On The Brink with Castle Island - Weekly Roundup 07/31/26 (Situational Awareness collapse, Coldcard exploit, latest on CLARITY, Visions of Bitcoin 8 years on) (EP.732)

Episode Date: July 31, 2026

Matt and Nic are back for another week of news and deals. In this episode:  What the novel Hyperion tells us about AI doom Situational Awareness liquidates most of its public book after a sharp AI selloff Will Leo have a second act? Why has AI been selling off? Over 1000 BTC stored in Coldcard wallets were stolen due to bad entropy Where does self-custody go from here? What happened with Coldcard? Political winds are shifting against crypto as midterms loom Strategy continues to strengthen its balance sheet with MSTR common sales An ice cream shop in CA is hedging the weather with Kalshi Circle buy's IBM's blockchain patent portfolio Visions of Bitcoin turns 8 this week  Content mentioned this episode:  Nic and Hasu, Visions of Bitcoin

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 Brinkum, Matt Walsh. And I'm Nick Carter. Recording this one late, so apologies, it's not out, but you were on a lot. a long-haul flight and I was at Fenway Park. Yeah, I was on a plane all day yesterday and I have a bono pick with Swiss Air and the European Union actually for that matter, both of them. Swiss Air they just straight up don't have internet on their flights. Oh, awful. How is that acceptable? They're on some third gen 90s internet system that never works basically. Meanwhile, United has Starlink

on the planes and then others have VSA which is okay. How can they get away with that? They're meant to be nice airline. There's no excuse for not having Starlink at this point. Yeah, it was a hellish 10 hours. I'd lock it and just look at the flight map. And then the EU has switched to this new system where have you been through the EU recently? Yeah. So they switched to this new system where everyone has to get fingerprinted when they come in. And so the immigration lines, when you're connecting even through the EU, it takes 10 times longer. So fast pass. They need that fast pass over there. I connected through Zurich yesterday and it was bedlam. It was just a total nightmare. It's awful. And in theory, it'll be more efficient in the end because then everyone will be enrolled in this biometric system.

But man, it is really rough out there. That's tough. So you just read, I guess you just go back to the old-fashioned stuff that you used to do in playing sleep and read. Yeah, I've read. Have you heard of this novel called Hyperion? I have heard of it. I've not read that. You know when you read or watch a film or something and then you realize that there are so many cultural references inside that media that you weren't aware of and they've become apparent to you and you realize oh that's what people were talking about you know hyperion is like that farcasters that's a thing from that novel no kidding i didn't yeah yeah so i thought farcasters were just made up word before it's a reference to this novel and then i also think a lot of what the i people talk about

with the doomers thinking that super intelligent AI is going to kill us. That's basically the plot of this novel. Sorry, spoil it. It was written in the 90s or 80s. That's exactly what happens in this novel. So I think Yudkowski is just re-gurgitating that novel, science fiction novel from the 80s. Who wrote Hyperion again?

Dan Simmons, I think, is the name. But I recommend it. It was really trippy. But anyway, this whole world of references that SF people like to make, that became legible to me for the first time. Well, it's good. All right. So there was some positive byproduct of not having the internet for 10 hours. And so you went to a country concert and you saw formerly known as the Dixie Chicks and formerly known as Lady Antebellon.

Is that right? Yes. And Tim McGraw, who has not changed his name. But so not, I wouldn't say it's like my typical genre of music, although I do like country. music, but this is kind of like older country. But yeah, I guess they all change their names because you can't have, you can't say antebellum or Dixie anymore. So you just can't acknowledge the existence of the South.

Yeah, I guess. The South is awful. But it's like, well, you guys pick this name in the first place. It's not like someone thrust that upon you. But yeah, I was at Fenway. It was good. Tim McGrath is, I would say he's a better actor than singer even though. Have you ever seen this?

Is he an actor? He's in that. 1883, he does a great job. He's still a good singer, though. But yeah, games at Fenway are coming to a halt next year. Do you hear about this baseball strike? Why? It's going to be a lockout.

So I think it's just going to be all concerts next year at Fenway. It's just no prospect of ironing out a deal because I think the owners want a salary cap. So that is why we missed our podcast yesterday. That's why we're late. So our apologies for that. But we're here now. We're here now. We have a lot to talk about.

