Blockstream Liquid's $320M Hack, Open Source AI Summit Takeaways, Vora Reveal

0:00Special Thursday edition of the Proceeding. Thursday night. Not football. Not the net, but Thursday night. PBJ. An emergency jam. And the emergency jam. And the emergency is we're all here. And it's good. We'll be talking about it. Well, we missed last week. Exactly. We missed last week. Yeah. We're missing the show. So we're here. And then we are at the Procedio Bitcoin open source AI summit. Yep. Today. Yeah. Tomorrow. Yeah. No jam tomorrow, but we're like, hey, let's do it. Let's do it. Let's do it. Evening. Evening jam. That's right. And DK's in town. So we got to take advantage to be back next week too. Oh, yeah. Give him better. All right. So watch makes for me. So you better go. There we go. Yeah. I'm here. I'm back to. Yeah. And we're going to try to keep it to. Well, maybe 47 minutes till six or so. Yeah. Anyway, we have a lot to cover though. Yeah. I want to talk about today's event. Oh, yeah. I want to touch on block stream liquid a bit because that was a huge hack.

1:04And I think it's interesting. I'm dynamic. There's some, there's some lessons learned there and some interesting conversation about that. You want to start there? You want to start with? Yeah. And then let's time boxes. I would spend the whole show talking about block stream liquid. But last weekend, Labor Day weekend. There. So for those that don't know what block stream liquid is, it is a. I don't know. It's a side chain. But it's a system in which you can deposit Bitcoin into it. It's a federated system where I believe they have like 15 different entities. Each hold a key 11 of the 15 have to sign to spend out of that system. So do you know who are the people who like who wants to put money in there? Like what are the other certain use cases where that kind of thing? Yeah. Confidential transactions. So if privacy, like really good privacy is important to you, that would be a reason you use it. Also just for you need your counterpart of your paying to also be using that.

2:06Is that right? Yeah. Yeah. So when you deposit your, when you deposit real Bitcoin into liquid, you get LLBDC or liquid BTC token, which then you can do things within that system. And then whenever you want out, you can exit as long as the 11 of the 15 entities in the federation, sign your transaction and then you can exit. I think there's some stablecoins on there. I think like Aqua wallet and Mano wallet and a few other wallets used it as well as like a scaling solution. You know, reach more users, lower fees, etc. And if you can hook up, for example, like lightning on top of liquid as well. So you can add potentially swaps between Bitcoin and Lightning, the other Lightning Network too. I don't know. It's a bit of a product block that's been working on for literally like 10 years. It's also like kind of a test bed of new Bitcoin concepts that haven't yet gotten consensus

3:07or enough traction. Since lock stream controls, the liquid system, they're able to add new features. So it's kind of an interesting, experimental test bed as well. In any case, I didn't know much money was in liquid, but I've now. We know now. How much money was not in? There's like 4,000. There's like 4,000. Around 4,200 BTC in liquid. Yeah. It's something like $350 to $400 million of value in liquid. So substantial money. It's more than I thought, too, actually. And then so the event last weekend, all of a sudden there are like red alerts on social media saying that 4,000 Bitcoin were withdrawn from liquid to one address, to one actor, which was certainly not an expected transaction to occur. So then it's like, oh my god, there's been a theft. How did it occur? And so what we know now is there was a vulnerability in the consensus algorithm that allowed for inflation.

4:18So LBTC was created that shouldn't have been created. And this attacker learned of this created 4,000 LBTC out of thin air then formed a valid transaction. So the designers, the 11 designers had to sign it are like, look at it. And the software looked at it. It's like it's valid. And so it exited. And then it exited to, the money can only exit to certain entities. I think the same or some either a super set of the same entities that are the functionaries of this federation. What I've learned is I think many or most have certain policies in place when it is withdrawn. Like there might be velocity controls where you can only withdraw so much out of there. I mean, once it's gone out of liquid into their own Bitcoin hot wallet or cold wallet, there might be velocity controls on how fast it can be withdrawn out of that wallet to the ultimate.

5:27And the standards themselves are some sort of like a financial services or a Bitcoin focus financial service like an exchange. In theory, one thing though, it's never public. It's not public. Yeah, they used to be public, but they're not public. So we actually don't know. I mean, like you can guess a lot of the major players, but you can guess like probably tell there was one and maybe Breeze on. I don't know. Like we don't know who they are. That was intentional. I mean, there's a trade off there. I think they want to reduce regulatory exposure. That's why they wanted to be more on the down low. But then as a user of the system, it's way less transparent. I mean, like in the most adversarial minded case, it could just be like block stream. It's all 15 keys. It's just one. You know, so I mean, I don't think that's actually the case, but you know, it could be. It could be. So anyway, but it turns out there was a service called SideSwap, which was one of the functionalaries. And they did not implement KYC or velocity controls.

