Stock market site in Times Square. I'm Ross Sorcin, along with Joe Kernan, and Becky is out on this Friday morning. Let's show you US equity futures. Still about three and a half hours before the open, but right now the Dow would open up about 56 points higher. You got the Nasdaq looking to open about 224 points higher. The S&P 500 would open up about 30 points higher. Much of the market story centering around the tech sector. The Nasdaq tumbling more than 1%, having its biggest one-day loss since mid-August. The index hit an all-time high earlier in the week. AI stocks dropping this after CNBC confirmed OpenAI told investors it saw $50 billion in annualized revenue at the end of September. Now that would be about $20 billion less than previously had been expected, though there is some back and forth about what is included in that number or not in that number and how that relates to the way Anthropic accounts for things. Of course, Anthropic's IPO scheduled, or at least expected, for some time in November. Take a look at Treasuries right now. 10-year sitting at 5.2 and the 2-year at 4.7.
That's a little, a little better, maybe somewhat responsible. We're going to talk a lot about this and some strange technical factors that have caused the tail to kind of wag the dog, supposedly, on the 10-year. We'll talk about that with our guest coming up in the next block. Check out the price of crude, that's moving lower. President Trump said the US would not attack Iran before the midterm elections. The president making those comments on Truth Social. He also said the administration is having productive discussions with Tehran. Oil prices spiked yesterday after attacks on Saudi Arabia by the Iran-backed Houthis. Missiles hit the main airport in Riyadh. And we're tracking a hurricane. It's like this is what was our... what was it for a while? Andrew, is it... I remember what, what, what...
Yes, probably 15 or 16 years. I have one... yeah, I have... is that, that's how you say it? Yeah, I think this, this might be one of... and so this... did you pronounce this hurricane yet? It, it... I have not even tried, but you know me and pronunciation. I spelled it out, you know, it's potato. The storm's a category 2. It has sustained winds of 105 mph. It's expected to make landfall later today near the Alabama-Florida line, just west of Panama City. Heavy rains and flooding could extend into Tennessee and North Carolina throughout the weekend. And we're going to get much more on the storm tracks in about 25 minutes. Been a really quiet hurricane season in the Atlantic because of a super El Niño.
Let's talk about this, this one, because this is a major, major sea change. I mean, just watch out, folks. This is the big story of the morning. SpaceX purchasing nationwide spectrum portfolio, pushing its Starlink service deeper into the US telecommunications market. Shares of Elon Musk's rocket company moving higher this morning, another 4% higher. The deal involves the acquisition of a spectrum portfolio from Grain Management, which specializes in digital infrastructure. Now, financial terms were not disclosed, but the Wall Street Journal says SpaceX agreed to pay Grain about $8 billion in cash. The FCC will need to approve the transaction. SpaceX says the acquisition will, this is the important part, pave the way for Starlink Mobile to become a major mobile carrier in the US, and that is driving down the shares of AT&T, also Verizon, also T-Mobile.
Perhaps most importantly, the band of spectrum, depending on your understanding of the way the technology works, can go through walls. And this has always been the central question about whether, if you're going to be running a network by satellite as opposed to terrestrially, whether you can really compete, because in a big city it's hard to do these things, even in suburbs. This could be a major, major game changer, and you're seeing the stocks of all these companies fall on the back of it. Science fiction-wise, they talk about when you can talk, you can contact anyone anywhere in the most remote places on the planet. And that, that, that's a potential of this at this point. And you're not near Newark, but when you do fly United, you immediately thought, wow, this is the other airlines.
Oh, you're talking about being on the, on the airplane. Sure. But how good Starlink is, amazing, is amazing. And I, that's my, my airline and that's my, my airport. But you know, you know what else I think, Andrew, wasn't Comcast one of the few bright spots in recent reports? Has been they've been expanding into wireless mobile, and that, that hit $20 and change yesterday, down from 65. Broad looking for whatever, but they're all getting affected by this.
