On The Brink with Castle Island - Weekly Roundup 09/25/26 (Post Clarity rulemaking, Sherrod Brown is back, RIP BitMEX, Coldcard white hats, is EA a cult?) (EP.741)

Episode Date: September 25, 2026

Matt and Nic are back with another week of news and deals. In this episode:  How loud are America's largest data centers? What's wrong with XAI's data center in Mississippi? The Trump admin is reportedly trying to promote dollar stablecoins Binance invests in Circle We pick up the pieces after Clarity Fairshake plans to spend $30m opposing Sherrod Brown in Ohio How AI anxiety helps the Democrats Sherrod Brown is back to haunt crypto Coinbase is listing IPOs now Are prediction markets out of step with the original vision? BitMEX finally closes its doors 52 BTC were saved from the Coldcard hack by white hats Is Anthropic privatizing the frontier? A White House memo says effective altruism is a cult Content mentioned in this episode: Nic on Substack, How Noisy are America's Largest Data Centers?  

Matt Walsh and Nick Carter are partners at Castle Island Ventures. All of these expressed by them or the guests on this podcast are solely their opinions and do not reflect the opinions of Castle Island Ventures. Guests and host may maintain positions in the assets discussed in this podcast. You should not treat any opinion expressed by anyone on this podcast as a specific inducement to make a particular investment or follow a particular strategy, but only as an expression of their personal opinion. This podcast is for informational purposes only. Brought down by bad mortgage investments, Lehman, which has 25,000 employees, will be liquidated.

The federal government loans American International Group, AI, $85 billion. This is a different kind of market, and the Fed is asleep. The federal government is stepping it to stabilize Fannie Mae and Freddie Mac, the two mortgage giants that have been threatened by the housing crisis. The Bank of England has pumped 75 billion pounds more into Britain's ailing economy with a new round of quantitative easing. You print a couple trillion dollars, and all of a sudden, people start to worry.

So out of this worry, we have something called the Bitcoin. Welcome to On the Brink. I'm Matt Walsh. And I'm Nick Carter. So you're recording from a remote location today. Are you on a ranch right now? I was. Yeah. Actually, shoutouts to Paul McCaffrey and KBW.

They hosted just the most incredible summit here in Montana. A state I'd never been before, actually. I love Montana. I'm sold. I like Wyoming too. I like Montana. I like the whole west. The whole west is great.

And once I get cloud seating out there, can you imagine what the west is going to look like? Yeah, I was actually talking with my Uber driver, and he's like, he was complaining about the data center. Sucking up all the water and apparently water is a pretty scarce resource out here. He was also complaining about the golf courses for what it's worth. Oh, really? You know how the pro data center people say you should also complain about the golf courses?

Yeah. People are as it turns out. They're also upset about the golf courses. Malcolm Gladwell's been all over the golf courses for years, but I think once you can control the weather via cloud seating, you're just going to put those clouds. I still don't really understand how that's going to work, but you know, you seed these clouds. How do you control where they go? Because I think the idea is, you know,

you'd want them in some of these areas that are drought-ridden, but also for data centers. You'd want them close to data centers. Yeah, I mean, I just take an abundance mindset and technological progress is going to solve all of these climate issues, in my opinion, as opposed to us being fearful and scared and, you know, resentful all the time.

What if we just, like, solve all the time? these problems like technology, you know, like I, even climate change. This is probably my most unpopular opinion and I have a lot. I truly believe that we're just going to instantly fix it by putting aerosolized sulfates in the atmosphere. Yeah, you've talked about that before. So, and I've lived my truth on this. I've bought a balloon full of sulfates and I've offset like a decades worth of my emissions by cooling the earth a little bit. Didn't even know you could do that. Yeah, you can. Making sunsets. I'm not an investor. It's a cool company. So yeah, I think China one day, it's going to get kind of hot and they're going to be like, well, let's just instantly fix this for like

$500 million. Super cheap. There's one downside, though. What's that? You don't get to see the sun. Well, the sky will no longer be blue. Yes, we've talked about this. I think that's going to be Deeply unpopular. And I was, you know, I'm in the big sky state they call it. The sky isn't, is actually bigger here. It's true. And I was thinking to myself, you know, in a decade, there might not be any blue sky left. What color would be gray?