6:27So if you use their service to swap out, then you withdraw LBDC to real Bitcoin through the 11 and that's an 11 and 15 multi sig. Then that went into a side swap wallet. And then from there, you can withdraw to your own wallet and that all that can be done instantly. No velocity controls, no KYC. So that's the path the attacker took to steal these Bitcoin. Now, I'm calling it stealing and theft. This attacker left a message in the Bitcoin blockchain claiming that they're a white hat hacker initially. Was that that was with the move of the 4000? Yeah, when this had exact white head, we're just here to help. I'm white hat hack. I think it was white hat hacker. So it was a guy, but definitely white hat. Yeah. And at the time, like, no one knew, right? All we knew is that 4000 Bitcoin were just withdrawn, holy shit. And it says white hat hacker.

7:28And then what transpired was fascinating because then there was exchange back and forth over the Bitcoin blockchain in operaturns between block stream and this attacker to communicate. And some of that was in plain text and some of it was encrypted where only, you know, those two parties could see the messages. So we don't know what everything that was exchanged, but to cut a long story short, the attacker did return 3,400 BTC. So a huge portion, but still kept 15% or 600 BTC. And you know, because a lot of the communication we can't see is unclear if that's just like the next cup of the negotiation or from the attacker's perspective, that's just like. That's their claiming that's their $10 million about. And as far as I know, that's sort of where things stand.

8:28And there's that drama back and forth know between them. Yeah. And then it's also interesting to see the the chattering commentary on the situation. Like there's some people on Twitter and in the Bitcoin space that that cheer this person on and don't think it's criminal and they endorse it. And I find that difficult to swallow. Like if you if you if you take someone's property and then you decide that you're going to keep 15% and give it back to them, like because they didn't like have the most superior lock on their door or whatever. Sorry, but that's still property by property rights violation and theft. You know, I think they went far beyond what any sensible definition of white hat hacker is and just from a purely like ethical standpoint. Yeah, the legal system can deal with as well. And I'm quite certain the legal system would call that a criminal case. Yeah, I mean, I'm curious. I think the sort of the other side would be and we don't know because a lot of this was encrypted communications.

9:31But if I guess liquid or a block stream agreed to that, those numbers now one can say, you know, it's one thing to negotiate when when you're sitting there. Hey, I've got all of your money. I'll give you back 85% if you agreed not to prosecute, which again kind of changes it from white to gray because it's I mean the cynical look on that right is, you know, if the hacker didn't have that and they sort of said, hey, I've got a bug bounty ahead of time. Like would you give me, you know, $50 million to not reveal like that's a very different negotiation. But I mean, no, I learned I mean, it would have been my intuition, but I learned also like the criminal case does is not impacted by a really if block stream agrees. Oh, there can still be a criminal case. Interest still theft. But why is that if they gave consent to keep it back or I guess that be a civil case or something. I suspect this is because like when you're under duress like that, I mean, the criminal law is like sort of independent of like the duress of the victim. But then the case that's where that lands now another interesting tidbit is over and you know, we just got done speaking.

10:31We just wrapped up the day one of the summit. Yeah. And a panel that I was on with three other folks, we talked about AI security scanning. Yeah. And and well, it turns out that the red team. So we had CEO Rob Hamilton on my panel super CEO. We love talking about on the show and super CEO doesn't really know how to code. It's getting Rob. But he jokes himself. He's a security professional now at the past six weeks, right? Yeah. Because ever since the cold card incident, he jumped into gear and has been you know, burning the midnight oil doing security scanning of many open source big one projects. Many other big winners joined joined him. So the the there were folks from the red team that found a vulnerability in liquid reported it. And that was the one we it's public information.

11:31That was a vulnerability dating back to like 2018 or 2019. So it's been in the code for a long, long time. And then the fix for that introduced a worse vulnerability. It was like easier to exploit. It was a similar type of so this has to do with the it's actually has to do with the confidential transactions and the range proofs. Anyway, I won't get into the technical details, but all good reason it's not implemented. But the fix didn't or anyway, the fix introduced a worst bug and that's what was exploited. And there's drama and Twitter today. You know, where I think the red team saying they also reported the correct fix. And then there's a dispute on whether that was looked at or not by by block stream. But reported the correct fix ahead of time. Yeah. Or meaning like explicitly noted the wrong fix was a concern.

