AI株の急落、AnthropicはRSIに近づくのか、予測市場vs世論調査

AI stocks get slammed as the market digests the midterm results. Anthropics models are approaching RSI. Maybe, maybe not. We'll talk about it. And also, are prediction markets accurate enough that we should take them seriously in elections? Or are they just piggybacking off of the pollsters models? That's coming up on a Big Technology Podcast Friday edition right after this. This episode is brought to you by Cribble. Quick question. Is your IT budget growing 30% a year? Because your data is. And now AI agents are piling on, running investigations around the clock and running up three meters at once. Storage, queries, and tokens. Cribble, the AI platform for telemetry, was built for exactly this. Own all your data, query it wherever it lives, and let Cribble route every AI request to the right model at the right price. The best part? AI inference for workloads on Cribble is included. Models will change. Agents will change. Your data is the asset that lasts. Own it with Cribble. Visit CRIBL.io.

How do you price intelligence as a service? How do you charge for delivering end-to-end work, not just access? Those questions remain front and center for AI founders and operators. And getting to the right fit business model, whether that's selling usage credits, outcomes, seats, or a hybrid combination, requires constant iteration. Chargebee is the billing and monetization infrastructure that powers just that for ambitious, fast-scaling AI businesses. Visit chargebee.com slash Alex to learn more. That's C-H-A-R-G-E-B-E-E dot com slash Alex. Welcome to Big Technology Podcast Friday edition where we break down the news in our traditional cool-headed and nuanced format. Well, in a year of many, many weeks with a tidal wave of news, certainly we've had it this week. The United States seems to have almost in unison voted against data centers, sending AI stocks basically in free fall until a little bit of a bounce on Friday and throwing Anthropics IPO's success into question.

We'll discuss that. We'll also talk about the latest advances that Anthropics is making on its models, with some saying that it's getting close to recursive self-improvement. And finally, we will discuss the role of prediction markets in the election. There's a big debate. Are they more accurate than pollsters or are they piggybacking off of pollsters' work? Ranjan and I will get into it. Ranjan, good to see you. As always on Friday, how are you? I'm good. Where are you in the world, Alex? That is not your normal Brooklyn studio. So for listeners, I am sitting in a very fancy chair. For viewers, you can see it. I am in London right now. Actually, I'll tell you a fun thing. I just filmed an in-person episode with M.G. Siegler, who showed off his iPhone Duo. We talked about whether that phone is going to work or not. So that will come on the feed next week. Forget the Duo. Let's just talk about that chair. That is the most London British chair I've ever seen. Actually, we dragged it right out of Buckingham Palace.

They might be upset. It looks like it. Oh, it looks like it. I have to tell you a funny story. I was in the airport on my way initially to Austin, then I flew to London. And I'm at TSA, and I'm the next person in line. And all of a sudden, three guys come in, and they're escorted by a TSA agent, and they cut me online. And it's two security guys, and the other person is Prince Harry. Really? Yes. You've made your London trip. You're really living it up. It's been royal. It's been royal. It's been royal. He plays commercial, which is interesting. But, okay. You've already left the U.S. One election result, and you're just fleeing. I'm out. No, I'll be back. I'll be back. But, okay, you might think that seeing Prince Harry at the airport, who, by the way, is very friendly. He was, like, giving me this, like, sheepish, like, Mom, sorry. I got to get ahead of you because otherwise I'd be, you know, mobbed for selfies. So I was like, that was cool. You might think that was the highlight of the trip. You might think seeing the iPhone duo was the highlight of the trip. No, it was seeing M.G. Siegler with none other than the B frontier around his risk.

No way. He had it. And he has it. We talked about it. And that discussion is coming up next week. So I don't want to give away too much. But let's just say it's been an action-packed week. And that isn't the only stuff going on in this world because you may or may not have noticed that not only did the Democrats sweep to power in the United States this week on Tuesday, the midterm election happened. Elections happened, and the Democrats took both the House and the Senate convincingly. But 38 jurisdictions either ban or restricted data centers, which we spoke about a little bit with Petty Schleifer on Wednesday. Not only that, the somewhat pro-data center governor of Texas lost his race to the very anti-data center governor-elect coming from the Democrat Party in Texas. So basically, we had seen all the negative polling around AI. Many people, anywhere from like 23 to somewhere in the 60s, a percent of people, depending on which poll you believe, said AI was crucial for their vote.