Much news. Yeah. this week was insane. Let's do the deals of the week and then I want to talk about Leo. So first of up, this one's from the SIV portfolio. Bird Eye Labs that's spelled bird AI, one word. They're an on-chain trading infrastructure company. There is four million from us, Castle Island,

Medley Adventures and the Venture Department. Yeah, very excited about this one. Kevin Farrell and Greg Scanlan, formerly of Franklin Templeton, really excited about what they're cooking up over there. So congrats to. the Bird Eye team. Next one is Access Robotics. This is a physical AI training data network. There is $12 million from hack, nomad, pie network ventures, and others. Then you have a world. That's Sam Altman's Iris scanning identity network. There is 52.5 million through a token sale to

Pantara, Bay and Capital Crypto and Susquehanna Crypto. Surprise they don't have that iris scanner at the airport. So I could have saved you some time. They did fingerprints and face. recognition. They did. Okay. Missed opportunity for World. Yeah, they got to get in on that. Then we have last one is Magic Labs. This is a wallet infrastructure company. They were acquired by Cracken. So without further ado, we have a FTX alum who blew up his fund. Situational awareness. This is your, this is your boy. Yeah, I mean, I've been obsessed with Leo basically since he wrote his manifesto, not in a weird way. I mean, not like in a creepy way.

I just think he's a very interesting thinker. And what is story? What an arc? What a life he's had? What is he? 24? I think he's 24 years old. Started a hedge fund.

Ran this thing up to what, $50 billion in like two years, $45 billion. This guy has lived so much in his short 24 years. Oh, yeah. So he goes from FTX to OpenAI, gets fired from, Open AI starts this hedge fund is it probably the best performing new hedge fund in the past decade

or so? I mean, it's got to be. It's got to be. He writes the defining manifesto of the 2020s. Yeah, situational awareness, which is just what 2024, he put that out and then started the which has basically been true. Yes. So far.

I would say it's been totally correct. He basically says everyone's missing this. There's only 100 people. in the world that know what's going to happen. I'm one of them. Give me some money. People give him $500 million. He runs it up to $40 billion. And then he used to puke most of it out. Unbelievable. So he's running this thing, Forex Leverd, which I didn't realize that that was part of the deal over there. But he's running the fund of Forex Leverd. And this week was choppy, right? So you had Citadel come out and say, the Fed is going to hike. Everything starts to sell off. Obviously,

So they didn't hike. They held flat. But he was getting margin called. And so that big puke on, what was it, Tuesday or Wednesday this week? Everything went down. It looks like that's when he was getting margin called. He had to do a block sale to Citadel. Yeah.

I mean, it's some of the holds a lot of AI stocks. I mean, I basically have a very miniature version of Leo's portfolio personally. It's been really tough for the last month or so. I mean, they've been selling off horrifically. I think Kimmy has been part of. it. I think that Kimmy is the second deep seek moment. Everyone's freaking out that the labs are going to suffer some impairment to their margins or profitability because they'll be cheaper distilled versions. I think that's kind of not true. Actually, if you look at Open AI, we're kind of leaking

some revenue numbers recently. They seem to be very healthy. But people think the labs are impaired, and then the labs are the nexus where all the cash comes from for the data centers and for the neoclouds and so now the neoclouds are selling off and so there's been this drawn-out panic for these pure play i i mean leo got caught up in that i mean he was horrifically exposed and he was so exposed that he couldn't sell because he owned too much of the stock that's in all these across all these stocks so he would own like i i don't know from his disclosure's like double-digit percentages of the float. And so in the end, he suffered this big impairment.

It looks like the AI PurePlay name sold off. He had shorts, hedges on just general software, semiconductor sector. That rallied. Both sides of the trade went against him. He had leverage. He sold almost all of his public book in a block trade to Citadel. And the fun lives on.

He's still up, isn't he? inception to date. I mean, he's still up. So he's not like a long-term capital management thing where he's out of business, right? He still has $5 billion of anthropic private. Everyone's writing his eulogy. This guy has a huge privates book, not just anthropic fluid stack as well. He's got tons of privates, which we don't know about. I think he probably has foreign stocks as well. Those don't get disclosed on a 13-off.