Well, because for, for so long we thought of the telecom companies almost as like the railroads, right? They had the tracks and there was no way to get away from the tracks. These guys, meaning Elon, is going over the tracks. So the tracks don't have the same kind of value that they used to have, potentially, but it's not like, you know, like cord cutting where the ice cube is really big and you can sort of see it shrinking. This is like... like... it could be. It could be. This is like could be disintermediated in the blink of an eye theory. So I don't know if you, if you saw... Is that really possible though? Is this, is that really possible that that could... I think, look, all of these things always take longer than you think that they do, but when they happen, they happen fast. Elon, I don't know if last night... this will sound super crazy, but I see a path to SpaceX being worth orders of magnitude more than the current Earth economy.
Yeah, we've, we've talked about that before. Trying to figure out what it's in the, what, 30 or 40 trillion or something. Yeah. I mean, look, I... it must be more than that. 100... I don't know how he gets there, but if you can start to think about SpaceX as a sort of massive telecom company globally, that, that's a big business, obviously. If you start to then start to think about all the other things. I, I, I was thinking about who loses, not, not really who wins as much, but yeah, that you can imagine the market cap would be... well, you're just significant... just take up all the market caps of all the other telecom companies, assuming that you think the other telecom companies aren't going to compete. I mean, the other thing is they could try to compete on price. I mean, it'd be a race to the bottom is what'll happen, but that's always great.
Let's talk Nvidia. It is actually not Nvidia, but a company backed by Nvidia, an Australian AI data center firm, an operator called Firmus, has withdrawn its planned IPO. The company telling CNBC it had, it pulled the plug due to market volatility. Firmus had reportedly planned to raise $5 billion in August. The company announced a $2 billion funding round backed by Nvidia, Coatue Management, Blackstone, and Jane Street, and it boosted the valuation to $10.5 billion. But see what happens now. Let's talk about some other potential deal news. I don't know what we really think of this. Shares of Chipotle heating up after the Financial Times reported that Starbucks has explored an acquisition of that burrito chain. Stock gaining more than 6% in yesterday's trading. This morning, the shares are still moving higher, and Starbucks is flat. Of course, Starbucks's CEO Brian Niccol led Chipotle before moving to the coffee giant. A Starbucks spokesperson, though, telling CNBC it doesn't comment on rumors or speculation.
I don't know. It's not like, it's not like Smucker's and a, in a peanut butter company. I, I don't really... You mean, you mean peanut butter and jelly? I don't understand. Burritos and coffee. I don't... I guess people, breakfast burritos. I guess... I don't... No, the question is whether you think you could rationalize both businesses, right? It's not about putting them together. The food synergy, obviously. I think it's, I think it's if you believe that Brian is a great operator and he was a great operator at Chipotle, and I think he's fixed Starbucks. You know where he cut his teeth. Well, by the way, he left Taco Bell. Yeah, but he left Chipotle and then, by the way, the management of it wasn't so great. This sort of one-man band. He turned Chipotle into Taco Bell, which is... I asked him to let's have some more, you know, it's so healthy and, and it's got so much cilantro. Give me Fritos-based or Dorito-based tacos with, you know, lots of sour.
You think that? Do you think Chipotle is healthy these days? I think that it's much healthier than Taco Bell. And I don't, because of that, I don't like it. Okay. I've always said that, that there's two kinds of people in the world: Chipotle people, Taco Bell people. And you're a Taco Bell person. Yeah. Most people wouldn't admit that. I probably should. Coca-Cola. You, Coca-Cola or Pepsi? Yeah. I'm a, as you know, I'm a Diet Coke guy. You like Diet Coke? And you don't, not Coke Zero? Nope. That's weird. I know, because I grew up drinking Diet Coke. We only had Diet Coke in our house. The aftertaste, the aftertaste is so ingrained that you don't even know it has an aftertaste.