12:35And here's the right. There's like a second email as he. So we'll see we'll see how that transpires. But that was the actual vulnerability. So again, it gets back to the AI security scanning like AI is able to find these vulnerabilities so so quickly and easily. Another thing that arose from this that I thought was found interesting is like the people were criticizing federations and the value of federations. And it actually got me to thinking in ways I hadn't really thought about before because in general, I just naively thought like, you know, federation gives you more security. I'm not just trusting you, but I'm, you know, it's you you and and a couple others are. Well, even even if it's also me. Let's say that we know that it's distinct entities, right? I mean, even if we let's just hold that as a given. But that's not necessarily the case. So in this situation, there's 15 different entities. But as far as we can tell, you know, they just they would whatever the software releases that block stream produce, they just updated and they didn't like.

13:41And I don't know this is fact, but I'm it appears to be the case. You know, they just like update the software, trust block stream and don't audit it themselves. Right. Don't do any kind of like, yeah, you know, checking there. And so that clearly that there's many aspects of centralization, but that that aspect would then be centralized because block stream. I mean, I ended up being, you know, one developer did the fix one reviewed it got merged and then I, my guess is no one else looked at it. Yeah. And they just sort of accept it as offer. So that seems pretty weak. And then another weakness that was revealed was this. Again, getting back to when Bitcoin's withdrawn from liquid. I learned that it's like best practices that are given to the entities or the functionaries or whatever to do spending velocity controls. Maybe do KYC as well, but like put these protections in place. Controls store.

14:41It goes from liquid. Yeah, but it's like, but it's a holding period in and there's controls in place. Right. And such that, you know, the entire liquidity and liquid can't be exhibited in real time, you know, in one transaction all in one go. And but one of the functionaries site, what wasn't doing that. So there was a weakness there as well. Like, you know, so I don't know what the right answer is. Should the liquid protocol enforces like add spending velocity controls and this. So what should they be that gets tricky. And I mean, you know, adding KYC, I don't think it's probably the right answer to the protocol. So, but it raised that issue as well. And then like I said, some people tweeting comparing like custodians to federations. And, you know, making coming of some formula where like, you know, if like two thirds of the entities in your federation are worse. The lower quality than like one custodian, you're better off with just one custodian.

15:43Yes. Some not exactly what you said in the past, but it's in the right in the same direction is what you said. Yeah. And I actually, I mean, I was going to kind of make that point again. I mean, to me, I think there's a lot of interesting, you know, points in this. But, you know, I actually think sailor actually put that in here that that nice like to the tweet storm. I guess I don't know if it was like the Bitcoin Maximus or more number ever. Sorry, he made the points like, I think we from the early days of Bitcoin, you know, maybe you know, maybe took a lesson that was important, but made it the only lesson, which is not your keys, not your coins. Or like, you know, the right to hold keys is an obligation to hold keys. And I think, you know, the way I'm seeing this move increasingly is like, it's not that like you should never trust a custodian or, you know, take everything yourself or never trust a sidechain. Ever like all side chains are bad or all lightning network nodes are bad or ever. I think the point this all raises to me is like in an open system with free competition and transparency, you should evaluate and really see which custodians encounter parties you trust.

16:44And obviously I'm still saying you should have the right to hold keys if you want, but for most people, which custodians you trust, which side chains, which lightning implementations, which lightning network nodes. And I think one of the big takeaways that's been interesting for me is I think as a community, you know, I don't think it's a bad thing. And look, we're the canary in the coal mine. We're the ones taking the AI stuff first. But I think that a lot of this has come out of sort of like sort of hacker culture side projects, maybe companies that I'd be curious within block stream and not to call them out or never. But like, I'm guessing liquid was not their number one focus at this time. And it's like, you know, if you're a company where maybe you at one point you were focused on something and then, you know, it's not like a fault of your own, but priorities shifted. There's new business opportunities and maybe there's just like one person kind of like left on the maintenance staff. Then, you know, that's still securing serious funds. Then that's a problem. So I think, you know, where I hope all this is going to go in the open system is the services that people using trust. Hopefully there's an easier way for people to, you know, to realize like, cussing with block or using something that came out of blocks off where which has, you know, whatever.