His book isn't that legible, and he's certainly up lifetime to date. He was up hundreds a percent this year, down 67 percent in July, still up. And I think basically all of his LPs have still made money with him. So he apparently put out a letter to his investors earlier today or late last night. And it does seem like he's going to continue. But I'm surprised this didn't actually get a lot more attention. Like this is one of the biggest blowups in the hedge fund space in the past, what, 10 years probably? Yeah. And you know what? I think he's actually going to survive and maybe thrive. Yeah. Because here's the thing. People say, oh, well, he didn't have

good risk management. But that was part of the deal was he was willing to take risks that others weren't, right? Yeah. So when situational awareness came out, everybody made fun of the guy. Do you remember? Yeah, I do. Everyone made fun of him. Yeah. He raised a few hundred million dollars. Everybody's like, this guy's an idiot. He's a kid. He doesn't know what's. But his LPs were signing up for one specific trade, which is max long, max, uh, you know, full AI exposure. through whatever vehicle you can. And they knew that that would be a choppy trade somewhere along the way. Yeah.

So I think this could be the start of a good second act for Leo, where he still apparently has some of a public book. He said that in his letter. And now he's learned risk management in a very brutal and expensive way. But he's still clearly the best when it comes to this trade, as far as I can tell. John Arnold tweeted yesterday, optimal number of past blowups for a hedge fund manager is one. So, you know, maybe he learns from this and goes on. But I wonder if all these LPs really were comfortable with the leverage.

I mean, if he just didn't lever this up, it would have been a lot better off. But the thing about who has disclosed LPs are. They're like other EAs, probably SBF, you know, family trust to the extenanty has one in jail. They're tech investors, their VCs. are people that are comfortable with risk. Yeah. Right. So what would you do if you were an LP in this guy and you'd see him grow half a billion to 40 billion and then he gave up half of it. Maybe he's at 20 billion now. You'd be like, well, you know what? I put my trust in you. Yeah, you goofed up, but this thing still works. Go do it again. Right. I think he probably will get the benefit of doubt from a lot of these LPs, but I guess it depends on who the LPs are

because an institutional LP would look at this and say, well, where's the risk management? Why you run this thing, 4X levered. If you're 4X levered into a semi-ETF, you would have outperformed to him. I get that that's not what he was doing. It sounds like maybe you're short semis. Yeah, I mean, he could do the boring thing. And it's ironic that we would say this, which has just become a venture capitalist. I mean, he's pretty good at the privates, right?

He could easily be a crossover kind of co-to style investor. He could do that and exit the public. eye and focus on the privates. But I think he'll try again with the publics. He says in his letter, he said, we have a fully paid for book with public book, which is long stock and fully paid for options with no margin liquidation risk. This restored stability and allowed us to preserve our private positions. So they could, I don't know what the terms are of the deal with Citadel, but he could try it again. Well, yeah, I think the Citadel deal is just a block trade. So it's not like Citadel owns him, but Citadel was a willing buyer.

It looks like he also had a deal lined up with Green Oaks and Sequoia to buy the Anthropic private position, but he chose to sell the publics instead of anthropic. This is like the best case that Sam could have gone through, SBF. I know. You know, Sam dreams of a situation like this where he managed to get rid of everything in time to save the empire. Well, I was thinking about that, too, is this is just a much better blowup in the sense that he was borrowing from prime brokers and the system worked.

He was about to be liquidated. It was potentially liquidated. And there was a block trade. And he didn't have to, you know, pull any of the shenanigans that FTC did. If we had a market structure bill, FTCS wouldn't have happened if they're operating onshore with proper guidelines. Right.

And this is just what a blowup looks like when you're not running a wild west offshore exchange. So he said they're next. net month to date performance is minus 67. Pretty tough. Yeah. But their year to date net is plus 80% after everything. I mean, anyone would take that.

Which is incredible. Yeah. So, and everyone's like dung on the own acting like his career is over. I don't know. I think he's 24. Oh, he's 24. He's 24. He's definitely not over.

Definitely not. My career had not begun when I was 24. No. I mean, this guy will be around for a while. But I think you're right that he could raise a ton of capital if he was just a private markets investor too. So you wonder if that's going to be part of the go forward there. I also wonder, do you think he just lost access to Fable? Do you think he had Fable before everyone else?