So we've actually seen this go from basically a technology many people hadn't heard of a couple of years ago to a defining technology that people have rallied against in the election. And obviously, we're seeing the fallout in the market. What was your initial reaction when you saw what was going on specifically with this AI as an issue? And where do you think it leads? We're going to talk through what's going on in the stock market, but let's go high level first. I mean, I think out of all the surprises from last Tuesday, people being vehemently against AI and data centers was probably the least for me. I mean, we've seen the numbers. I know it was pulling worse than ICE. I hope it was not pulling worse than ICE this. I think maybe they might have been on that same poll six months ago. Not looking good for any of the three. No, no, no, no. We all know. I mean, we've heard this. We've seen it. It's been very clear that the public is especially the data center debate,

and the AI industry has done itself no favors in terms of how it's positioned itself. I think the really interesting part, and we're certainly seeing that in the markets, is what does that actually mean very concretely for the anthropic IPO, for the hyperscalers? Like this is going to add a lot more unpredictability around what their actual economics look like over the next two to three quarters. And I think that is exactly what triggered the market. And I think we're going to be seeing a lot more of that volatility. Well, you know what's interesting? It's been very difficult to sort of put your finger on what's happening. So when the election results were coming in Tuesday and Wednesday, Wednesday, actually, in the wake of it, the tech stocks went up and the market went up. Some people said it was because there was a belief that finally the voters were going to restrain the tech giants from spending so much. And that was a good thing. Then they digested it another day. And on Thursday, they're like, oh, God, wait, they kind of need these data centers to continue this AI story.

And so the market went down 4%. And then Friday, there's a bounce back. We're not at the end of the trading day, but it does look like a bounce back. And I think what's interesting is, and I'm curious to get your take on this, one of the defining features of the AI trade is that it is somewhat inscrutable. It's amazing how many smart people really don't seem to have an idea of what's going on. You rarely see the market do this up and down and up again, And usually maybe there's maybe a fall and then a bounce back the next day. But the fact that it was up, down, up, and no one seems to have a handle on where this is going is an interesting feature of this trade. And that has trickle-down effects across everything that we watch. What do you think? I'll disagree there. I actually think it's pretty clear what's happening. I know NASDAQ's up 1.6% as we record. after being down over 6% in the biggest fall since COVID, like March 2020.

So, I mean, we're always going to get this little bit of a bounce back, but, I mean, it's set, it's clear. Like the data center build out and the absolute scale that was required to make any of these numbers make sense almost certainly will not happen. So to me, I mean, I think I saw like what you said was kind of funny around the idea of like it'll actually restrict spending. And that's good for stocks because it's adding a layer of discipline to hyperscalers that they've not shown on their own. You could even argue maybe Anthropic now out of its $518 billion in commits on compute might be able to get out of them because if data centers actually aren't able to be built, then maybe that'll be more attractive. I mean, that's like 4D chess type of logic in terms of any of this. But I think the main thing to me, like not only have voters spoken and Ben Thompson, I like he said that before the constraints were around first to compute and then like actual physical capacity.

But now political capacity is the major constraint. And I think we're going to see that going forward. So I think there is no way the assumed build-out of data centers and compute capacity can continue as at least it was forecasted to. So then it just becomes, how does that change each individual story? All right, I'm going to challenge you on that because, you know, I'm trying to figure out how real the restrictions are going to be. And I keep thinking back to the initial tech clash where Congress went after tech and the FTC went after tech and the DOJ went after tech. And you even saw meaningful wins from the regulatory agencies and cases they took against these companies. And where are we? Has anything changed as a result of this? I would say very little. And so I am always tempted to double check my initial reaction on the headline to what's actually going to happen five years down the road.

And you know what? Like you're already starting to see some of the stuff come in here that's leading you to believe that maybe this is also another blip. and that the votes and the legislators and the governors won't get in the way of the AI build-out the way that we might have anticipated. Here's two things that I want to bring up. Number one, there have been lots of 38 jurisdictions or more that have voted for restrictions or bans on data centers. We agree with that. But that does not limit these tech companies from either, A, expanding where they already are because the groundwork is already in place, or B, finding new areas that would be maybe potentially eager to welcome the growth. Right. You're going to see competition among the states, even among towns and cities within the states to get these data centers in, because, you know, no matter how ugly they are and no matter what they represent in terms of AI job loss, the trade unions want them there because they do mean at least short term jobs.

Right. Then you think about the governors and even governor elect of Texas, who ran so strongly against data centers, is now saying she is effectively not anti data center. She just wants them to play by their rules. So what we're seeing is not the end of the data center build out, but it's maybe the beginning of a negotiation. I'll agree that I think it is the beginning of a negotiation It's going to be an ongoing negotiation But again, I think like I keep thinking about this not in terms of You know, is it going to happen or is it not going to happen? All that matters to me and I think all that matters to the market Is it going to happen the way it was assumed to have to happen To make a number of valuation stories come true? That's it Like, and it won't. It can't every if Gina Hinojosa is saying, actually, and I think she had said that if the groundwork has already been laid, then maybe those and their schedule for completion into 2027, that those can go ahead.

And there's going to be all that kind of nuance around this, which is good. I think like I actually think that's a very good thing. And I think we're going to maybe we do end up in kind of a world of some kind of debate and rationality around this. But that is not what the story requires for everyone from Anthropic, the hyperscalers, the oracles of the world. That's not what will work for them. Yeah, and it does go back to some of the developments or some of the volatility, I would say, that you've seen from these companies and basically from the whole trade. And to me, I think part of this, and just to go back to my comment about the volatility here, is that so much of this is narrative-based. It's a story about potential, and it's less of a story of what's happening today. So let me at least give you my theory of the case for why I think we're seeing the volatility, and then you can let me know whether you think that there's merit there or not.