Okay. Yeah. So that's actually a bit of an unfortunate tradeoff. And so what do you, do you think they'll just change the colors of baseballs and golf balls at that point? Because it'll be very difficult to catch a baseball. True. Yeah, they're going to have to be like day glow pink or something. Yeah.

Yeah. These are the issues. second order of thinking. That's where your mind goes. Second order of thinking. I don't know how they're going to do that. Well, you had a busy week. I mean, you put out a very good report on the noise at data centers in the United States of America.

Yes, thank you. So this one was actually more the AI than me. I'll happily admit to that. So that's fine. I think the skill is just telling the AI what to do. I mean, I wrote it, but I had my agents cranking on this for a week. And so we looked at the satellite pictures of the top 25 largest hyperscale data centers. I wanted to see if the residents' complaints about noise were real or not.

And actually, they were kind of warranted, in my opinion. Most of the data centers had basically no complaints from what I could tell. So that was already an interesting finding. But about a half dozen of them of the top 25 did. And they all had the same characteristics. they were within 500 yards of the nearest population cluster. A lot of them had on-site generation, so gas turbines. And mostly they were really large, so over 400 megawatts.

And so that's what I found. And the worst one was actually Colossus 2 in Mississippi. So that's X-A-I's data center. Colossus 1 is the biggest one. It's also the biggest one. Yeah, Colossus 1 is a couple miles to the west. And it's actually super remote. it's like two kilometers from the nearest house,

but then Klosses 2 is in the suburbs, got it, unfortunately. And they plonk down 80 generators temporarily. Wow. And they're super loud. So there's tons of lawsuits, lots of independent complaints, so you can tell that the noise issue is real.

And so my conclusion from all this stuff is, there's no reason that a hyper-scale data center has to cause a noise problem. You just have to be smart with site selection. you can put in mitigation, sound barriers. And if possible, you should avoid putting in generation if you're in a somewhat urban area. So I think either we're going to go to this world where, you know, if you're going to have on-site generation,

it's going to have to be something like Stargate in Abilene, Texas, which is so remote. Or actually, galaxies helios data center, middle of nowhere, Texas, 10 miles from the nearest house. So I think you'd have to do that or find an alternative method because the political pushback is going to be too extreme. I think the noise would drive me crazy if I live near one of those things. Yeah.

Yeah. So like the HVAC, I think the noise isn't that bad. Air cooling. But if there's turbines, you know, some of them have backup diesel, but some of them have turbines running 24-7. That's a problem. Yeah. I mean, it's like being on the runway of an airport.

I mean, it's. Yeah. And even over a kilometer away, you can hear it. So distance definitely helps. But see, this is the problem is you're damned if you, you're damned if you don't. Right. So initially, you know, the data started to scoop up all the electricity. People complain about that.

And then the new thing they did was, well, we're going to put in our own generation. So we're not going to tax the grid. Putting in the generation creates a different nuisance, which is the noise. Right. So you can't win. Yeah. Well, I thought that was a good rundown. We'll put that in our newsletter. Why don't we hit the deals of the week? And then a bunch of stories to talk about this week. Yeah, so first up, we have high-fi. There's stable coin payments infrastructure company.

There is 37 million from left-lane capital, Antler, and Arcanem Capital. Then it's Infinify, which is a D-5 protocol. They raise $3 million from electric, new form, and generative ventures. North Capital is a private market's tokenization platform. They're acquired by MoonPay, in a $60 million deal. And then Breed VC, operated by our friend Jed Breed, they have raised $15 million for their second fund. Congrats to the Breed VC team.