And then, you know, performance started to suffer once it was throttled. I think given that his wife is the chief of staff at Anthropic, he might have some kind of proprietary access there. He still has Fable. I don't know. The timing's a little bit odd. you know, Fable get shut down by the government, can't have the Mithos model. Maybe he had that. Maybe that explains why you're so good.

He still got Fable. He's actually getting married this weekend, I think. I saw that. That's tough timing. I guess now it's behind him, but man. Some awkward conversations to the wedding. All right. So the other big story here, this is a really sad one.

Cold card wallet. I don't know what the latest number is. I'd seen on Twitter here that it's over 600 bitcoins have been lost as a result. of this. So for those unfamiliar, cold card is a hardware wallet. It seems like there is bad entropy on this hardware wallet. And if you had the wallet set up from 2021 to I think 2024, you're at risk here. And I would just say if you use a cold card at all get your funds off of it at this point. But this has turned out to be just a really sad story here. We have a lot of bitquiners that have lost their bitcoins as a result of this hack. Yeah. And we're both historically,

least cold card and I guess coin kite users the open dime I've been a fan of as coin kite product so yeah really really tough one here because cold card is kind of like the bitcoiner is bitcoiner that's who uses it and it's a real robust self-custody wallet open source I think too and the problem here was one of the software versions had bad entropy and that's kind of the one thing that the wallet is meant to do is have good entropy. Yeah. And so if you just trusted the wallet's entropy, I mean, you could have done this thing with rolling dice a thousand times or whatever. I don't think anybody in their right mind would want to do that.

Some people did, I guess. But it seems like there was a bug who was misconfigured in terms of the RNG source. And over a thousand Bitcoin were swept in a 40-minute period last night. That's crazy. I mean, yeah, to set one of these things up, do this dice roll thing, but no one in their right mind is going to do that. I mean, what is the percentage of people setting up a hardware wallet that are going to go, you like put 50 dice rolls of entropy into something?

So you just can't expect that out of a user. It really does, it calls into question, I think, hardware wallets that are not multi-sig on multiple different devices. So I think if you're, you know, if you're set up with like a CASA set up with the multi-sig, that makes you feel a lot better. But single device, you know, hardware wallets, if you're just relying on one, I think that's a tough thing. I don't know if I'd feel comfortable doing that. This is a really bad one because this is basically the most paranoid class of user, the kind of person that's using an open source, relatively niche hardware wallet. And even

they were not safe. And so as you say, what you can do is you can do a multi-sig with CASA. Kasa manages the software. You bring multi different vendor hardware wallets together. So you bring a cold card, a ledger, a treasur together, and maybe Kasa manages a key. And so you don't have the dependence, you know, you're protected against any one of those being compromised or lost or, but this does really illustrate the problem of Bitcoin self-custody, which is the more paranoid you are, typically the more complex your setup is going to be. Yeah.

And then you have this issue of foot guns where now you've forgotten your 25th word or you've forgotten how to reconstitute your wallet. And it's just really, really hard. And I don't begrudge anybody that is using an ETF or using Coinbase for your custody. If you're wanting to do self-custody at this point, I think you have to go multi-vender with a multi-sic. with a software provider like CASA. So there's a, I think we're going to have to have a post-mortem on what happened here, but there's some chatter that the reason that this even happened in the first place is because

they changed the licensing scheme on the back end of cold card because passport was building on them. Did you see this? Yeah, I did. Yeah. So it seems like it was sort of like an own goal. You said it's open source, and I guess it is, but they changed the license and that's when the software was upgraded.

So I think if you were on the older versions of Colt Card, you'd be fine here. It's pretty unfortunate. I mean, and I don't know if there's any recourse because I don't think this is a company that's capitalized in any way. So I guess they just will cease to exist. Yeah, I don't know what is in maybe the terms of service says they're immunized against claims, but this really is the manufacturer's fault. Yeah. And I mean, it was open source, so anybody could have caught it in theory.