So I just wrote a couple of bullet points here. So here are some givens. I think AI is a technology with potential and a great story, but that story is not commensurate to its current earnings. So you have trillions of dollars in public company traded and trillions of dollars in public company valuations or stock traded on the expectations of effectively two companies with a hundred billion dollars in combined revenue So the AI labs potential was so great that they ended up leading to subsidized demand for AI chips and that demand went through the roof But to live out that story or give it a shot of working, they need the data centers to be built. And all this is basically hinged on these small but growing rapid companies. and the market is used to some sort of certainty around it. But again, new companies, new technology, and small size that are getting bigger is never going to be a normal trade

until it starts to settle out and it hasn't settled out yet. So any hiccup and everything goes haywire because of the small size of the business, the real estate of technology, all the hope, and data centers in particular are the linchpin here. Once they go, the story evaporates. This is exactly my argument. I appreciate this. You just laid it out even more eloquently in a structured way than I did. I mean, this is it. It's exactly. It's a narrow set of companies that the entire trade is built around. And, again, we've talked about this even on Anthropics specifically, and we're going to get more into the current state of the IPO, but everything assumes perfect execution. That's on their own internal businesses. That's on growth. That's on revenue. That's on unit economics. But that is also on the data center story to be the infrastructure for this gigantic story and all this promise.

So everything depends on perfect execution, and we're seeing even this little hiccup can actually completely derail that story. I'm a big hiccup. Yeah, yeah, yeah. Yeah, right. But this is why I'm going back and forth on it, and I haven't fully made up my mind on it yet, is to just go back to, is it really the end? Like, you know, I understand. I understand why the market will see a sign like that, and it wants to buy in, but I don't know if it's really the end or not because of all the different times we've seen this negative thing happen and it not actually manifest the way people are. So I'll push back on, in the past, the whole, like, idea around regulation and the tech clash was very limited to the actual singular companies themselves and like can the FTC go after Meta can and again we've seen I think was it New Mexico going after Meta for 25 billion dollars like we've seen more and more things happen again under

Lena Khan and the FTC there's certainly like actual action taken against large big tech but by the same token, these companies have gotten significantly more powerful than they ever were before. This is a little different to me because this is the regulation is or the government involvement becomes so much more small and targeted in a locality and whether they're going to actually allow a data center to be built. And that's going to happen on a one by one basis across the country. So it's not these massive, like centralized federal and central company issues of whether Meta's lawyers can beat the FTC's lawyers. This is going to happen throughout the country, and I think that is what makes it far more unpredictable and actually, in an interesting way, makes it more likely that something can actually happen to these companies. Now, here's some, so as I've talked through this, let's talk through some more evidence that I think would agree with your point.

These are major planning projects, right? To be able to build a data center, you need to go through land approval. You need to buy the land. You need to find the land. You need to hire the company that's going to go build it for you if you're working with subcontractors. And then, remember, we're dealing with power issues, so you have to figure that out. You have to figure out who the electricians are going to be, who the tradesmen are going to be to be able to put this together. So, you know, as much as you could say and as much as I've brought up this idea that you can shift the data center from, let's say, Ohio to Wyoming, and maybe you can. You also run into the fact that you're going to have to do this from step one. And so, you know, the question is, you know, how many years do you end up delayed because of it? Now, let me bring something up to you, Rajan. Some people have said that because of this, we're about to see the peak of AI CapEx. I just don't see that being true. I just think it might be drawn out across more years than it is, than we've seen it today.

And it may, you know what the funny thing is? This might be the thing that paces the frontier, you know? Well, no, but at the end of the day, that's what I, that's why I'm not over. This does not make me more or less bearish or bullish on AI in general as a technology. I mean, as we speak, our recording almost stopped because I've clawed in the background editing a separate video that I'm working on. And I just got a notification from Riverside saying it's constraining my compute on my MacBook. And I can only imagine the amount of compute that that is consuming. And like those kind of token heavy use cases are going to continue to explode. And I strongly believe that. But it's not going to happen in a straight line over the next two to three years in a way that a very small number of companies will actually be the beneficiaries in a way that has been predicted so far. And I think that that's going to be the – that is sinking in and the market is bouncing back a bit today.

but these things never go in a straight line down. Even the trade war and liberation day was like two days down, one day up, two days down. So I think today and the last two or three days is not going to be fully reflective. This is going to take at least a few weeks to start to get a sense of how much has it set in, but I think it has set in. I definitely do. You know what's interesting? So if you have less, remember, they have been trying to expand like crazy. I mean, when I was with Sundar, for instance, he said, I wish I had double the compute that I have today. And that's the case with all these companies. They've been trying to expand because they're trying to do two or three things at the same time. And now they'll have tradeoffs among those. So if you have the compute that you need, what do you do? You serve the current product with inference. Right. So what you're doing with Claude, that is compute intensive. That is coming from one bucket of compute. While you're doing that, you have to already train the next models.

