still smaller than XAI's cluster in Memphis. If you didn't believe it was possible, you weren't even working on it. Grok 3 should take the lead if scaling laws hold in January or February.

I thought it was... a really shocking statement from Mira Marotti that she resigned during a fundraise. That's the only way she can express disapproval of what is going on there and still probably get her money.

So I think, I think there's, there's a lot of reasons if scaling loss holds to be, to, to be optimistic about crock three. But I think, and then by the way, on the power question, and they are, you know, 23 and 24, it was just a panic to get GPUs and get them plugged in. Now we're, you know, trying to make them efficient and thoughtful, and to your point, re-architecting them.

They have a little more compute and a lot more memory, which really matters. So per kind of unitive effect of compute, they're a lot more power efficient.

It was used very effectively, as it should have been against the Harris administration, that they'd approved $42 billion for rural broadband, hadn't built anything, had approved $40 billion for EV chargers, whatever it was, hadn't done anything. And it's not that they didn't want to dig trenches and do broadband. They didn't want to make EV chargers. Their own regulations prevented them.

No, the H200s, probably not 2X, but a good increment. And the H100 was a great chip. So 50%, yeah. Yeah, Blackwell's just around the corner, and that's an entirely new architecture with an entirely new set of networking technologies.

Right now, you have a friend in your pocket who has an IQ of 115, 110 maybe, but has all of the world's knowledge accessible to it. And that's what makes it amazing. I think this will be like you have a friend in your pocket, and they sometimes make things up. Again, they're very human. And a lot of humans, when they don't know the answer, they dissemble. These AIs do it too.

So you will have a friend in your pocket with an IQ of maybe 130 that knows everything, has more up-to-date knowledge of the world, and is more grounded in factual accuracy. And it is interesting for any question involving real-time information, mostly sports and finance. You know, I always, you know, if there's a stock down 25%, ask every AI, why is the stock down 25%?

Generally, Grok is the one that knows. But I'm obviously biased.

Grok knows.

In the world today.

I find these debates also very funny. There have been articles written about multi-hundred billion dollar ROI questions. Those are very strange to me. Why? The biggest spenders on GPUs are public companies and they report financial results every quarter. And you can calculate a metric called return on invested capital.

And ROIC, ROI, has gone vertical since they ramped their CapEx on GPUs and actually just started to level out in this latest quarter. So the ROI on AI has been very positive thus far. Just a fact. It's a really good question Will it continue, particularly if it's going to cost $100 billion to train a model in two or three years, which I think is a realistic estimate.

Yeah.

Yeah. I mean, I personally have not had good experiences with Copilot, but I would say that, and I'm sure both of you have come across these. There are lots of companies that are just these thin wrappers over a foundation model. And they go from zero to 40 million instantaneously. Yeah. And they're profitable. And for their customers, they're replacing labor budgets. Yes.

I'm sure you guys are noticing this too, but startups today at a given size are employing fewer people than they would have three years ago. And just like, you know, it's funny. People were very skeptical.

So I just think... deregulation and simplifying regulations and the tax code is going to lead to an immense amount of growth, which is something that all Americans should be happy about.

Yeah, and that's the ROI on AI. And like, you know, I went to the first AWS reInvent conference and no big companies were using cloud computing. It was all startups. Startups always adopt technologies first. So outside of the ROI on AI that you're seeing in Google and Meta from...

you know, using this across their businesses, you were seeing real ROI on AI from startups, the same way they saw real ROI from cloud computing before anyone else.

But I don't think these companies are in a classic prisoner's dilemma. They all believe to varying degrees that whoever gets there first to artificial super intelligence is going to create tens or hundreds of trillions of dollars of value. And I think they may be right. And if they get there, And they think that if they lose the race, their company is at mortal risk.

So as long as one person is spending, I think they will all spend, even if the ROI decelerates. It is a classic prisoner's dilemma.

Well, first of all, it's a tragedy. And actually, I'm in New York as we record this, and it happened one block from where I am. And I mean, it is absolutely, it's a human tragedy. Taylor Lorenz was not the only person on social media who reacted that way, which I thought was deeply troubling.