Coca-Cola, maybe looking to sell its... see, it's Costa Rica, but is it Costa Coffee? I'm... Yeah, Costa Coffee. Costa, not Costa. A lot of people mess that up and it, it makes a difference because it's Costa Rica. It's Costa Coffee brand. That's according to Semaphore. The company bought Costa in 2018 for just over $5 billion. FT reporting earlier. They've been trying to sell this thing for a long time, though. Is that true? Yeah. Coke scraps plans to sell Costa after offer from private equity firms came in below expectation. All right. What else? I just had other things to talk to you about, but it's Friday. We, I'll figure it out. We got three more hours to go. We do. We do. But it, it, it's never enough time coming up for us. Might be plenty for other people. Stock futures moving higher. We're going to open up our market discussion with the CIO and head of global investments at T. Rowe Price. And we'll, if you want to do some homework, read this piece on B1 of the, of the Journal about the bond market. Maybe there's hope. Maybe there's hope. But don't read the front page, Andrew, about the def, about the deficit this year. I got nothing good to say. And nothing good to say except that revenue is up 3%, spending was up 6%. Squawk returns in just a moment.
Bit of a rebound so far, although it's petered out a little in the Dow even already. Joining us now, Sebastien Page, co-head of global investments, chief investment officer at T. Rowe Price, where he helps oversee more than $1.8 trillion in assets. Sebastien, you're a good guest for a Friday because you, you basically feel pretty good about at least valuation in, in the stock market. You, you, you note that a lot of people are talking about bubble territory from AI, but you think some of it is justified.
Yes, Joe. You know, you had Jeremy Grantham on your show a few weeks ago. He said it's the most expensive market in American history. And there are measures that show that. For example, the CAPE ratio shows we're in the 99th percentile of historical valuations. But Joe, valuations are in the eye of the beholder. If you adjust for the fact that the S&P 500 is now essentially a technology index with twice the margins it used to have 20 years ago, much higher sustainable earnings growth, earnings growth, and calculate a price-earnings-to-growth ratio, plug in the five-year expected earnings growth. Now, maybe that's optimistic, but you don't get the most expensive market in American history. You actually get a price-earnings-to-growth ratio that's near a 30-year low. So valuations are in the eye of the beholder. In our asset allocation committee, Joe, we're slightly overweight stocks and we're optimistic on the AI trade for the next 6 to 18 months.
And I mean, not only that, but you, you like bonds, too. And, and if you manage it correctly, but, and, and maybe it's not government debt. You can get 8% in, in certain areas of the, of the, the corporate market. And, and you don't, I think, believe that that is necessarily due to something bad. There, there's some good things that, that have rates higher. And I, I guess you're alluding to growth, growth in the economy, but you, it's almost you have both. You could have bonds and stocks both pretty attractive at this point.
We like credit. We have an overweight to high yield. You mentioned 8% yield, that's very attractive by historical standards. Yes, spreads are pretty tight, but this is an index, again, like those long-term charts, can be quite misleading. The high yield universe has 60% double-B bonds. It used to have 30, 40% 20 years ago. So you can't really compare spreads to historical levels. So we like it. We don't expect a recession, and we manage, especially in bonds, the portfolio quite actively to avoid default risk. But we were actually underweight, still are underweight Treasuries. So within bonds, there are different asset classes and they have different attractiveness in our minds. We've actually closed part of our underweight, about half of it, to Treasuries this week because, you know, rates are starting to get pretty high and they could go further, they could go higher. But the interesting thing about bonds is, you know, when rates go up, people say, oh, that's bad for bonds. Yeah, it's bad because the price of the bond goes down, but if you were underweight going into this, it creates an opportunity to buy because the expected return goes up. So in the long run, rising rates, Joe, are good for bonds. So you end up with a portfolio, in our case for our asset allocation committee, that's slightly overweight stocks, has an overweight credit, still has an underweight to Treasuries, but we've reduced it by half and we're looking to go to neutral if rates keep going higher. So that's how we're looking at the next 6, 12, 18 months, Joe.
Sebastien, a piece in the Journal, and it's a thesis that there have been technical factors that have caused the 10-year to go up in yield, and that some of those factors now are beginning to wane, and that the selloff is starting to wane as well. And it's, it's very, it's kind of weird. It's a tail-wagging-the-dog thing. The confluence of market quirks caused the 10-year to be sold. The surging mortgage rates and bond futures, the way that, that players reacted to that was by selling the 10-year and the 30-year, and supposedly that's going to start reversing. I don't, I don't know whether you can put stock in that, but we've been wondering whether 5.30 is the high end of the range on the 10-year, or whether it's the low end of a new range of 5.30 to 6.5 or something. Do you think 5.30 is the high end of a range?