Right. So that's another bucket of compute. And as you do both of those, you also have to do the proper test to make sure those models are safe. So you have product user inference, next generation research or training, and then the security thing. And what happens? I'm just going to throw this out at you. Actually, this is a very interesting tradeoff. if those data centers that were anticipated can't be built, what happens? Where do you make the compromise? Do you make the compromise on security? Do you make it on inference so you can make less money and you have mad users and Anthropical will risk them going to open AI? Or on the third bucket, do you not get the chance to train the next models? And if you can't train the next models, remember this is a story of potential. So the potential of RSI or AGI or super intelligence, you compromise on that. Where do you compromise? It's not clear to me that there's an easy bucket to sacrifice. No, there's definitely no easy bucket because, as you said, I think from a pure operational standpoint, the training of the next frontier model would be where I would sacrifice and just kind of build out.

And as I'm on team product, I think that always makes more sense that there's still so much opportunity room. But the investor story is not that. The investor story is the potential for the next frontier model and the next frontier model after that. So if you're solely trying to please investors and then you start falling behind on the benchmarks, that's – it's a tough decision. Maybe – you know what? Maybe here's my optimistic look at this. The same way innovation on the actual efficiency of compute side coming out of China and other areas where compute was more constrained was kind of the way things were approached. Maybe we start to see that in America right now. And this election actually brings efficient computing innovation. And that's my platform. That's what I'm running on two years from now. I missed this one. John John, 2020. Efficient computing innovation. Well, that is what's in there.

That is what Sundar said. He said the more – if you're a little bit constrained, you have to be creative and not just brute force it with compute. Yes. So that can be our optimistic look at this, which is still not good for a lot of the economics of how the trades that have been built up. That's the irony of it. You know who would hate that? Jensen would absolutely hate that. Oh, Jensen, yeah. He's like doing the math. Well, they're going to be more efficient, less GPUs. Yeah. Jevons Paradox. You could Jevons Paradox. But you can't Jevons Paradox if you can't build more data centers because they're banned. Then you got a real – oh, yeah, that's true. That's true. Then I think that – so the one thing we lose in all this is Jevons Paradox. It turns out Jevons Paradox was not a rule. It was just a thing. By the way, just to go one last bit about it. You referenced it in the beginning, and I think it's really important to sort of talk through. we have companies like Anthropic

on the hook for all this compute spend, right? If the person you're supposed to spend the compute with is unable to build the project, you're out. I mean, I'm not a lawyer but you no longer have to spend the money because if you want to buy 10 popsicles in the freezer breaks and you show up and they're like, you owe me 10 bucks and it's a soup, you're not obligated to buy those popsicles anymore. It's a soup. I want in like a Delaware district court, I want to hear that argument made by a fancy lawyer from a New York City law firm. I think that – Your Honor, we thought we were buying popsicles from Nebius, but as Exhibit B will show you, it was a bowl of hot sugar water. It was a bowl of hot sugar water. I think – I mean that's the kind of like irony of this. Maybe this ends up actually helping Anthropic in the near term. Maybe that becomes part of the IPO story because everyone who is a little bit concerned about those commitments now, potentially there's a way out.

Even though I've always argued that they're never really legally binding, they're always kind of, you know, wink, wink, nudge, nudge. At least how I found this playing out. You're sending the lawyer to squirm out of them, those commitments. You were ready. You were just waiting for your rules. I was ready. I didn't need a popsicle soup. I just needed – I wasn't hungry anymore, and that would get me out enough. That's right. Turns out those popsicles aren't as nutritious as you thought, which maybe you should have thought before you made those commitments to buy them, but whatever. We won't relitigate that contract. Let's actually talk about what this means for the Anthropoc IPO, because the reports are now that – this is from the Wall Street Journal – major investors are skeptical. Anthropic can achieve a 1.5 trillion valuation. It kind of goes back to my theory of the case that I showed that I laid out before. The only reason Anthropic was considered for a 2 trillion,

1.5 trillion dollar valuation was because of the story. And now look, it had a lot of facts to back up the story. But again, like if some of the spokes are getting knocked out, then what do you do? So what do you think is going to happen? Oh man. Oh man. Okay. I, I just have to go off a little bit here that, I don't know if you did you see today in the Wall Street Journal that there was a leak and a person familiar with the matter leaks that now Anthropic for October is showing an 80 billion dollar run rate. You cannot, you should not leak financials this close to an IPO. That never happens. That is not supposed to happen. That is not how any of this works. and still after all of this, these targeted leaks that are trying to kind of reinforce a narrative that totally make every audited financial in this entire process meaningless and a joke continues. I'm sorry, this one just killed me.

I thought, you know what, your targeted financial leaks leading up to the S1 and then the S1 amendment, good. But very smart communication. But still doing that after all those meetings and after all that and still kind of like obfuscating any possible chance of us having an understanding of what Anthropik's actual story is and what their financial story is. This is killing me. I hope. I don't know. I was the entire breakdown of all that is good in the U.S. financial markets and the norms that have been established for decades upon decades just getting shattered. Well, let me ask you this. Would you leak those numbers if you thought the IPO was imminent? Or would you leak those numbers because you thought, given the current environment, you were postponing the IPO anyway and you wanted to get the narrative out? because this is an interesting headline from CNBC.