I was completely unaware of it.

We're all very lucky on this podcast. And I can't imagine how I would feel if someone I loved had been denied medical care and died, and I felt like it was unnecessary and due to some corporation trying to make more money. But I just, I cannot believe, I'm sure I'd be outraged, I just can't believe I was deeply disturbed as a person by the number of people online celebrating this.

Yeah. I thought it was, it was awful. And I guess the last thing I would just say about that chart is I think it is a little misleading. I think that is initial denial.

you know, maybe the, you know, in other words, maybe the company that, you know, only denies 7%, that's a final deny. You know, they, that's a company that.

And if you're oppressed, the corollary to that is if you are considered oppressed, you can do no wrong.

And that is... Obviously, there are people who are oppressed on this planet, but they can still do wrong. They should still be held to a moral standard.

Two months ago it was 60% there, now it's 70 to 80.

I tried nine months ago to make a multiplayer.

Sorry, go ahead, Gavin. I tried nine months ago to make a, uh, multiplayer, um, app for Skyrim, um, which I was very excited with Skyrim. But I do think I failed. Um, maybe I should give it another go, but the, um, you learned, you didn't fail. I learned, I learned, I didn't fail. I learned. Yeah. I have a growth mindset, but we learned the limits.

But I think next year, the human language will be the dominant programming language. Totally. Totally.

Necessity is the mother of both technical and creative invention.

Why not just make them automatically sunsetting after five years if they're not renewed? Exactly. But make the process to renew them difficult and data-driven. Exactly. So if you fought for an allocation. The process of renewing them will take so much time that it will slow it down. Exactly.

It raises average IQ by two or three points. It really matters. But look, I think at the end of the day,

I mean, more nuclear, more better. Like, I mean, I agree with everything Joe said. And the only thing I would just add is it is the most environmentally friendly kind of energy source. In some ways, it's even more... Like, I think it is highly likely that in my lifetime, the world just runs on solar. Like, if you just...

know we all know compound interest is the greatest force on earth but if you just look at the rate at which photovoltaic cell efficiency is compounding battery efficiency is compounding and people make these balance of system arguments but it will it will never be as cheap as nuclear but it will likely approach coal and i think a lot of the world will run on solar but that's going to take 50 years

I think there's too much money in politics now.

I might feel very, very differently if I didn't agree so profoundly with the largest donor in this political cycle. I would say the reality is, as long as the money raised on each side is roughly equivalent, I don't think it really matters. That would be my take. You just want some equivalence. That's all.

Yeah, it's pretty wild, actually.

Like what's your sense of this? Yeah. Yeah. I want to get the G1, which is their humanoid robot.

Profoundly agree. I think year of the robots, inclusive of FSD. I think FSD works today and it's going to It's going to cross into mainstream adoption where I already notice, particularly if I'm taking an Uber late at night, I really, really prefer to have a Tesla. Just sometimes you get an Uber driver who's tired. And I just feel a lot safer if they have the FSD running.

And then obviously using it for yourself is amazing. But I think it's going to continue compounding at an accelerating rate. I think in a lot of ways, what I'd say broadly is, I think for a while, it's going to be big businesses are winners, big businesses that use AI thoughtfully.

And the reason is, is that with what is happening, what O3 showed us, OpenAI's new model, a combination of reasoning and test time compute, Like I think if you're a big business and you can pay a million dollars to let an AI think for six weeks about the most important question for your business, that's gonna be a profound advantage relative to small businesses that can't afford that.

And by the way, we are the other, like inference compute is gonna be the kind of derivative winner of that. We're gonna run out of GPUs, accelerators, compute, in 2025 the same way we did in 23.

Absolutely.

I do think that they might be very good for the world. But I do think they would be very bad for America if they replaced, if some constellation of stable coins became the new reserve currency. It is such an advantage to be able to borrow in your own currency and to give that up lightly. That means you control the real size of your debt at all times. oh, wow, we have a big, big debt.

I mean, this is a little bit like what we just did. Let's just let inflation run hot. And oh, wow, that debt's a lot smaller than it was. Effectively. So I think it would be a big mistake.