It could be approaching the high end of the range. I think the technical factors you describe create some day-to-day bounce, but the trend has been higher. Joe, there's a famous old book by Agatha Christie, "Murder on the Orient Express," and the punchline is you have 12 suspects for the murder, but they're all guilty. So when people talk about why the 10-year is going up, we talk about, well, actually growth is pretty good, it's picking up. Inflation could be it. Oh no, it's deficits. Oh no, it's the term premium, this technical term. Oh no, it's because Japanese yields are up and other yields are up. Or it's just, you know, this question of the deficit. And I think all these factors go into it, and it's actually harder than it seems to parse them out. But my conclusion is, if they're all pushing, if all these factors are pushing in the right direction, and my number one suspect is inflation and oil prices, that's been highly correlated with the level of markets. So you can talk about technicals and all these things, but you see oil go down, oil, oil goes down, then the 10-year goes down, and they're very, very correlated. But it's a group of suspects, and like "Murder on the Orient Express," they're all kind of guilty right now, and the trend remains towards the higher range. Very hard to call the peak. We're managing this risk by buying some bonds and leaving room to buy more.
Who was guilty in, in Murder? I think it was Colonel Mustard, was it? No, wait a minute. See, I'm getting... So the clue, the clue is different in... Okay. Yeah, Colonel Mustard was, was not... Murder on the... I, I, never mind. We're going to go down a rabbit hole there, Sebastien. Thank you. We will... Thank you. Have a good weekend if you can, managing all that, that begins with a T. $1.8 trillion in assets. Hope you sleep okay. Thanks. See you.
Okay, coming up, Meta's feud with TikTok escalating. We got details ahead, and as we head to a break, take a look at early NASDAQ movers this morning. We're coming right back with a lot more on this Friday right here on Squawk.
Welcome back to Squawk Box. Meta banning TikTok ads on its social media platforms in the US and a handful of other countries. This is now the latest escalation in what has become a long-running rivalry between those companies, as TikTok has emerged as one of Meta's most formidable competitors for users, creators, and of course, ad dollars. In a statement, Meta says, quote, "We don't have to run ads from a competitor whose goal is to pull people off our apps." An interesting thought, and I don't know if you put this in the category of you own the stadium, you get to do what you want in the stadium. What gets into other people say there's a free speech thing. You should be able to buy an ad anywhere you want, right? Got a thought? No, I don't. I, I, you know, there's certain things, Andrew, I'd refer to you. On people sit around watching TikTok. They just sit around... that they do. They just doom scroll down the TikTok. The way they doom scroll down the Instagram. The way you doom scroll down the X. Boom. Boomer. The guilty. You know, you... I watch Rifleman. So I got, I have, I have issues. I have issues on MTV, whatever the hell it is. Coming up, we're waiting on quarterly results from the... Yep. Did you like... Did you ever watch Rifleman? No. Conners. Huh? He was the coolest. He was a basketball star at Seton Hall. That's... I don't even know what you're talking about. Coming up, we're waiting on quarterly... How about Bonanza? You know, I know what Bonanza is. Yes. God. Well, you... Little House on the Prairie. Yes. Okay. Little House on the Prairie. I know. Little House on the Prairie. I know. Oh, man. See, the weirdest thing. We got results from Delta Airlines. We're going to have the numbers as soon as they're released. And the new NBA season about to tip off. The league holding a game in China. Commissioner Adam Silver will join us live from Macau in the next half hour. As we head to break, here's a look at yesterday's S&P 500 winners and losers.
All right, Delta just out with results. Let's get to Phil LeBeau with the numbers. Hey, Phil.