Anthropic IPO announcement odds drop on Kalshi following midterms in AI sell-off. In the last week, odds that Anthropic would announce an IPO before December fell from nearly 90% on Kalshi to 57%, which is a massive move. So what do you think Do you think that there is a chance that Anthropic pushes this thing off And if so what would that implications be for the rest of the AI trade Because again we talked about how there are these two companies that are basically holding this market up And Anthropic is one of them. And if that IPO postpones, look out. I think that they have to. I agree. For the sake of everyone and for the sake of the industry, they kind of have to. But also for themselves. I mean, we had gone back and forth on Twitter a little bit where you were talking about how, like, you know, you had your good Q3, your record Q2, a little bit of a nice amendment on Q3 without the full picture. But, like, everything, every headwind is kicking in right now.

So to me, that normally means you shouldn't go out because in reality, like, that is actually asking retail investors to hold the bag. And it's like, let's let everyone cash out and have others actually own that risk and hold that risk while we enrich ourselves. I think I still think that's not a great way to look at it. But I think I mean, think about it like the initial two trillion dollars was all off of that Q2 and kind of that momentum and those growth rates. Already we had talk and it was going down to one and a half trillion. Now there's reporting yesterday it could be between one point two and one point three. I do think if it falls below a trillion in terms of what they're seeing appetite around, do you think you do pause? Do you stop the process? Because in reality, like this is all momentum. Anthropic had a historic eight month period and is still a very, very successful business.

But like this story will never be the same again. It won't. So I think you go out. I think you still go out. I mean, as you're saying, and as I said yesterday on X, and you came and fought me on it, you're going to go IPO with this unbelievable growth surge, the token maxing era, a bit of a slowdown in Q3, but not terrible, and a big October. So however much we might hate the way that it was. Yeah, a supposed big October. I think if you remember, if you pause, what you're doing is the growth rate will inevitably slow from this initial S1 flip because the growth has been extraordinary. And law of large numbers, you can't keep that up forever. And you also have, and I think this is potentially underrated, you have OpenAI, which has made this big enterprise push, which is doing well in enterprise. And which I think most likely has more compute than Anthropic does.

because remember over the past year, we've talked about how OpenAI bet much more on compute and took this riskier, what Dario would say, YOLO path. And in an era where the cost to build data centers are higher, the political obstacles to build data centers are higher, OpenAI is potentially set up to go on a nice little run here. And if you're Anthropic, do you really want to wait until that run is over? or your mid-run to then say, okay, we're going to now go with our delayed IPO. I don't think Anthropoc has any choice but to go out now. I agree. I agree. I don't think they have a choice. It's still crazy to me because, again, IPOs and buying a stock is supposed to be forward-looking, not backward-looking. The incredible growth only is interesting if you're supposed to extrapolate out that incredible growth. If the idea is, holy shit, you had an incredible run of growth and that is going to be declining, so we need to get out to the market and have others buy our stock, does not make sense to me.

I'm not saying it's not going to happen. The markets have gotten pretty weird for a while now, but I agree. From a purely self-interested, just what matters to every individual at that company, they need to go out and fast. And I also agree. Remember, Anthropic was nowhere relative to open AI just two years ago. Things moved so fast and changed so fast that the entire narrative could shift for them very quickly. So SpaceX is at $1.8 trillion. I think Anthropic can pull off $1.5 trillion. Even though it's like slightly different businesses, if SpaceX hasn't been, you know, I mean, slammed harder than it, then, well, it's trading underneath its IPO price, actually. And also, well, that's right. It's at $133 now. SpaceX has one thing that is very different than Anthropic.

Elon Musk, and the greatest capital raiser of all time, and that is his gift. And he can kind of like will the retail investor community better than anyone else, and Anthropic does not have that. So I still think – I don't think that's the right corollary there. Let's just do our bottom line on the Anthropic IPO. Do you think it goes out – it's supposed to go out week of November 16th? That's what people are saying. Do you think that happens? And if so, what do you think the valuation is? It's so tough. Look, I mean, I think we're going to see a lot more volatility over the next three to four weeks, definitely. I think people, as we said, we don't know already. Just that uncertainty, the first 10, 15 minutes of us kind of trying to understand where things are right now and realizing we have no clue is how investors everywhere are going to feel right now. And if the market keeps tanking, that makes the decision to go out really, really difficult. And I don't know. I still think even, yeah, even like anything above a trillion, I think they should be very, very relieved and happy with.

But also remember, when we keep throwing these numbers out, that assumes there are buyers on the other end. If the market is tanking, especially the retail investor community, the willingness to jump into a rich valuation just because a stock was hyped, that has worked for the last number of years. That's going to be the first thing to go. I think it goes out, I'll call it November 19th, and I'm going to say $1.3 trillion. And it'll be a good outcome for this company. and, of course, it's just the beginning. Well, not the beginning. It's the middle of the story, right? It's not the end. So, okay. Wait, that's not Thanksgiving. Oh, no, it's a week before Thanksgiving, all right? A week before Thanksgiving, yeah. I mean, I've just canceled a trip to the Bay Area that week in anticipation for the thing happening. That's why, Alex. Let's see if I'm right or not. I'm going to try to manifest it. Manifest it. All right. Do it for Alex.