Jason, shouldn't you just go with the vibe of the podcast? Shouldn't you explain what it is? Oh, sorry.

Yeah, I think it's great.

I would take the other side on pro sports teams. It feels like that market is about to be institutionalized and funds are going to start buying pro sports teams that ultimately and ultimately capital markets have way more resources than kind of the wealthy individuals who've done it to date. So I just think it's the amount of money that can go after pro sports teams is about to

Looking great.

When you say interacting, what do you mean when you say interacting? Wait, hold on.

No, I just want to just come back to Chamath because for sure these private equity guys are DCFs. The NBA has been terribly managed. I think even LeBron said they have a big problem. know and i'd separate the nba is the worst managed nfl is probably the best managed but the one kind of counterpoint i would say on the tv rights is google bought sunday ticket and are extremely happy with it

Amazon and Netflix also both bought NFL games and are really happy with them. And so I just think those three companies, it doesn't really matter to them if pharma ads go away. In 18 months, you're going to be able to dynamically create an ad for each individual person and show it to them. And if you're a high value user that Google knows that you're about to book a vacation,

They'll dynamically generate whatever hotel destinations you want and show click here. So I just think sports, as long as they command the eyeballs that they do, and for sure, if the NBA doesn't fix it, the value of those franchises will start to decline.

But just the fact that all of the biggest tech companies are so happy with the sports kind of rights that they bought, they're going to buy them all. Every sport will go to one of those companies.

I agree. And they, you know, I mean, they all, all except for Apple, really. And Apple, they have, I think, an MLS deal, but they all kind of stuck their toe in the water. And I think they feel like, wow, the water's warm. Let's dive in. Okay. Yeah.

Government service providers. You do not want to have the United States government at any level as over 35% of your revenue.

Crazy thought. Crazy thought.

Atreides Management. Atreides Management.

Yeah. Cost plus is terrible. No, just cost plus is terrible. That's all right.

I only ski in the morning.

I would rather not talk about specific positions. I would just say I agree 100% with Chamath. All right, there you go. They're gonna lose their Chinese business, because they don't make competitive products anymore. If there's not massive protectionism, and we're seeing signs of that, they'll be caught between Tesla and the Chinese OEMs.

And I think the only way that this doesn't happen, Chamath's, the only risk to Chamath's prediction is just government intervention, because they're such big employers and often seen as national champions. But absent really significant government support, They're all in deep trouble.

Get a little work done. Need a healthy mix of apres, ski, and ski.

I think the biggest business deal is that they're going to be deals. I think you're going to see a tidal wave of M&A after four years of not being able to get anything done. I think there's an enormous amount of pent-up demand. So kind of point number one. Point number two, something will happen with Intel, and that will be big, and hopefully it's good for America.

And then I do think you will see a lot of these frontier AI labs that are independent be acquired. I thought your points were well taken, Jason, about everything that Google is bringing to bear and XAI owns their own compute.

I mean, you guys can choose whichever one of these three you like, but the ultimate AI winner will be the one with the lowest cost of infrastructure costs, the lowest cost of compute. And definitionally, you can't be the low-cost provider if you're renting your compute from someone else. Because of markups. Yeah, because there's a markup.

Well, I would say I'd make two observations. First, like I'm in the bottom percentile in terms of natural athleticism, but I do have many, many thousands of hours skiing. Great. Okay. And I would just leave it at that. Yes. Oh, okay.

If you're buying your compute from Azure or AWS or Google, you are at a disadvantage relative to their internal services over time.

only thing i would add and i think dovetails with what you said and and um friedberg's kind of year of the robot is autonomous drones there's a company called zipline which my firm is an investor in yeah but autonomous drones they really are the best way to deliver almost anything to suburban america and then i think over time it will be sorted out so that they could deliver in cities as well

So I just think that that might be a little bit of a wild card for some of these delivery services.

Apple intelligence is even worse than Copilot, which is saying something.