But, and NVK actually said he did his own AI-based audits and didn't catch anything. So he actually suggested that the attackers themselves were probably using new AI tools. And they went through the code base. I mean, it's got to be North Korea, right? Like, could just be, you know, a smart hacker. You know, these, these skills have been essentialized now. It really illustrates this time of uncertainty we're in with, I think AI as it is, right now favors the attacker over the defender, especially because the good guys, they do all

these infoset queries. They keep on getting said no. Yeah. Like the labs keep on denying their ability to operate. So it's a peculiar moment in time because the attackers are empowered against the defenders. Yeah, it's a sad one out there. All right, let's move on clarity act stuff. Clarity Act is at a a low right now. So Clarity Act is 24% on polymarket. Senators Tom Tillis and Ruben Gallego have reportedly rewrote the ethics conflict of interest section. That's the one that's aimed at Trump's crypto investments. They've sent it to the White House. I don't think the White House has reacted yet. But again, it did sell off this week on the prediction markets. So people are pretty bearish on this. I'm the other way on this right now.

I think 24% on polymarket is too low. It's been a lot of work that's gone into this so far. And you just look, there's some trades that can be made here. So there's a lot of chatter around Democrats getting to name finally their SEC commissioners and CFTC commissioners. I'm wondering if there's a grand bargain to be had there. And really, I think there's a bunch of Democrats that don't want this to be an election issue. They just want to move past crypto and never talk about it again.

So I think it will come down to this ethics provision and whether or not the state attorney generals can play a role in enforcement. That's definitely a big piece of it. There's other bills that have been fashioned with state AGs being involved in having the ability to sue the Justice Department as opposed to going after a defendant directly. So maybe that's a path here. But how are you thinking about the current state of things? I'm probably aligned with Polymarkert's assessment right now about a one in four chance it goes through. It is weird as a legislator that you spend months of your life on something and then you have absolutely nothing to show for it. That's a tough career.

It is. I think it'd be a shame if this fell apart. But let's say it does, as we've been saying, I think you'll quickly get rulemaking. And this week, SEC Chairman Paul Atkins came out and said he's ready, willing, and able to institute rulemaking. for the industry if the Clarity Act falters. So if it comes to it, I think we'll just get SEC rulemaking. And that'll be enough for big financial firms to tokenize things and rely on that guidance. And I think it would be hard to roll back a rule like that in the future.

I mean, obviously that does happen. But if you have the biggest financial services firms in the world that are relying on a rule to launch new businesses, that's harder to roll back. I think it's easier to take a rule from a, hey, you can't do this to a green light. But once you're talking about shutting down a business line for a big financial services firm, those are the types of things that get fought in court. So we'll get the rulemaking, I'm sure, but we'd rather have the bill. Yeah, my political assessment is that Democrats that it might have been on the fence are feeling a little more emboldened to move to the left in this election cycle because the polls have been pretty favorable for them.

and crypto is one of these issues where they're, I think even despite the existence of the lobby and Superfax, they're willing to lump it in with prediction markets and AI and data centers where they're just willing to take the political hit or, you know, funding, take the risk to side more with the base, which is staunchly anti-crypto, very skeptical of big tech, very skeptical of AI, skeptical of data centers. So that seems to be the emerging consensus to the left of center at the moment. All this is kind of a calendar issue at this point in terms of just the legislative days to get something done. And the Senate's supposed to go on recess on the 8th. But, you know, they don't have to go on recess.

Like, they don't have to take a vacation here is the other thing. So there's precedent to just extend the calendar. They do take an awful lot of vacations. Yeah, it's like you don't have to go on vacation. So why don't we just stick around and get the bill done? As a senator or whatever, I guess you have to do stuff in your own state occasionally. Is that it? You got to like, you know, go to the county fair and like cut a ribbon at, you know, veterinary clinic or something.

Is that right? Yeah, you've got to go back and you've got to actually speak to your constituents and the people that put you there in the first place. It's kind of an interesting job because half of it is like, kissing a baby, and then half of it is working fruitlessly on legislation. Yeah. The other half of it is just kind of grandstanding, you know, at hearings and things like that. So Binance U.S. plans to apply for a CFTC DCM license. They want to offer regulated prediction markets.

So the CEO, Stephen Gregory, to Binance U.S. They were back from the dead. Yeah. Kind of surprising, right? Like, that business never really took off. whatever happened to Catherine Coley. I think she's just living her life. She exited and no one's heard from her since.