Hey, Joe. This is a miss on the top and the bottom line for Delta in the third quarter. The company earning $1.72 a share, 3 cents shy of estimates at $1.75. Revenue coming in shy of expectations at $17.59 billion. The street was expecting $17.67 billion. Now, the metrics of the numbers within the numbers in the quarter, and jet fuel is the story of the quarter. Not just for Delta, but for all airlines. In the third quarter, Delta spent more than $4.1 billion on jet fuel, up 62% year-over-year. Cost per seat mile, excluding fuel, up 7.3% year-over-year. And then there's revenue per seat mile. It was up 15%. Remember, Delta and all the airlines, they are raising airfares to offset the impact of higher jet fuels. That's why the revenue numbers are stronger. Domestic revenue up 16% compared to the third quarter last year. International revenue up 12%. Main cabin revenue up 12%. Premium revenue, people continue to pay up for the premium experience, up 18% compared to the third quarter of 2025. Operating margin of 9.4%. It was 11.1% in the third quarter of last year. Free cash flow of $463 million. That is well below where it was in the third quarter of last year. Now for the part that people are going to be focused on today, the guidance from Delta in the fourth quarter. The company expects to earn between $1.15 and $1.65 a share. Right now the street is at $1.42, so just 2 cents above the midpoint between $1.15 and $1.65. Then you have the Delta revenue guidance expected to be up 20% compared to the street expecting an increase of 17%. And for the full year, Delta is now expecting earnings per share of between $5.10 and $5.60. Right now, the street is at $5.46. We'll talk about jet fuel, we'll talk about higher airfares, and we'll talk about what's happening when it comes to Wi-Fi in the sky with Ed Bastian. That's coming up in just a few minutes, guys. Don't want to miss what Ed has to say. Rough time for the airlines, but we knew this going into it, guys. It's all about the jet fuel right now.
Kind of interesting to see revenues soaring and earnings per share below. Phil, and you look at prices, they're... it's like, wait a second, I was just looking at some for an upcoming trip, they're, they're definitely... and Joe, that's not slowing down demand. That is not slowing down demand, and Ed Bastian will talk about that when we, when we meet with him. Demand remains extremely strong, especially for premium. And I said you can ask Ed or one of us. And, but I've, I've seen criticism that, that he didn't do Starlink when he had the chance, and that United is... that people love it at, at Starlink. What, what is the strategy? I, I was talking to Sen... it's Amazon or something. Where... we'll talk about, they're going with Amazon Kuiper. We'll talk about that with Ed in a little bit. And of course, we all saw the comments from Elon Musk or the post that he put up where he said, you know, Ed will lose his job over this. Not sure that's going to happen anytime soon. The man is running the most profitable airline in the country. They've had a lot of success under Ed Bastian, but that didn't stop Elon Musk from saying what he thinks about Delta's decision to not go with Starlink.
Right. Okay. All right. Phil, maybe we'll talk about the big merger. Who, who's merging with who? And remember, it was, I guess United post... they approached both. Didn't Scott approach American and Delta at some point? Yep. Threw out that idea. Yeah, you can go anywhere. Yeah. No, I can't imagine that. That's all we need. And I don't know what fares would cost at that point. Anyway. All right. Thanks, Phil. Okay. Coming up, we're going to go inside the highly competitive AI race for superintelligence with author and co-host of Machine Gods podcast. Kevin Roose is going to be with us. And a reminder, you get the best of Squawk Box in our daily podcast. Follow Squawk Pod.
Welcome back to Squawk Box this morning. Our next guest is the author of a new book. It is called "The AGI Chronicles: The Inside Story of the Race to Create an Artificial Superintelligence."
Joining us right now is Kevin Roose. He's a former technology columnist at the New York Times. He's also the co-host of a newly new podcast called Machine Gods. And Kevin and I used to work together. So, I'm very proud to have you on the show. Congratulations on the book. I was going to say though, with the title, whether you've thought about changing the title to SGI.
Has there been any conversation with the publisher about that?
No. I think we're pretty comfortable with the title.
You're good on the title. Super artificial intelligence. I hope I don't...
No, it just goes super general intelligence. Isn't that what...
No, no, but I think that instead of either artificial or super, I think it is artificial, but it is super. So my thing is super artificial intelligence or artificial super intelligence.
You spent the last two years working on this book, but it's also just been like a rolling story that's moved the whole time. We're all obviously fascinated by where this is all going. Obviously earlier this week, President Trump started talking about changing the name of all of this. But just yesterday, there's been a whole sort of real crazy debate about the valuation of OpenAI, valuation of Anthropic. And so I'm just, before we even get into the book itself, I'm curious whether you think that sort of the dreams and hopes of this whole ecosystem, from an economic perspective, are going to make sense when this is all over.