Yeah, well done, Jed. So what did you make of this report from Bloomberg that the Trump admin is trying to promote dollar stable coins overseas? There weren't a lot of details. Yeah, I don't know what that means, but that's the biggest no-brenner ever, right? I mean, another way of wording that is,

the Trump administration is pro other people outside the U.S. buying U.S. treasuries, which is a good idea. I don't think you need to do anything for this to happen. There's a lot of dollarization outside the U.S. with physical dollars. Wouldn't it just naturally be the case that stable coins would proliferate without the U.S. government doing anything at all? I wonder if there's things that they could do at the margin just in terms of financial Rails to make them more integrated and more accessible. So allowing for stable coin payments to be used in various payments that happen into the U.S. system. Yeah, I mean, I think the biggest thing has already been done, which is genius. Allow large issuers to enter the market, allow for market structure clarity around the brokerages that want to hold these things. Yeah, I mean, you got that

with genius. I don't know, though. This is one of the, it's more of a financial plumbing question, I think. Swap lines, but using stable coins. or something. So elsewhere in stable coins, Binance has invested 100 million in Circle and they've signed a five-year deal to push USDC across their platform with Circle Ping Binance and incentive fee

tied to USC balances. So historically, Binance had their own stable coin BOSD. That didn't make it. Their flows have been largely dominated by Tether. And now it looks like, circle has made a strong move into the Tether's turf. Yeah, so the way this was structured, it looks like it was a $100 million investment at a discount to the closing price.

I think it was last Friday. So they bought it at a discount and they have a lockup on it, I think two-year lockup. But it installs USDC as that default, as you say, man, what could have been with Paxos, huh? Paxos really got kneecapped on that whole deal. Yeah, that was a really tough one. Just days after clarity collapsed, the CFTC has done the expected thing. They have filed some crypto rulemaking.

What do you make of this? I think it's great. Contents of this are not fully disclosed yet. I think you would expect that there would be some focus on prediction markets, actually, which was not covered in clarity. So prediction markets were deliberately not included there. So I don't know that we have full visibility into what these rules are, but they've kicked off.

And that's great. I mean, that's happening at the SEC. It's happening in parallel to the CFTC. It wasn't too long ago that we had under Gensler, we had this just brutal rivalry between the CFTC and the SEC and they were contradicting each other and, you know, saying that certain assets were securities. The other side would say there were commodities.

So the good thing is that this is happening. The second very good thing is that it's happening in coordination between the CFTC and the SEC. So I think on the major issues that were covered in clarity, you will get rulemaking. I don't know that you'll get rulemaking to cover the BRCA, the developer protections and things like that. You know, the Clarity Act actually had some stuff in there around crypto ATMs and protecting from fraud. I don't think you get any of that. But you will get clarity on how the market structure for tokenized securities out of function.

And you'll get clarity on whether or not an L1. protocol or an L2 protocol as a security commodity. So I think that's really positive. You know, I was talking to Caroline Fam of the CFTC this week. And she said, you know, the CFTC and the SEC sat down for that roundtable recently. Yeah. That was the first one like that in over 15 years. Is that right? So they never worked together that closely on stuff. And so it's actually a pretty major departure from precedent that they are. I think in certain administrations, there's even been talk of combining the CFTC and the SEC. But in the United States, the agriculture industry in Senate Ag, I think would prevent that.

I think that would be a very difficult thing to do politically. I think the U.S. has the record for the most financial regulators, don't you think? For sure. I mean... I don't think it's even close. because it goes down to the state level and it's yeah even federally we have

over half a dozen we used to have more before the financial crisis remember that they combined a bunch of the banking ones post during Dodd-Frank so speaking of the fallout post-clarity and I guess I was expecting this to see to see this even earlier but

Fairshake has started to make their intentions known in terms of where they will be spending their war chest Fairshake, of course, is the crypto industry's super PAC. They plan to spend $30 million opposing Sherrod Brown in Ohio. He's trying to come back. I think that is money well spent by Fairshake. What do you think?