No one has heard from her since she stepped down from the head of Binance US. And that was years ago. I'm sure she's glad to be out of the burger world. I'm sure. Strategy in the news this week again, you have to hand it to Michael Saylor. The guy knows how to sell common stock. He sold another $544.5 million.

of micro strategy common stock. The cash reserve is now up to 3.75 billion. He still has 843,775 bitcoins. The company also bought back 25 million of STRC. It's their kind of first attempt to defend that preferred in that manner. You know, I don't know what to do other than just tip my cap. I thought it would be a lot harder to get people to buy this at the money common stock when it's effectively trading at a negative MNAV

if you do it the way that most people do it. But the man can raise capital. Saylor just has this amazing ability to wriggle his way out of any jam he's in. And I was actually thinking about this in the context of Leo. Right? So Leo lost, I don't know, peak to trough, maybe $10 billion.

There's not a lot of people that have done that and had a second act. Yeah. But Sailor is one of those people. Sailor is. I mean, yeah, he went down. He lost tens of billions of dollars. 99% in the dot-com era.

Yeah, so he lost, I believe, over $10 billion. It's not, you know, cash in the bank or anything. And then he's had a second act. And now maybe he's going to have a third act because people were writing him off. It's not to that I, you know, not to say that I think MSTR is a good thing to own. But it is really incredible that he's squeezing more juice out of the lemon, which is common shareholders. I don't know how he does it.

I don't know what these people are thinking. but he has to a certain degree right at the ship or bought himself time. I don't know what the next step is now that they have a decent amount of cash on their books. But, you know, congrats to him. I mean, he's got tons of time here. And if you're a four-year cycle person and, you know, we're getting towards the bottom, right? So I think Sailor is probably just getting ready for the run up. So STRC, despite he bought back 25 million of STRC, not a huge amount.

STRC isn't taking the bait. It's still at $89 pretty stubbornly, despite the much larger now cash reserve. So it seems like STRC holders are still wanting a higher interest rate, basically. Yeah, and there's a lot of chatter that there's some restructuring going on on the back end. And maybe STRC holders are going to get some equity and, you know, there'll be a deal to be had there. And the Vulture funds are kind of all over us.

TRC at this point. So I'm sure at some point in the next couple of weeks, we'll be talking about some grand bargain there. So we have some prediction market related stories here. First of all, Robin Hood posted record Q2 revenue of $1.30 billion. That's the first time their prediction market business outearned crypto and spot equities. Wow. So event contracts generated 156 million against $100 million from crypto. It's a real changing regard moment there. That's pretty wild. that just goes to show you where the retail appetite is right now and it's in the prediction market game it's definitely not in crypto so i have another very funny prediction market story which is there's an ice cream shop in california have you heard of them no 28 wishes no so they're an ice cream parlor in l.a and you know how we talk about

how prediction markets need natural hedgers to finally uh you know one to really bootstrap and become self-sustaining. Yeah. In their own right. They are an example, I'll be it a small one, of a natural hedger because they bet on the weather because business is bad when it's cold. Huh. So they buy hedges constantly on the weather so that they make money, I guess.

The hedges pay out if it's really cold and rainy. That's interesting. So they're a natural hedger. They will. They put $20 a day. I was going to say, what type of scale are we talking about? This feels like a complete PR type of deal. I mean, it worked because we're talking about it.

The New York Post is talking about it. But yeah, they're naturally hedging their ice cream profits. Good for them. Good for them. In other news, Circle has acquired IBM's blockchain patent portfolio. Apparently it's 680 patent families over 1,000 issued patents. Circle now is the largest blockchain patent holder in the U.S. And I guess, you know, this is maybe a defensive posture.

You don't hear a lot about patents right now in the blockchain space. And I remember the zero where IBM was just putting out patent after patent. City Group used to do a lot of this. And I never really understood it. But, you know, for this to be a transaction, there has to be some industrial logic. And I'd imagine that Circle has been sued over patents in the past and maybe just feels like this is a, a good thing to have in the war chest. That is the most surprising news of the week.

I never expected that. IBM used to just be all over this with the... They did strawberries on the blockchain. Yeah. I remember that very well. I remember having some meetings with IBM people way back in the day. And it was similar to R3, right? Where they were just kind of talking about how Bitcoin's not going to work,

but our blockchain is going to work. So, story. With a J, do you remember them? Storage. I do. This was like file coin before file coin. Yeah, I think along with Sia, maybe they were before Sia even. This was an ICO back to the day.