Do you guys think Tyson threw that fight? That it was part of the deal? Yes. Thousand percent. I so agree. I would love... I think somebody said Saudi Arabia was offering to host a remake or whatever, a rematch, a rematch of the fight where the winner would make 50 million and then we'll see what's what.

Yes. And the loser gets zero.

Well, first we should use Gemini deep research to just ask it to look at the total Pax Americana debt outstanding, apply the current market interest rate to it, and then look at that interest expense relative to GDP and generate that chart over time, that would be a cool chart. That would be interesting. I don't disagree with Chamath.

Anything is possible and there could for sure be a problem at a big bank. I don't know that I would say it's likely, but anything is possible. My most contrarian belief, so I think that America over the next, at some point over the next four years, will print at least one year of greater than 5% GDP growth, real GDP growth.

I think productivity is going to go vertical because of AI and deregulation. And I think there could be a world where this is the late 90s. And it doesn't sound like a big difference, you know, 5% or 6% versus 2% or 3%, but it is a massive difference. you know, at five or 6%, you know, the economy is doubling. You'll call it every 12 years, roughly. First 24 years at 3%.

I mean, it's just, it's a massive difference in terms of kind of the wealth of the country and individual Americans. As far as a specific prediction for 25, because I don't know when that will happen, I think you will see the frontier labs stop releasing their leading edge models to prevent knowledge distillation and their IP effectively being stolen.

Gavin, what do you think? I think it's right. I think there were a lot of people, not just centrists, but a lot of people who had been lifelong Democratic voters who voted for Trump. So I do think Trump won the centrists. Trump won everything. essentially.

You know, DeepSeek from China, it was really impressive, but it thinks it's GPT-4. So I just think you'll see the labs keep their best models in-house, distill them, and put small models out over time.

I totally agree. I think AI, you know, people are fond of saying that we're in a world of AGI or ASI, money will be meaningless. But for a short period of time, money will matter more than it's ever mattered before. Because the amount of money you can spend on AI on test time compute is going to give you a massive advantage, whether you're a company or an individual.

So I just think AI is going to really amplify inequality for some period of time. I hope I'm wrong. I hope AI leads to all sorts of opportunities being created for low-income people.

Socialism. Socialism is going to rise.

I think it's pretty interesting.

But just to build on what you said earlier, like wokeism and progressivism will decouple from socialism. I think wokeism and progressivism you know, we'll be on a declining trend, but socialism in terms of the government. Like policies, you guys are saying. Yeah, yeah.

Thanks for having me here, Jason, Chamath, and Dave. Atreides, we're a crossover firm. We invest publicly and privately in consumer and tech, and we go from Series A to mega cap.

I think the companies that make high bandwidth memory, going back to we're going to run out of compute, it's actually a pretty shocking stat. High bandwidth memory is a bigger part of NVIDIA's cogs on GPUs than Taiwan CME is. And today there's two companies that can make it, Hynix and Micron. We'll see if Samsung gets their act together.

But HBM memory, it's in NVIDIA GPUs, AMD GPUs, Amazon Terraniums. And particularly in a world of test time compute and inference being so important, HiBen with memory is arguably more important than it ever was. And it's been sold out for the last two years. So HiBen with memory would be my pick.

No, no. I was just thinking. I mean, if Chamath's prediction of a bank failure is true, you absolutely want to own CDS. I mean, you'll forget 100X. You'll do 1,000X, 10,000X.

Really cheap. I've had a, I would call it a no China guideline ever since the long top financial fraud more than 15 years ago. It was just amazing. It was, you know, a lot of great investors owned it. And, you know, they talked a great game and they had a Western auditor.

But then it turned out that the fraud was happening at the local post office where they were opening up the documents that the local auditors had signed off on, changing them and then sending different documents, you know, whatever, FedExing them. Yeah, so I just think it's a hard place, I think, if you're not Chinese, to make money. But I tend to agree with a lot of what David said.

I think Trump and Xi both want to deal. I think there's a deal to be done that leaves Putin out in the cold, as I referenced. And if that happens, Katie, bar the door. I mean, there are really, really high-quality companies that are – some of them are mid-single-digit multiples. Yeah.