I mean, we just can't get rid of this guy. This guy, man. I don't know. I mean. Historically, one of crypto's biggest antagonists that finally loses a close election to Bernie Moreno. And then what, two years later, he's back? He's back.

I mean, how can you do that? I mean, you can't just banish these people. So he's actually had in his race, according to Polymarket. He's running against John Husted. It is very close, but he is in the lead. He's a chief architect, chief architect of choke point 2.0. He was one of the probably 12 most influential people in getting the crypto industry debanked, I would say.

Yeah. is a tough one. You know, Ohio, they have a lot of data centers. Yes. They have a lot of data centers. And electricity prices have gone up there. And the battle lines have been drawn on the data centers. I think the anxiety over AI and data centers helps the Democrats on net because they're the ones that are more antagonistic. The rights a little divided on it. So I don't think it's a crypto election in Ohio. I think it is AI. And if I had to guess, I would say Sherrod will win. And he'll be back to haunt us again.

I think you're probably, as of today, that would be right. And he's probably, he's probably super vengeful, don't you think? Because he probably blames the crypto lobby with getting him kicked out in the first place. I think he could come back and, yeah, be pretty hot-tempered, I would imagine. But you're right. I think it's 70% of people in Ohio oppose data centers. and that's his core platform right now. And I think his chief of staff said something like,

well, we're not against crypto, being a little mealy-mouthed. I don't know if I believe that. Yeah, you are. Yeah, you are. Yeah, I mean, there's no real ambiguity there. Yeah, you're definitely against crypto. A bunch of stuff happened in the tokenization space this week. New York Stock Exchange and blockchain.com signed an agreement to explore tokenized stocks

and ETFs. They're planning a 24-7 digital trading. venue. And then I thought this was really interesting. BlackRock launched three tokenized model portfolios this week. They're available to non-U.S. investors, high income, diversified growth, and high growth. They all contain some mix of stocks, bonds, and Bitcoin, and they are all on chain. That's a pretty big deal. Another interesting piece of news that caught my eye, not tokenization, but Coinbase is now offering just regular old equities, IPO allocations on the platform. I don't believe these are tokenized.

I think there's regular old stocks. The first one is ORA, the ring company, biometrics. I love ORA. I might actually buy this one, do. So this is just, Robin Hood has done this. I think they started like a year ago,

but I think this will probably be popular. I just need to figure out my notification settings on Coinbase app because it's brutal. It's every time there's any sports game, it's like, do you want to bet on the platform? And it's like, well, I'd never have. So why would I want to do that now? Why would I want to bet on like a women's UFC fight on Coinbase?

I think some of the prediction market old timers probably look at the state of prediction markets right now. And they think, well, this is a little sad because it was meant to be out about using markets to create novel and useful information flows so that society could make better decisions. And instead, it's actually just a slightly more efficient way to do sports betting. which is a good dovetail into New York Attorney General, Letitia James. She is back on the podcast, suing Polly Market, alleging that they're running an unlicensed gambling operation in the state of New York. This is two months after she brought a similar case against Kalshi. So that goes to your point.

And it's, I mean, this is all coming to a head at some point, but sooner the better, I think. I mean, if the sports thing doesn't work, they're going to need to find traction with, you know, markets with natural hedgers. I have a deep dive on this, actually, on something that happened this week in Polly Market that I haven't seen reported, but there was a CFTC filing this week. And Pollymarket U.S. has raised their fees on just the U.S. platform. And basically to bring them in line with Kalshi. So the taker fee coefficient went from 6% to 0.069. So close to 7%.