One last thing before we... Do it for Alex and his travel schedule. He's going to be in New York, right by the stock exchange, ready to go. I don't want to have to bring this seat anywhere else, okay? It's a big chair. You need to get a chair like that back at home. That thing, it's regal. He said royal, regal. It brings a different feel. Why don't you describe it? I think you need to describe it for audio. It's a kind of reddish tan. again looks like it very well could have been nicked as they say from buckingham palace um there's kind of those two uh two wings on the side that stick a chair you would smoke a cigar on smoke a pipe on that you would uh sign a contract colonizing parts of india back in the 1800s as Lord Mountbatten. That is definitely a chair that did some colonialism at some point to me. Well, I can tell you, I will be doing none of these things.

No cigars or colonialism coming from Alex today. Not tonight, at least. I have other things to do. All right. Before we go to break, I just want to get your perspective on whether you think that we talk a lot about the comms problem that the AI industry has. And this is obviously All these problems right now are potentially a manifestation of that problem or downstream of the problem. Do you think that the AI industry can turn this around? And it's very important to ask because we already have some news from OpenAI saying that it will avoid NDAs when it negotiates data center deals. It's going to say that residents will not pay for electricity to power those projects. It's going to say it's going to turn these things. It said it's going to turn them down during peak electricity hours. And I wonder if we'll see an about face because up until now, you know, though the AI industry has said that everything's above board, clearly they know that there's work to be done because they're coming out with these reforms. And, you know, you look at places like Louisiana, there was this New York Times story from July that I think we talked about where like data centers have started just appearing in local communities.

that they're being, you know, they've been placed under secret code names and they just show up. And that may or may not, you know, well, that's clearly not sustainable. So do you think there's a way for them to turn it around and if so, how? What are we calling the AI industry here? Is that Google meta, open AI, anthropic? Are we calling it, like, what are you referring to as AI? And NVIDIA as well. Anyone with a stake in AI data centers being built. How about that? Okay. So it would include the NeoClouds. But we know who really is, you know, who drives the narrative here. It's the OpenAI, it's the Anthropics, it's Jensen, it's Mark Zuckerberg. I think there's an opportunity for some number of politicians, some number of kind of like untarnished voices and faces and people to actually start to kind of assert themselves as the face of the AI industry.

Again, we have joked in the past like Jensen could be the good guy of AI. Like again, Dario and Sam and Jensen and Ilana will never be the good guy in the face of the American people for the AI industry. I think maybe there is room for both politicians and other leaders to kind of step up and fill that void and gap. But I don't think for how brutal the effort of trying to actually not care about what anyone thought over the last couple of years, I don't think you're just going to walk back from that that easily. Do you think they suddenly friendly Sam touring America, going to farms and stuff like that, and doing one of those tours they used to do back in the 90s, do you think he might do that? It's hard. I mean, because data centers are ugly, ultimately, right? You could put a mural on them.

They don't grace the landscape the same way farmland does. So it's going to be tough. Wait, did you hear there's a, I think it was in France somewhere, like a beautiful data center that they tried building and like tried to like, and there's still a lot of community backlash on it. You know, it's happened in different places. I think, okay, you know, speaking of manifesting, I think the AI industry should follow this hot take that I've had for a while. And it doesn't seem to be picking up any steam. No pun intended, because what I'm about to say involves that. I think that the AI industry should turn the middle of these data centers into spots and have all the heat that the chips generate used to power hot dubs, steam rooms, saunas, and maybe they could put a cold plunge in there. And I think given the trends that we're seeing right now, the wellness trends that we're seeing,

if I'm OpenAI and I say, listen up, Ohio, I'm bringing a data center to your community. It's going to be somewhat ugly, but I'm giving anybody within a 25-mile radius access to chat, GP, spa. That doesn't work. The GPT spa, then I think that will improve the popularity. All right. So one thing I will note, this French data center, I just looked it up. It was TVPN had shared it, but it was a prank and not real. So we'll already retract that one about anything in France and beautiful data centers. But in terms of Alex's beautiful data centers, spa combination, it is interesting. Like, you know, don't go to Iceland. Go next door and get into – maybe they carve out and build these kind of lagoons that are getting heated. But also, as we're saying it, like being that close to that mass generator of power and saying that that'll heat up the water that you're going to put your body in feels a little off to the point that even your amazingly named GBD Spa, which obviously would draw in consumers from every walk of life.

I think I'm starting to feel a little more ick around this. I get it, but listen, it's not nuclear, okay? It's not like you have this radioactive waste. It's just heat coming in because these chips run hot. So I don't think that this is like you're going to be in the byproduct of some industrial process. There's natural heat. There's got to be a way to funnel it in, too. too, I promise you, if they put a data center spa down the block from me, I'd be there all the time, especially if it was free. And you know what? There's something that's severely lacking in society today which is common places for people to get together and bond socially Loneliness epidemic This could solve loneliness and we need to be more well And this could solve all those problems You just solved the data center problem the loneliness epidemic wellness and longevity in health. All I'm saying is try it once. Listen, try it once.