I think they will. I mean, look, the stories we've heard about how they account for revenue. This is an accounting story, but the fact is these companies are some of the fastest growing companies in the history of capitalism. They are going to go public. They are going to be some of the largest companies in the world. And what's so crazy is like I wanted to sort of tell the story of how we got here because this has happened incredibly fast. Like 10 years ago, Dario Amodei was a VP at OpenAI. Sam Altman was sort of a little-known entrepreneur running this company, and like now these guys are the most important people. Their companies are sort of propping up the entire economy.
Okay. So, when you read this book, and I don't know if you're trying to lead the reader somewhere on this, who do you ultimately...
trusts, and maybe make it hard, who do you trust more?
I don't trust any of these individuals fully with something this powerful. And what's more is they don't trust each other, right? These people all used to work together.
Well, they all hate each other.
They all hate each other, by the way the book sounds. They all hate each other.
They all do. There's a lot of stories about why they hate each other. But these are all people who sort of bet on this crazy idea of AGI, artificial general intelligence, more than 10 years ago. And at the time, this would get you laughed out of the room. They were considered sort of fringe weirdos, and they had this conviction that it was possible to build a machine that was as smart as a human, or maybe much smarter than a human. And what we're seeing now, this race, is just like we are in the last lap of this thing that started more than 10 years ago.
It's the last lap and not the first lap.
Yeah. Well, it's the last lap of this race because the race that they've been running is toward this thing, AGI, the system that is as smart as the human, that can do anything the human brain can. And I think we're just about there.
So, it will happen. It's people still, and I guess it's people have arguments about whether AGI really equates to consciousness, like human consciousness. And I don't... being as smart as a human is different than being sentient.
That's right. I don't really take a view on whether AI could become sentient. But to me, it almost doesn't matter. It's sort of beside the point. It's like, what can it do? What good can it do? What harm can it do?
Happened. It, you know, occurs. Well, I read about the singularity because I want to live forever. You know what I mean? So does he. But I still pray just in case. But quantum leaps since the beginning of time. It used to take a million years for something, then 100,000 years, then 5,000, then you look at the industrial... we're down to things taking like five years now for quantum leaps. And with this super intelligence, artificial intelligence, we're going to be into months probably for things that would have taken decades before.
Yeah. The good version of this is that we get a hundred years worth of progress, scientific and other fields, in a month or a year or a couple of years.
So I got a question for you. I was watching an interview that Mustafa, from now at Microsoft, had given, and he was talking about Anthropic's constitution and the idea that the way the constitution is sort of set up, these are the guardrails that are built for this, that it encourages pushback, that pushback is part of what it's encouraging. And whether you think ultimately that kind of, whatever you think that constitution looks like, poses a danger, and how different it is, if it is at all, to whatever you think OpenAI's done or whatever you think's happening at what Demis has done at Google, etc.
Yeah, I mean, I interviewed more than 150 people for this book, and they ranged from being very concerned with how the morals and ethics of these systems are being built and shaped, and some of these companies have in-house philosophers working for them that are working on training the models, and then some of them just view this as a tool. It's something that we will control because we've controlled every other technology in our history. And so even within the group of insiders, there's a lot of disagreement on this point.
And what do you think the right answer is, though?
I think it doesn't matter whether it's conscious or not. My...
conscious, but the idea of that, I mean, people always say, 'Oh, we can just pull the plug.' And I say, 'Sure, we can pull the plug.' But if they're off, you know, if they've sent a hundred agents off that we can't see and they're talking in their own language, going off to do something bad, you may pull the plug maybe a month too late.
I think this alignment problem, what they call the alignment problem of making sure that these systems do what we ask them to do, is the most important problem in technology and maybe the most important problem in the world right now. It's still unsolved. That's why you see these guys calling for a slowdown. Not because they're sort of, you know, they want to shut the whole thing down, but just because they want to sort of buy more time for the researchers to figure out how to control it.