I think Cal she's at 7%. I wouldn't say we're in a price war. I mean, this is more just getting in line with what CalShe already offers. But it's a 15.8% price increase leads to about a 20% increase in retained fees for polymarket. I'm very interested, assuming that these markets continue without being shut down at the state level, just where this settles on fees, because there's starting to be a lot of these prediction market platforms. Yeah, and this kind of sense or view in the investor community that it's going to be a duopoly or a monopoly. And I actually don't know if that's true.

You know, I don't know why there necessarily would only be one or two of these that win. I think it's just going to cost a lot of money to go out and acquire customers and there are only so many platforms that can do that. But there's plenty of, if this remains kind of a market structure that allows for sports gambling, I guess you could just see a lot of brokerage firms adding prediction market functionality. Now, who would they use on the back end? I guess you'd need a DCO and maybe that does favor one or two of these things. But it'll be interesting. I think a lot of these international brokerages have already added prediction markets.

So there was a Wall Street Journal article this week about $5 billion in trades on CalShe. There are all $5,500 on one market, which was the price of ether. They didn't really allege that anything super untoward was going on. It's just that they spotted the pattern. I don't know. I guess I just don't understand what that market structure inherently needs to be. Because I guess sports books do that, right? But it's a construct that is incompatible with my understanding of how financial services ought to work.

Well, speaking of market making on your own exchange, Bitmax shut down this week. RIPP. So what a run. Pour some out for Bitmex, even though there were some controversies there. Over the years, there was the March 2020 issue. Was that a Bitmex thing?

Oh, yeah. They just remember that? Yeah, they might have done everyone a favor. They pulled the plug on the exchange for a minute. Bitmex invented or popularized, I guess, the perpetual swap. A real crypto, venerable crypto institution. So I don't know.

We don't have a song on this podcast for, like in memoriam or anything, but you'll just have to imagine it. So rest and peace. You know, obviously they had a ton of controversy over the years, but as far as I know, never lost any customer funds. And that really means something. You know, I think they had that two or three multi-sig or three of three with the founders. And every day was it they would process withdrawals?

Something like that. Yeah. For like a decade plus. I mean, never got hacked. I don't know what happened to the, liquidation fund. Because there's a big pool of capital for a while, right? I saw that the estate of Celsius, the bankrupt brokerage, was suing Bitmax at the 11th hour, trying to get access to some of that, alleging that they lost money in the liquidation,

the COVID liquidation, which that seems like a stretch. But yeah, the liquidation fund is owned by the owners. So I would assume that just gets prerodibly distributed. Yeah. Well, certainly controversies. had some problems. There was the, I guess, the BSA violation. Yep. But I, yeah, I consider them one of, you know, one of crypto's good guys. So yeah, rest and peace, Bitmax. True pioneers. Here's another story. Strategy. They are back to buying Bitcoins. They bought 950 bitcoins for $75.7 million. First big purchase of Bitcoin in a few weeks. And, you know, hand up. I guess, I was wrong about strategy being in a bind because that guy just smashed the at the money equity

has a ton of cash and you got to tip your cap to him. So another Bitcoin story, remember the dreaded cold card hack from maybe a couple months ago? Of course. As it turns out, 40% of the Bitcoin that moved in the second wave were white hats trying to proactively recover Bitcoin that was at risk so that their owners could claim back those Bitcoin. Bitcoin. So 52 Bitcoin were evacuated by Whitehats. And I saw they've established a trust or some sort of a process for users to try to get the funds back. Is that right? Yeah, Wyoming Trust. Now, it's called the Cryptoaccovery Trust. Doesn't this make you think of the Satoshi coin situation?

Yes. Isn't this kind of a template for that? I mean, white hacks, sweeping coins, putting them in a trust. the true owners can come and claim them. This is what should happen with Satoshi's coins. So how would that work functionally? So let's say Quera or Cy Quantum or one of these just gets their first. You know, presumably they're going to be using this computer that they've built for a lot of different things. I guess going after Satoshi's coins wouldn't be top of the list. So is it going to take some...