If your community still votes to put a moratorium on your data center, then I'll apologize. I'm in. You're right. I'm in. I'm sold. I like it. Let's take a break because, you know, even if there might be some financial up and down to the technological process, progress continues unabated for now. We'll talk about that and then we'll also discuss the role of prediction markets in a selection. We'll be back right after this. You know those pieces in your closet that you find yourself wearing on repeat? A sweater that goes with everything, the perfect pair of pants or the tea you somehow always reach for first. Those are the pieces that make getting dressed easy. And that's what Quince does best. Quince focuses on high-quality wardrobe staples made with premium materials like 100% Mongolian cashmere, organic cotton, and merino wool. The styles are timeless, versatile, and designed to become the dependable pieces you reach for day after day. Personally, I'm loving my Mongolian cashmere crewneck sweater from Quince. It's actually, unbelievably, making me look forward to winter. Find your next fall favorites at Quince. Download the Quince app for exclusive offers or go to quince.com slash bigtech.

Get free shipping on your order and 365-day returns. Now available in Canada and the UK, too. That's Q-U-I-N-C-E dot com slash big tech. And when Quince asks you where you heard about them, let them know it was Big Technology Podcast. It's the best way to support the show. This episode is brought to you by Intelligence AI. Is your investment in AI coding tools helping your team ship better software? Intelligence connects your AI spending to code quality and engineering outcomes, from individual sessions and pull requests to entire projects. You can see where AI saves time, where it creates rework, and how to improve. And its model router selects the right model for each task based on quality, speed, and cost. See how Intelligence AI helps enterprises optimize AI-powered delivery with connected visibility into cost, quality, and outcomes at intelligence.ai. And we're back here on Big Technology Podcast Friday Edition with Ronjan Roy of Margins. All right, Ronjan, I don't know how deep we can get into this, But it's very interesting. Meter, right, which is the the organization that monitors autonomous AI, did an analysis on Anthropics R&D.

Let me read it to you so you can react here. This is from the information. Anthropics R&D speed may have doubled due to AI researchers find. AI use within Anthropics may have doubled the company's pace of AI development. A new report from AI research nonprofit Meter says the rate that Anthropics models are improving as of this October is roughly twice the rate at which they improved between March 2024 and October 2025. Meter estimated around 80% of the acceleration in Anthropics research pace is due to internal use of AI within Anthropics. is, of course, after Anthropic has given Meter employee-like access, though Anthropic did restrict a lot of what Meter could publish because of what it called proprietary information, and Meter complained about that. Are we not fully taking seriously how fast this technology is getting better? This is a little bit crazy. I don't know. I'm sorry. Like, I know I've gotten – it's just every action taken in every one of these cycles

just makes me more skeptical and cynical, even though I know this technology is incredibly powerful and I'm sure the unreleased models can be dangerous. Even in this example, again, so much information was redacted and there was a backlash against, from one of Anthropix's co-founders who was trying to kind of like laud their work that was being done and how they're getting ahead of it. And again, as he said, Sam McCandless, he was like, we've redacted a lot And I accept this makes it harder for others to check meters work. But we determined that sharing exactly what is driving the speed up isn't the right thing to do. So they go out saying there is this significant acceleration that we're entering scary, uncharted territory once again, coincidentally, right, as they're trying to IPO into a very dangerous market. But then we will not tell you exactly what any of this is or how it works or what it means because – and again, using security as the blanket to actually try to like cover this, it's still – I don't know.

I just have an honest, open conversation. This is truly terrifying. And I just want to note – and I loved when Alex Stamos was on a couple of weeks ago. Well, I didn't love the fact that I left that conversation more terrified than I've ever been. But the dangers and the risks are bad actors using AI, not agents going rogue. And I still – all these media reports, all these anthropic security kind of like reporting, it's always around agents going rogue and painting this picture of that rather than like how does North Korea not get access to these things. All right. I'll just have one thing to say here, which is if I were to push back on you, so I'll just give it a shot. I would say that these companies have been saying that the releases are going to accelerate year after year. I remember being at the Anthropic Dev Day, Quote with Claude last year, which is crazy.

That was just last year. And they said the releases are going to start to come faster and the technology is going to get better faster. they kept their word. The releases have come faster, and technology has gotten better much more quickly. And so Mead are going in there and saying that they've almost doubled the rate of progress, to me, is not a stretch. And, you know, as they've been able to make these leaps, we've seen, like, new products become available or possible because of it. Like I'm thinking about Dodds or Muse, right, which was not possible last year. So I am curious what we'll see after this batch of improvement shows up. And I mean towards believing. Well, no. Okay. Here's the thing again. I don't not believe it. I was actually thinking about this the other day that like part of me finds it ridiculous because when like Dario and whoever else was saying the threats to humanity back with like the GPT-1 days,

they clearly saw where things were going and were right. We're completely right about where things are going and how this technology would evolve. To me, still the gap, though, is these meter reports and this kind of work is not actually useful at all in this entire conversation. It's not going to help solve any of these problems. It's not raising any awareness to a point where we can do anything about it. So that's the part that bothers me. Again, I'm not discounting. Again, video editing blows my mind. What has happened in the last six weeks? Like, it is crazy what is available right now and how much better it's gotten. Wait, say what you've seen. So now you can, like, motion graphics on video, you can just ask AI to do. And Claude has already done it, and I'm sure it's going to be everywhere very soon. Like the most complex thing, even prompt-based video editing of like being like, go through this, eliminate all the jump cuts, improve the audit.