17:49A lot of really cool off qualified people top of the line models is not the same as like the weekend hacker project the story 4000 of your BTC and there's a wide spectrum in between. And the, the, I'm not saying that the block sure is that. But it's the last aspect that got brought up that didn't really get enough discussion. I don't think but that someone raised the point about liability from the entities. So there's 15 different companies. And it's unclear how much liability each of those companies has with this staff. Right. Like could there be lawsuits for people who lost money? And if that were ever the case and it was established that there is liability with those entities, is there enough value in it for them to like would that dissuade anyone from. What's the benefit of being a signatory or whatever? Because there any benefit to running one of those? I mean, mainly that you at least get your sort of vote in terms of like on you, you want to, you might be using the system.

18:54So you wanted to, you want honest actors. So you want to at least be one of them. I mean, Spark has the same. Or you know, Spark is federated. Fetamin is federated. Those two other systems in the ecosystem that have federations. And so with Spark, there's one entity which is light Spark, but then there's Spark operators. And I don't know where they're at now, but they started with two. And now I think they're at three. Maybe they have more now, but there'd be the same question there. You know, why do you want to do it? Now with the way Spark works, like let's say we built a wallet. And we, and it was a Spark based wallet. Our incentive to be an operator in Spark would be that Spark only needs one honest actor to make sure. And so then we just have to trust ourselves, right? And that's easier than trusting a bunch of other companies that you don't have. You're not able to observe their, their, their, you know, right.

19:54So that would be a reason. Also potentially business model. They can cut fees and stuff like that. So they're definitely can be incentives, but it'll be interesting to see if that's tested around liability. Now where do things wind up? Blockstream just, so they paused the liquids. System last weekend. They just started it again today. I think glass I checked with thralls are still not enabled, but the actual blockchain is producing blocks again. Or so this liquid is a block based system. Not block ink, but blocks and transactions. And chain of blocks. And they they started they restarted that and out of back said that it will be one for one. Interesting. So they're in the whole then. So somehow somehow be one for one meaning anybody's going to be made whole. Yeah, because other because with liquid, if if it's a 50% haircut, the way liquid works is is not evenly applied to everyone.

20:56It's a rush for the exits. Like let's say we're all in liquid and the date like the moment they turn that on, it would be a fight for who gets their money out first. Because in the last 15% gets zero. It's like the banking system. So it would like it could be unfair and it would be a very, it would probably be the death of liquid. I mean, we can. It's liquid going to survive anyway. I don't know, but like that would definitely be the death. But no, he knows that it will be one for one to somehow someone plugged it some way. 600 BTC or being plugged apparently independent of whether they get it back from this attack. So yeah, very interesting story. That's liquid story. It's kind of settled and done now. Is are we going to keep following this? There's more stuff. Well, mainly I mean, even through today, there's still offer turn messages going back and forth between. Yeah, with the hacker. Yeah. So like are those 600 Bitcoin returned as this person apprehended? I mean, that's a future thing.

21:56And then who made who made block stream poll? Is it from block stream treasury or individuals or investors like tether has been a historical backer. So, you know, they they'd have they'd have the funds to do it. But what's, you know, what's their incentive if they did? Yeah. And like with, you know, buying it or sorry, Bitfinex had a big hack long ago, like 2015 or 2016. Yeah. And which way more than 4,000 Bitcoin or stolen. I think it was like. There's a lot of 100 or I think it was like 100,000 or 180,000 Bitcoin or something. Because Bitcoin was not, you know, it was like $200 a Bitcoin then. And they got themselves out of that by issuing a token. I actually think it was one of the more, I actually, it's one of the, it's one of the few tokens. It's called the USD. And, uh, I think that's how they're. Yeah. Yeah. This token actually is like a reasonable use case in my opinion.

22:59And that it's, it's almost like a futures contract. I mean, you know, it's kind of like you think they're going to eventually make it whole and I don't. Yeah. Maybe I sell you my, and then I was a depositor. So they gave me some tokens for the, for the haircut I took. And then I, and I'm like, I don't believe in this future. So I sell it to you. And it seems like, and it just, you know, gets around, you can probably do that through some bank dropsy core or regulatory core or they take 10 years. And like, I'm sure this, this I could just do right away. So I don't know. It's actually interesting. Yeah. But anyway, did they, did they get any, is there any kind of fingerprint on the hack that suggests it happened because of a brand new AI model or it was just kind of sitting out there hidden in plain sight and anybody could have done this or. Um, well, that fix that ended up having a worse vulnerability than what it fixed was made open source September 1st. And then September 7th or 8th is when the hacker occurred, I think.