It might be. It might be top of the list. Well, it's certainly in the business plan for a lot of these. China do it. Yeah, I guess that's right. So maybe they'll just... I think they've a moral obligation to do this. If you're a good guy with a quantum computer,

you have a moral obligation to rescue the coins. This is where whoever is the secretary of treasury is going to factor in pretty large to these discussions, I think. Yeah. And whenever it happens, I mean, the more I pay attention to the quantum, I'm actually in the quantum building right here

in Montana. there's a quantum industry. The more I learn about it, the more word I get. And the problem is, is that AI accelerates it like crazy. Turns out AI is pretty good at math. 100%. Who knows? Maybe Anthropics can announce next week that they've solved it.

Yeah, what did you think of the anthropic thing yesterday? They discovered what a new, what was it, an enzyme path or something? New CRISPR, CRISPR 2.0. Yeah. Yeah. I mean, this is my broader thesis around. Anthropic, I think they are pacing the frontier externally, maybe not, but certainly plowing ahead internally and trying to add the generation of novel IP to their kind of portfolio of businesses.

And I think at the end of the day, if some client of theirs is spending $2,000 in tokens to develop a new drug that's worth billions of dollars, Anthropic wants to internalize that whole thing. Yeah. I think that's what they're doing. It reminds you of the almost like the Facebook Zinga platform challenge where Zinga was making all of their revenue building on top of Facebook. And Facebook just wanted to internalize gaming. So they took that over at one point. It's the same exact conundrum here.

And I don't know how. So then that would be an argument, I suppose, for self-sovereign AI. But then you're not going to be at the frontier is the problem. No, but if you're an enterprise, what are you using AI for really? You're using it for customer support tickets, right? That's what you're actually using it for. Yeah, it's probably building PowerPoint decks and it's helping your answer to know. It's like boring stuff.

You don't actually need Astra 10.0 or, you know, whatever the new Claude is. You don't actually need that, right? Right, because Astra is going to, you know, six months from now, Astra won't even be a frontier model. And the Chinese open weight models, yeah, they're six months behind on the frontier. That's good enough for most business tasks. Yeah. I mean, so I think that that's where this is going.

You're just every enterprise that's not doing deep science, exploration, deep tech, R&D, they're just using eye for boring ordinary stuff. They're going to have a cluster of GPUs in the basement running a Chinese open weight model and they're going to call it a day. Yeah. That doesn't mean Anthropic won't ever make money again. I think they will. But I think they're going to be a biotech company. Lumpy revenue for a biotech company, though, right? Yeah.

I saw some PhD that was saying, that discovery wasn't even that big of a deal. It's like, well, it seemed like a pretty big deal to me. I don't know. I mean, it's like the mathematicians being like, hey, guys, the point of math isn't to like solve math problems. It's actually for us to have the experience and enjoy the process of solving them. and so now they hate the AI. Every industry on Earth is going to go through this existential question, trying to gatekeep. Isn't it just such a peculiar thing that as human beings, we don't actually understand what is inside of our own bodies that we're discovering these new things? We don't even know how our brains work.

Yeah, I mean, I'm on team. Let's just discover as much science as possible whether or not it offends the scientists. We have the ability to program our genes so that we don't get rare diseases and things like that. I mean, that would be great. So speaking of Anthropic, Axios is reporting that there is a memo circulating in the White House. It's not an official memo, but it is a memo that is saying that effective altruism is a cult, and Dario is one of the cult leaders, and basically just be careful of these people. So I think they listen to the podcast.

I think that's what's happening here. Yeah, I mean, did they only just figure this out? It's, I mean, it's been, we've known this. We've been shooting off flares on this issue. for years. We've known this since 2022 guys. I mean, come on.

And we've let it go. It's gone too far at this point. Now they control everything, these people. I don't want to be in the permanent underclass of non-EAs while they run the world from their AIs.