You just prompt it now. And this did not exist like two months ago. All of that. So, yes, the pace of innovation continues. And that is heavy, compute, complex type stuff. And it's working. So to me, it's less about is the model front, the frontier evolving at a pace that is still dramatic. To me, that's not the question. It's are you giving us some useful information about what to do about it or just kind of maintaining your position as like the god in all of this ahead of your IPO that is in a bit of trouble? I know I shouldn't look at it just like that, but I can't help it. No, I hear you. I hear you completely. All right. Before we leave, give us your thoughts on the – tell everybody about the prediction market debate that has sort of spiraled over the past week and what your thoughts are. So election night last Tuesday, this week, I guess Tuesday.

It's been, again, another long week. um so my biennial foray into actually watching cnn and abc news and kind of network news which i don't do at any other time i was kind of shocked because even they are all partnering with the of the world and like showing on the screen and like bringing prediction markets into everything the prediction markets almost in every race i think it was all the toss-ups effectively called them well before the networks did themselves. And by calling them, the predictions, the price went up to like 97 cents, 99 cents in a lot of these races. So the prediction markets were ahead in terms of basically saying who won each race before the networks had officially called them. Now, on one side, it's very interesting because it's like, can you bring together the masses and through kind of that price discovery actually be faster about kind of parsing information that is

out there in some way or then there's the other take of just you know networks actually want to get it right whereas prediction markets don't care either way so um you know like but but to me this was it was both fascinating having always been really interested in the space but also it It was kind of depressing and terrifying again because I still view them as generally gambling platforms and not some other type of thing. Where do you fall in the – are the prediction markets the future of political news? Well, I think there's a debate going on this week that sort of will – whichever side you land on will sort of point to the answer there. So, Tarek Mansour, the CEO of Cal, she said, prediction market favorites won 479 of the first 487 races. That's 98%. And Nate Silver, famed pollster, former guest on the show, said, give me a break. The polls and models provided the direction in contrast to vibes and dumb insiders.

And yes, it's good that traders were smart enough to trust them, But the better election models were more confident about the outcome in the markets and got there more quickly. So basically, he's basically saying, yeah, you were able to, you know, hit your accuracy, but you're piggybacking off of the hard work upholsters. Chris Hayes, I'm 100% team Nate in this beef. It's obvious to anyone paying attention that the betting markets just moved over time towards the models. That's the info they were working off of. They didn't discover anything. Well, actually, I mean, on that, I don't know if you saw this. Annie Andrews in South Carolina, who had the shocking win to actually be a senator now from South Carolina over Darlene Graham. It's that kind of reflexivity concept of like the numbers driving the numbers that come after that. Again, in this case, they're saying it's the polling numbers that, of course, then guided the bettors on Kelsey to then make that prediction. I don't know if she her internal polling had actually shown her three points ahead a month ago and they hid it because they knew that if that because that would be kind of this like massive story around, you know, South Carolina, Lindsey Graham's sister in trouble.

So to avoid additional resources and money coming in, they actually hid those internal numbers. So rather than letting them further kind of drive the narrative over what was to come. So, yeah, I do agree that the polling is critical for those kind of markets to actually find the right outcome. I don't know. It is it's it's really interesting to me in that academically, again, I think it's polling has always been kind of weird to me, like especially with advances in technology today. the idea, especially when it was still kind of like phoning people to ask them questions, was crazy. By the same token, I just worry about people gambling their lives away on these things that are not supposed to be gambling apps. Yeah, I've said it on the show. I don't think that the gamblification of everything is a good thing. But it is. Look, if you're watching it, it's a useful data point.

So you can see it. You can see it in real time. You can't see polls in real time. So they got Nate Silver on that front. All right, Ron John, it was great speaking with you. We did it. We made it work. I'll make sure that I'll have cool furniture from now on. I didn't realize it made you so happy. I can see you gambling on that. That's a good gambling chair, colonization chair, cigar chair, all of the above. That is the most multifunctional chair for any vice you could think of. Yeah, including speaking with you. On that note. On that note. On that note, I do want to say on air, thank you for coming out to the marathon. And I know I was too fast, but I appreciate you coming out and supporting. You made it. You made it, man. Congrats. Finished. Knees are barely working, but I appreciate you being there. That chair is also a good marathon recovery chair. It is just good for everything. I'm bringing it back. At this point, I have to buy it and ship it.

It's not part of the show. It's three of us. You and this chair. All right, Ransha. Have a great weekend. Thanks for coming on again. All right. Thanks, everybody, for listening and watching. We'll see you next time on Big Technology Podcast.