24:01So within a week. So those are the facts. So I don't think there was a new model in that week. I don't think that would explain it. But I've got to believe probably just taking, well, I think something new models and yeah, I mean, there's. Yeah, if there's a lot of, if there's 4,000 Bitcoin Bounties, I'm going to be looking at it and I want to steal it. I'm going to be looking at every new patch from liquid and toss it into my security scanner and then I'm like, okay, I'll take 4,000 Bitcoin. Right. I mean, do you guys have any opinion on the white slash gray hat thing here? I mean, do you think, I mean, it's kind of unprecedented. Like do you have any strong strong feelings on that? Yeah, either. I think this is a gregious and criminal and unethical and should be condemned, not encouraged. Yeah, I think if you're saying you're, I think you can't claim white hat and then decide how much reward you get to keep. Yeah. Like if I, why don't I say I'm a white hat and I'm keeping 99%. Exactly. You're just, you're going to go ravage your restore and I totally agree.

25:04Just to feel man, I mean, someone took the other side, they would be like, well, it just needs to be reasonable, right? So like if they took 1%, then it would be, that was a reasonable bounty. So then then it gets to be a split, split is slippery. It Well doesn't pay you any attention, so I think that's what it looks like when you work in the office and get days off and and they're sorry for a. And then you know OK, what I mean and I guess, if that mole establishing is violation of writing in Wiel, whatever is a division of the terms you guess next, I mean I mean because they're not exactly my major crystal back Shared front. Yeah, they like why are you being so. So it's like you know. So, yeah, to me, it's black and white, not gray on that. But it's actually really odd to me that like, dick pointers who are always about, you know, taxation as theft, it's like, oh, but it's only fine to steal 4,000 Bitcoin and keep 15% of it. That's not dumb, it's a tax.

26:06All right, let's talk about, yeah, let's start about some open source AI stuff, because that was today. So, where do we begin? Yeah, we're doing, there's a two day event in the Presidio hosted by Presidio Bitcoin, the open source AI summit. Yep. First of all, let me give a shout out to the people who have made this happen. Oh, yeah, yeah, yeah. Haley, my team, Austin, who's Spinal Grantee, and then Alex, he leads CB and the three of them work here out of, out of PB. They've done a lot of work. They've done a lot of work. We've put on a lot of great summits, including your type one energy summit, but this one is like, next level, it's bigger. We're not even hosting it here at PB because we have, we had over 130 people today, and then the after tomorrow, we'll have like, over, you know, like around 150 unique people who are coming to the summit. And yeah, today was out of the Golden Gate Club. Beautiful. Beautiful venue, really nice.

27:06And we were able to get like an all-star set of folks as well. I mean, we had talked about in the show maybe a month ago that David Sacks agreed to speak. So I said, I'll have a great conversation with him and we can dive into that more if you'd like. But not, I mean, obviously it's a big name, but we've had many others as well, including tomorrow. I mean, although you guys talk a little bit about the speakers today and what you found interesting. Tomorrow we have like a VP from Hugging Face opening up tomorrow. And then a VP, the guy in charge of Nimo Toronto, the open source model for NVIDIA closes tomorrow. So pretty big deal. It's a really good cast of folks. Yeah, and I was super impressed to see, I mean, just really in the last couple of weeks. Yeah, I mean, it came together and the, you know, it started being, you know, obviously with the good summit, hopefully this is what happens is it starts off, you know, it's like a little bit slow to get going. Then as you get going, you know, everyone wants to end and all of a sudden, sorry, the scare seats are taken, but it was really cool just to see how much demand it was.

28:07It was really cool to see that this summit was a mix. You know, there's people from the, you know, sort of Bitcoin world that are mixing with like the actual, you know, people that are building the open source AI, you know, the, you know, sort of Vanguard here in Silicon Valley. And this is the first event that I've seen where like that was like really mix. And I hope crossover. I hope that's really good stuff. I think the number like two thirds of the attendees are like not Bitcoin. They're AI people even better and a third are Bitcoin, but even the third that are Bitcoin are people who are very much in AI using AI or very AI pill. Yeah. Yeah. So that was cool. Who are some of your like favorite speakers or topics or panels or people you may. I like. Yeah. I want to make sure I get his name right. Is it Ramesh? Nom? Yeah. Yeah. Mesh. Remains. Okay. I liked his, I wasn't familiar with his work, but I guess he's a sci-fi author. So he got three sci-fi books. And I liked the way he kind of showed the transition of what's happening with kind of, you know, I think a lot of time, you know, in 2022, everybody said, oh, open AI, fully central,