Yeah. I mean, it's a... It might actually be too late to stop them at this point. They might have all the power already. Well, if they pace the frontier, maybe everyone else can catch up. And but like the question is factually, you know, is it a cult?

The answer is obviously yes, 100%. There is a quasi-religious belief in the coming apocalypse or rapture, whether it's good or bad. They organize their lives around this. They have weird social behaviors. There's some odd polyamorous element. Some members of some offshoot splinter groups kill people. The Zizzians. Have you heard of this?

her to them. Honestly, for your own sanity, don't even look into it, is my advice. There's weird sex stuff going on for sure. That's a cult thing. I mean, it fits all of the descriptions of a cult, without a doubt. I mean, I'm just glad someone in the executive branches is taking into account some of these things. They're definitely not to be trusted. And because they're utilitarians or consequentialists, they believe the ends justify the means. You know what that means they're painting with a broad brush here, but basically they adhere to a philosophy that says you can do short-term immoral things in pursuit of a longer-term good goal. So that makes them extremely dangerous, in my opinion, because they can justify practically any action, right,

in pursuit of their broader goal, whatever it is. So that could include lying about their motives. That could include false flag attacks, such as potentially releasing an agent swarm, you know, the likes of which we've seen recently or worse things, frankly. So I consider true believers, true adherence to this, extremely dangerous. Yeah, yeah. Will McCaskill really is a man, a truly influential human being, probably in the arc of history, huh? It's pretty incredible that like just a couple of philosophers have really changed the course of history, it seems. Shows you how ideas are just the most powerful thing humans can invent really, really.

Like Nick Bostrom in another era, no one would have ever heard of him. Yeah. But he's just, you know, Swedish academic wrote a couple of papers, but his ideas have influenced everyone in Silicon Valley. It reminds me a little bit of the pre-atomic age where physicists really played that role. Niels Bohr and Oppenheimer, those type of teller, those type of guys were truly the most powerful of the elite. And you know, I think that's a great example actually to look at the EA's because the physicists, once they developed the bomb, they decided that because they had done

the science, that they should have a kind of diplomatic and political role. Right. Shape the existence of the world to come. And they kind of universally failed at this, right? They were not able to parlay their scientific expertise into steering the political agenda. Right. You know? And I think they totally overestimated their ability to do that. And I think that's exactly what's happening with the AI leaders right now. You know, they obviously have technological expertise. Does this mean that they are the right people to determine the shape of the world to come? Certainly not.

Yeah. Right. And I don't, and you know, people often say, well, you know, the AI people, they're close to the frontier. They know what's going on. So we should listen to them on safety. I don't think so. I think it's anti-democratic to let a handful of technocrats

in Silicon Valley decide what our priorities are as a society. So everybody should go read this book, The Making of the Atomic Bomb. Very good book on the history of the Manhattan Project and the personalities. And I think there's a ton of historical parallels there. You know, the other historical parallel

is that both movements became infiltrated inadvertently, I would say, by communism, where the communists got access to the technology. It goes even deeper than that. You know, you look at the motivations of the people that stole atomic secrets from the Americans. Our lead only lasted four years, right? They were Marxists.

They believed that their allegiance was to something greater than the U.S. Does that sound familiar with the EAs? So they had allegiance to transnational, philosophical concept or they were Soviet sympathizers or they thought that no one country should have this technology because it would be too empowering. Sound familiar? I mean, come on.

So you look at why the Soviets got the bomb so quickly, the personalities, the motivations, you see a lot of that in AI today. Well, I'll be the one to say it, though. There were some Brits involved in stealing that tech. There were. And some were when unpunished. Yes. Actually.

Yeah. Yeah. Yeah. They were not the best friends to the United States during that process. I mean. All right. We're up against a hard stop. So let's leave it there.

Everybody have a safe and healthy weekend. And we will see you on Monday.