Enterprise application software. This is basically just about... 2025 is going to be, I think, the year of agents, particularly in the second half. Agents just being... AI models that can take action on your behalf, that can do anything online that a human can do online.

And if the labs and the big cloud providers dominate agents, which seems likely going back to, you know, the low cost producer is going to win. Enterprise application software, I think, is going to be in a lot of pain. And some of these companies are talking a big game about agents. But at the end of the day, they don't have their own models. They don't own their own compute.

And I just don't see them being ultimate winners in the world of agents. You know, I could easily, I could be wrong because they do, you know, they have some customer data, maybe a little bit of a data moat. And they do, you know, have strong customer relationships. But, you know, most companies have relationships with AWS, Google, or Microsoft as well.

So I think enterprise application software will be in pain.

I would say Trump and centrism.

And just to add just one last thing is just, you know, fundamentally, these enterprise application software companies, the software industrial complex, and I agree with everything Chamath said,

know they're they're fundamentally based on making white collar human employees more efficient and what the ai companies are going to do is just say hey we're just going to replace that worker and it's just a fundamentally different mentality shift yeah and it goes to like there may be a lot of pain that accompanies this and you know david's points around the you know potential rise of of socialism

Jeez, this is scary. It was an obvious...

They've all started competitors.

But also 6 billion in losses too, David, against that 12 billion in revenue.

I think the 6 is for next year, but I don't know.

I think AI, and Nick, I sent you a chart. I don't know if you could flash it up, but I think AI is going to make more progress per quarter in 2025 than it did per year in 23 and 24. And the reason is just with 01 and 03, this is Arc AGI, it's designed, you know, we keep changing the goalposts for the Turing test and AGI, I'm sure we're going to change it again.

we're going to blow through this and i just think what has happened is we were scaling around round one on one axis which was pre-training and then we started scaling around inference time compute and it's very clear that we have now added a third axis of scaling performance and that is reasoning and what this is is these models the internet is composed of answers people giving answers and what the models really benefit from

is kind of the internal monologue of somebody getting to that answer. And this is, you know, in AI terms, they call it a reasoning trace. And it's one reason, you know, 18 months ago, everybody was like, oh, wow, the more code you train a model on, the better it does and all sorts of things that have seemingly nothing to do with code.

But all code, you see kind of the reasoning, the internal monologue, the thought process, the step by step. And so what's happening is you're using models to generate synthetic data that contains these reasoning traces. So you ask a model, you know, solve this problem, it has to be a problem that is functionally verifiable, that has an answer, or you know, the answer, we say show your work.

And we have to do that many, many different ways and times. And then you pick the best ones, and you kind of feed those back into the model, you apply some reinforcement learning to it. And so now you're scaling along three axes that are multiplicative with each other. And I just,

I think Putin. I think Putin is going to lose bigly. So if you are Xi Jinping, you know, Xi Jinping, Russia is a client state of China at this point. If you're Xi Jinping, what is happening is now a disaster for you because Europe is starting to rearm, which will only accelerate this year.

A guy on the Google team tweeted maybe five days ago, it's going to be a straight shot to ASI, artificial superintelligence. And I think that might be right. Ilya gave a talk at NeurIPS, Ilya Sutskever, one of the original pioneers of this field. And he said something that I thought was...

scary and he said these models that reason are inherently unpredictable so the best reasoning models in the world today are the ones that play games the kind of alpha go alpha zero style models and they are constantly making unpredictable moves that no human grand master ever could have come up with and now these models are going to be making similarly unpredictable leaps in all sorts of domains

which hopefully will be awesome, but it might not be.

That will allow America to take resources out of Europe and put them in Japan and South Korea and all over the Pacific. That makes it a lot harder for you to do what you most want to do, which is

Unquestioned, 1923, season two. I'm normally a science fiction kind of fantasy spy.

of guy when it comes to tv but 1883 and 1923 were the first tv shows that kind of hit me the way game of thrones did and i'm crazy excited for that to come out fantastic are you watching land man i'm enjoying land man i haven't watched it yet i'm excited to try it it's quite fun quite fun turns out this guy taylor sheridan he's i mean he yeah it's incredible he's in the zone

reunify china and taiwan or invade taiwan uh let's call it what it is and i just think xi is going to begin decoupling from putin if you're trump you want to show that you're independent that you're not um enthralled to putin in any way so i think trump is going to be a lot tougher on putin than people think and i think he's going to get a very very uh a deal that's very very bad

Yeah, I'm excited to hear Day of the Jackal, Sicario, yeah. Absolutely.