29:10they're going to just run the table with this. And he sort of showed how I think in 2022 or so they had like a 20 month lead. And then in I think 2023 or four, they that trunked to five months lead. And now it's basically, you know, every few weeks, it could be like a new leader. I think that's like a really promising idea to have illustrated that way. And especially to do it in, you know, instead of AI being this huge centralizing force, it could be a plurality of AI's. And so I thought it was actually interesting to see him bring that data to the story. And I mean, I'm optimistic. I think I've been optimistic and can see we have to mistake that there's going to be intelligence everywhere. It's not going to be a single monopolistic winner. I don't think, you know, even if there's recursive self improvement, I think there will be, I know Max made disagree, but I think I think there's going to be a lot of AI and it's going to be benefit of humanity to have so many different alternatives. And he created a really compelling case that that is going to be true. Yeah.

30:11So yeah, I thought I totally agree with that. I thought he brought great energy. He opened the event. Yeah. He brought great energy. He also opened. He also opened type one. Yeah. And Naz is probably, I mean, certainly one of my favorite living futures. He was the guy that actually solar-pilled me back in like 2015. He was the first one to really show the Moore's Law of Solar, thinking like 2011, an article for scientific Americans. He's really always ahead of the curve. And I mean, yeah, it's a very hopeful message that I mean, I want to believe I'm getting there, but regardless, it's a hopeful message. Yeah. Well, yeah, we're lucky to, we're fortunate to have him come to our events and help out that way. And just for those listening and you're maybe curious and want to watch the summit, it was recorded. It wasn't live streamed though. So we're, I think it'll probably be next week when all the recordings are made available, but we will be publishing and including remesses. Sweet. We're trying to rush the David Sacks one to get it out by tomorrow midday because I mean,

31:17we're excited to have him be part of the summit. And then, you know, what happened yesterday, right? I was offline. I missed this. I, I, I, it just give me the setup. What happened? Well, so there is this employee at Anthropic who just quit and so X employee just quit Anthropic, tweeted out talking about like he's worried about the end of humanity greater than 10% chance that we're all going to die in the next couple of years. You know, very dramatic standard anthropocard. Very dramatic and apocalyptic message, but with no substance or it's not like, I mean, as David, you didn't reveal any new information. Yeah. I mean, as David Sacks sort of jokes like, I mean, usually if you're a whistleblower, I mean, being framed as a whistleblower, number one, usually the company that you left is not on your side and supporting you. In this case, many anthropocard employees were retweeting it and supporting this apocalyptic message. And also this person only worked their six weeks.

32:19Only worked at Anthropic for six weeks. For six weeks. Yeah. He had like no Twitter history. And I don't know. I don't know how old he is, but I mean, look, he looks quite young. I mean, I think early in, early in his career, Sacks called it a style. And then he's, and then like, and then in the same, or I think he tweeted maybe Tuesday night or something, but then like all day Wednesday, he's on like all these mainstream media interviews. You say he was like on national news and the end Rogan? Rogan covered it. So the Wall Street Journal, or either the New York Times or Wall Street Journal also published an article like 18 minutes before the tweet that was like about or the tweet. Yeah. And then it was Joe Rogan. It was like clearly orchestrated like dozens of politicians retweeted this with their message to their constituents and so on. So clearly, I mean, obviously the stuff that wasn't just spontaneous, I think, but in any case, it created a firestorm. And I mean, the concern here, at least from a, if you're a supporter of open week models and distributed versus centralized, then the concern here is that this is going to create

33:26rapid momentum where Congress acts and introduces legislation and possibly regulatory frameworks that are very draconian, where companies like Anthropocopia and OBI have the, you know, just creates a regular recapture and create the moat to them and everyone else in open models might not even be able to comply. So specifically, I mean, I think it's worth stating it kind of, I don't know, the exact words to use. But if, if regulatory capture, if they see you create a panic, then the regulators say, okay, we have to deal with the panic. We had to deal with panic is to talk to somebody, the leaders of open AI, leaders of cloud, you know, Anthropoc. And then you build legislation and regulation that supports their needs, but then any of the open source stuff is sort of out in the cold. There's nobody to talk to. They're not involved in the conversation. And so I think part of the story he was saying is, let's figure out how to as a kind of open source, open, you know, plurality of models, how do those things sort of have a voice