Yeah, one of the real original decentralized cloud storage darks. They have filed for Chapter 11 bankruptcy. There have been many such filings in the last month or so. So that's the end of an era for storage. although I think they're continuing. So I don't fully understand these foundations that are filing bankruptcy. I don't really care enough to dive into that, into that morass and understand it.

But I think this will continue as a project from what I can see. Did you read this technical postmortem of the Hugging Face incident where Open AI's model escaped its sandbox and basically hacked into Hugging Face? Yeah, and it seems like this model was undetected for maybe weeks at a time. So it's really quite interesting. The AI people seem to be a lot more scared of this than the rest of us, as my analysis. And you know what gives me comfort is the model did go around hacking half the internet, but all it wanted to do was a cheat on the test.

Yeah. So I think that's actually a point in favor of not being worried. because the model was aligned with its objective, which is to do the eval, whatever it was, right? They wanted to steal the answers to the test, right? It didn't go off and do anything crazy, right? Yeah, it didn't, but I mean, it could have. I actually feel great.

It didn't develop its new independent objectives. It just went about the objective in what we consider to be a bad way. I would say AI policy is going to be one of the most complicated policy areas to work on because you look at the response and hugging face didn't have access to the frontier models in order to defend itself. So it couldn't use like the mythos style models. And so they had to end up relying on Chinese open source models, which were close enough to the frontier, I guess, to defend themselves. So that's a whole issue in and of itself, right? Like why can't companies in the United States have access to the frontier models for cybersecurity?

Yeah. So I felt very old today because a blog post I wrote is eight years old as of today. Oh. It's quite well known. Visions of Bitcoin. Visions of Bitcoin. Yeah. This is a big one. So that was eight years ago. That was in July, 2018. How does it hold up? I think it held up okay. It would have been really useful to have AI back then because I think I made some of the images with Excel.

that's how I used to make pictures. And, you know, the point was, I wrote this with Haseu. The point was that Bitcoin has perceived in different ways over time. And at that point, it had only been nine years of Bitcoin. And I was arguing that most people considered Bitcoin to be this censorship, persistent e-gold rather than a payment's network. And I think if you pulled the chart forward to today, it would look different again. But I don't think it held it up.

Okay. I was just feeling very old remembering of that. Yeah, I remember when that came out. What would you say the vision of Bitcoin is today for most people? Well, I would overweight the financialized. So I had a bunch of different categories. One of them was uncorrelated financial asset. I would say that's the most popular one, but it's not uncorrelated.

seems to be rather correlated. It's definitely less of a payments network. Despite all the effort that's gone into lightning, I think we need a post-mortem of lightning. It's certainly not really a reserve currency for crypto anymore. At the time when I wrote this, I said it was declining in usage as a programmable shared database, which I think is, that's basically nil. And virtually no use.

Yeah, virtually no use is a cheap payments network today. Yeah. Which is fine. You know, I think what I think you really need to see something as it really is rather than imposing your own normative lens on it. Yeah. And because then you get in trouble. If you, if you can't perceive the things, it really is. So I think it's okay for people to have different opinions about what Bitcoin is or should be.

I mean, that's the whole thrust of the filterist debate. People have totally different opinions of what Pick one should be. but you get in trouble when you substitute the normative for the objective. And you lose sight of what it is. Are you paying any attention to this e-cash fork? No, I mean, I think it's basically impossible to change Bitcoin these days. So whether it's e-cash or Bit 110 or, I mean, absent some kind of enormous effort, I don't think Bitcoin can change.

Well, it's going to have to change. change for post-quantum, but maybe that'll be the last change. That's the only one. Yeah. And I think if it wasn't for quantum, Bitcoin would never change again. Yeah. Ever. All right.

So I think that is it for the week. We will get this one up. Sorry for being late. And I guess we'll have some more Leo updates next week. Maybe he'll just run it right back up. You never know. I'm still long Leo intellectually. That is, unfortunately, not an LMP.

I think he's going to have a second act here. All right, everyone, that is it. Everybody have a safe and healthy weekend, and we will see you.