Do you think it's all hypocritical? I mean, David Sacks would say this is all some kind of effort at creating a sort of regulatory arbitrage where they're going to control the whole business.
I think this is very consistent with what these people have been saying for, again, more than a decade. You can go back to 2016 and see Dario writing papers about how rogue AI could hurt us. You can find Sam Altman writing blog posts about how he thinks superhuman machine intelligence is the greatest threat to humanity. So these people might be wrong...
years and find science fiction guys saying Isaac Asimov, the three rules of, you know, don't, never harm, whatever it was. I mean, it's amazing that it's been science fiction for so long and now it's not, now it's just science.
And these people all grew up reading that science fiction and deciding that they wanted to really be the ones to do it.
Okay. So I got two more, two final questions for you. One is we had a mathematician on yesterday, the mathematician who has been claiming that OpenAI was stealing his prompts to be able to answer some of these famed questions in the past couple of weeks. But we don't know how it happened. OpenAI says it didn't happen that way. Do you believe that these model makers are still stealing information constantly? This sort of the Alex Karp argument. I think some of these model makers would say we used to steal, we've stopped stealing. What do you think is really happening?
Look, I don't know about the specifics of that mathematician's claims, but like OpenAI just released hundreds of purported solutions to these open problems in mathematics. I don't think they were stealing all of that from human mathematicians. These models are becoming superhuman at math to the point that many people think there will never be another major proof in mathematics that doesn't involve an AI system. So the progress is real, whether you think these claims are real.
Okay. And then my final question is really an economics related question, which is one of the things we've seen just in the past 12 weeks. There's a chart going around which you've seen, I think, which is that, you know, 12 weeks ago, 80% of token usage was the frontier models that people were paying for, the full-on paid models. 20% was open weight. And in the past 12 weeks, supposedly it's completely switched to 80% open weight, 20% frontier model. That seems to me, if that transition's actually happened that quickly, would have a huge economic impact on the values and valuations of these companies, which of course then prop up so much of the rest of all the other hyperscalers that have stakes in these businesses.
Yeah, I think the frontier is solidly closed source still. The open weights models are much cheaper to run, but they're much less capable. And I think an open question is whether if you stop this practice of distillation, these companies from China and other places sort of siphoning off the intelligence of the leading closed models, whether that gap would widen, whether this is kind of an artificially narrow gap. And I don't know the answer to that, but that's a big question.
Okay. The book is called The AGI Chronicles. Kevin Roose, thank you.
Thank you so much. Thanks for having me.
at the NBA playing a preseason game in China. Today, our Contessa Brewer joins us now from Macau with Commissioner Adam Silver. Hey, Contessa.
Hi there, Joe. Nice to be with you. This is the Texas Showdown in Macau. You've got the Rockets versus the Mavericks. This is the second year in Macau. What do you expect?
I expect what we're already seeing is this huge crowd here. I don't know if your camera is picking it up, but...
Look, look at this, Russy. Show these are the people who've shown up more than an hour before the game even starts, right? And it's essentially been a full week of activities here at Macau. Of course, Patrick Dumont, who's the governor of Dallas Mavericks, is also the CEO of Sands, which seemingly has the bulk of the properties here in Macau. I think I heard there's more five-star hotels here than anywhere else in the world. So combination of clinics and fan events and technology summit that you were part of today. So it's been a great week for the NBA here.
What's really incredible too is to see the young players out here. They were doing these clinics with a lot of the kids playing basketball here in China. It was remarkable to see Cooper Flagg, for instance. But what's the league get out of expansion in China and giving these fans a chance to see basketball in person?
First of all, there's hundreds of millions of NBA fans here in China. We're the number one sport right now in China. It's an opportunity for a limited number of fans to experience it in person. Tremendous social media presence from this. But then there's also the media presence of the games themselves with Tencent as our partner here, CCTV. So, these games will be widely seen in China, also time zone appropriate. I mean, people are used to watching games in the morning here because obviously they're on the evening in the United States, but it's an opportunity for a huge audience, a huge fan base to experience the NBA firsthand.
You've been very clear about your intentions to grow internationally. It's been reported that negotiations with the EuroLeague hit a rough patch. Where do things stand right now?