34:31at the table and sort of avoid the regulatory capture from being able to be completed? Right. Yeah, and I think that's an important point that it doesn't necessarily have to be as draconian as open source as man, but like, you know, for example, if you have like, you know, a thin tack regulation that says we have to speak to your CEO, it's like, well, Bitcoin doesn't have a CEO. So I mean, I'd like to believe that, you know, people are going to see through some of that, but I don't know. I mean, people are really, you know, the sci-op, the orchestration, the media, you can create a lot of confused. And also you, and you already have, you know, it's common, it's, it's, the, the AI sentiment is quite negative already among a lot of mainstream folks. Right. Right. So you have that. And then these extreme messages just confirms their existing priors. Yep. So if it's, if something confirms your priors, you're less inquisitive, you're less like what towards the data? Where's the evidence? That's what I thought. Yeah, that's what, yeah, it's like, yep, that's what I thought. Yeah. AI is bad for us.

35:31So it's, it, it is kind of a setup for a disaster where then, of course, that, I mean, why are the politicians flamming onto this? I actually don't think they really care about AI safety. You're, I think, they were all going to die into years. They want votes and they want donations. Well, I mean, to be fair, just play devil's advocate a little bit on this. I mean, I, I think it's probably true for a lot of them. I do actually believe a lot of people at Anthropic really do believe this. And I think increasingly, if you keep saying that, some of these people, they're telling their story. I do think a lot of people probably are even in, you know, politics, generally afraid of that message. So I mean, I think there's both sides to that. I agree, but, I mean, I agree with that, but I just think that that's lunacy. In the sense that no one, I, no one's answering, what about the Chinese open? Yeah. What? And I asked David that. And that was a great question for me. In fact, I thought that was one of your best questions was like, okay, you know, if we try and ban open source and just see it into the Chinese, then what? And like, I thought his response is also pretty, but I asked him to steal that because I know he doesn't hold that position. Yeah. Well, no, like, steal man, and, you know, he did try to steal man, which was, you know,

36:35they would say that we need, he didn't say it this way, but like sort of a nuclear or non-proliferation treaty with China or something. And I tend to agree with them on this point. Yeah, exactly where he was. He was global pause. I mean, he basis these people are living in, you know, a fantasy land. I tend to agree that there's no chance to like, people are going to agree with pause. Because one data point, you point out. You point out, everyone in a lot of us trying to do the global pause on it. You want? Yeah. Yeah. When before XAI had a seat at the table, he was like, hey, we got us low in there. Let me catch up. Yeah, sorry, man. You're. Yeah. Well, that hurt your fanboyism a little bit there. Hey man, listen, I don't agree on everything. I don't agree on anything. But a, a stat that David mentioned was 80, some survey in China, like 83% of the population is like, you know, has positive sentiment on AI. And I don't know what the stats are in the US, but it's more of the opposite. Right. So if you have a population of over a billion people, 83%, they love pro AI.

37:36And you're, you know, doing well with AI, you see the future of it. Like, are you going to pause because the US has a chat with you? Yeah. No, I think it really, it comes down. And I think that really reinforced for me this idea that China, you know, at writ large one, has seen, you know, a lot of people come out of poverty. They've seen technology be great for them. So they're like, okay, this is going to be more of the same, more welfare, like, no, more economic progress, let's say. But, but I also think it's interesting in like, you know, I'd be curious to see more of like a debate between, you know, Chinese lab leaders and lab leaders in the US, because I really do think there's a fundamental breakdown. And you could be totally cynical and say, no, it's all just like lying, right? It's like, after maybe you could be totally there, they're, you know, really believing it. Maybe it's a mix of both. But it really does seem like there's a fundamental divide of, you know, in China at least from everything I've heard, they see it as a tool. As a new technology, like all these other technologies akin to something like the web or a train or, you know, airplanes, whatever. Whereas, you know, this group of sort of the EA people. And to be fair, I think a lot of them really, truly fundamentally believe it.