Yeah, it's reasonable.

Yeah, it is interesting. Obama deported, I think, more than 2 million people. It's like people have forgotten. But he was actually really, really tough on the border.

I mean, Google is. I would say they're a lot smaller, which makes it easier for them to grow faster. Yeah.

Well, it probably won't be Google because if they had the largest dollar gain, it would mean they were growing at insane rates. But I think Azure versus AWS will be interesting.

I think like most things, it'll take, like immigration, it'll take time to get going. Yeah. It just, it's going to take a little bit of time, but I think they will make progress. By the way, at some point, if I'm here on the all in again, I want to talk about UFOs and the drones. Let's talk about it now.

So I don't know. And I think it's very clear. If you look at all the statements from the New Jersey mayors and governors who've met with the police, the FBI, the Defense Department that they don't know. Now, you know, Trump said someone in the government knows what they are. He also said he was not going to go to Bedminster anytime soon, his resort there.

for Russia and Ukraine, and you've lost half a million people, and for what?

I just think over the last, I'd say eight years, every 18 months, there's a big story in the New York Times, the Washington Post, the New Yorker, the Atlantic, very credible media sources that with dozens of interviews with fighter pilots and commercial pilots talking about seeing things with advanced sensors that made no sense to them. And then there's kind of a taboo that went away.

There was an article in the New Yorker where they quoted from a bunch of people who were at, you know, kind of the skunkworks laboratories in the 1950s and said they absolutely saw extraterrestrial materials that have been recovered from a crash. That's one reason the US made such big leaps in material science.

And, you know, there was the Uemana thing where essentially a lot of, you know, astronomers said, hey, this is clearly a UFO of some sort, including the head of astronomy at Harvard. And then just the conspiracy theory I would have is if, if, if, if, if, These are actually UFOs and not government drones, either from the United States or China.

It seems like the most likely explanation in New Jersey is that some sort of a drill. And obviously, once this hysteria gets going, people misidentify commercial airplanes as drones. And why would UFOs have blinking green and red lights? But it is just interesting that there was a big concentration of these reports

as we were scaling into nuclear technology, 1945 to 1960, and then now that AI is getting going, which is the next kind of technological phase shift for humanity, there's another big kind of bolus of reports.

I mean, I think it depends how deeply it's classified. And I'm sure if the government doesn't want to give it up, they won't give it up. But, you know, there's all the, you know, there's all this.

Yeah. And, you know, they call it UAPs now to make it, you know, not flying saucers. But it just feels like something funny.

At least 20%. 20% from Chamath.

I would say 25%. I would take the over on the 20%.

I don't know. It is so strange the way there are, you know, in so many, I would say, cultures all over the world, pre, you know, we are closer in time to Cleopatra than she was to the time of the Great Pyramids. Like, that was a long time ago. But it's just weird that in so many cultures all over the world, there are these depictions of what looked like astronauts.

And then there's all these Renaissance era paintings that clearly show what we would call UFOs in the skies over cities. So I don't know. I just think it's- I'm with you.

2025.

But I think Trump said something very interesting about John Bolton. He said the guy was absolutely crazy, but it was awesome having him in the room when you were negotiating deals. Because people looked at this guy looking angry, red in the face, so excited to hit the nuclear button, so excited to go to war. And you ended up with much better deals.

So I think you're going to see probably a lot more bellicosity from the Trump administration than anyone expects. And that's just to get a good deal done. for, you know, between Russia and Ukraine. And then it's to get China to kind of decouple from Russia. So I think the in reality, I think you're right, David.

But I think there will be a lot of rhetoric that is at odds with what you're saying before you end up being right.