38:39I think that this could run away. The Gregu paperclip maximization, you know, biological weapons, whatever, pick your flavor on there. So I'd be curious, like, I would love to see some of these leaders from China or whatever lab people from China and the US actually talk about it and see if they can like move each other's minds at all or not. But that does seem to be a stark difference. Yep. So we're going to do our best to get the David Sacks conversation out by Nadeid Amaro. Ashu, are you doing that work? I think you just got assigned while he's here producing. Shout out to Ashu. Ashu bot. Begin that thing. No, he's like, yeah, you need like three clones or eight to three yourself, Ashu, because he's been doing tons of work for the summit today. Ashu's the man. Then I'm like, yeah, Ashu, can we get this David Sacks video edited and post produced like ASAP? And then, oh, then Ashu, let's do it in prom too, PBJ. Yeah. I was just saying, finding Ashu, like, best thing that ever came out of the summit. Yeah.

39:40So, thank you. Yeah. Let's see. What else about the summit? Either today or is there anything? I guess you're not going to be there tomorrow, but the two of us are. I mean, you had two panels yourself. Yeah. I had two panels. The last one was very interesting. I guess this top of mine for me was on Open Protocols. And so that was super cool because, you know, we talked about Open Protocols for identity, commerce, communications, and memory. And it was cool because I actually had four folks on there. And I got kind of a, it was both a variety of perspectives, but they also had like approached some of these different standards from different angles. So we had Elon Strauss, who's more of a sort of like philosopher guy that's written a lot with Tim Aurelie. And it was good because he brought that perspective of kind of like, you know, free markets, high-tech, move, you know, information processing to the edge. And then we had three builders. And it was cool because, well, three of the four people had actually played with Noster on identity. So, obviously, I turned a bit to a Noster commercial. And the fourth was very open to it. And so it was really cool to hear Tom from Buzz talk about his experience building with Noster why they chose it.

40:42Obviously, Roland from Albee and Task Jewel had a lot of experience. Perhaps the most exciting thing was that this guy Sam Nightwing from News Research, News is, you know, they make Hermes, which is I think the most popular agent right now, even more so than OpenClaw. They're very sort of open sort of open source, open mind, you know, ethos. They haven't made a lot of these decisions yet. So it was interesting for him to get to sort of, you know, learn and kind of ask some questions before and after as well. He seemed really interested in that. The mission open source memory protocol. Oh, yeah. Yeah. I don't know much about it. Yeah, I mean, I, to be frank, don't either. That was something that Elon brought up to me before the panel. And he's been pushing for it. Yeah, you know, I guess they're trying to get some coalition around this. I mean, I don't know the details. Is it a question memory or? Yes. And I think in general, the idea is like you need to be able to have an easy way to export your memory or context wherever you go. The more I thought about it, which is still limited, that's extremely important. Because like that's the way they lock you in. That is the locking, right? So if there's like some kind of easy standard and I like, I do not expect the labs like

41:44to embrace that. But if you can get more of a community sentiment or like more of a sort of showing pointer on how to do that, that to me is probably the most important thing to keep the ecosystem open. But it's too like buzz would be able to maybe even instigate, but certainly use an open source memory protocol. 100% I mean protocol. Yeah. So you guys have talked about it. I don't know. Yeah, I don't know either. I don't know a lot to talk about. But I am ready for our next topic. Oh boy. He's going ahead. But I do think it's interesting. I mean, you guys should chat with Elon I'm glad. I'm also glad they all got to meet each other because like, talk now gets to talk to him a little bit. But I also think they were on the buzz point. They were all very interested because this is one of the first environments where, you know, the Hermes age can talk with the open claw agent. And I was trying to kind of get that point across of like if you have, you know, a forum, like buzz open, you know, that's all based around open protocols. They talked about A to A and how it was easy to, you know, put Hermes in there, even though Hermes hadn't done any work to integrate with buzz. So anyway, so I thought that was super cool. It showed the importance of open protocols. And I listen a lot of it, but not all that.

42:46Did you cover like standard traces, like formats for traces for so that not on that panel no. That's a good question. Yeah. But on the other panel that I did, the, which was kind of focused on the economics of open weights models. And then it really, really we covered a lot of stuff. I just basically asked Christian Catalini who I like a lot. And this guy, John Comkov, who's, you know, really, really fascinating thinker. A lot of thoughts are on autonomous agents. Both of them are very interested in, I guess, that question, how to like, sort of like sign and validate, I guess, traces as well. So yeah, I mean, I think it would be worth going even deeper with them on that. Cool. Should we talk about for a little bit? Let's do it. Yeah. So Vora is a company that operates out of PB. Folks, we know and I guess both were both interested in investors. But Jesse and Eric, they've been, they've been at PB from beginning, I think.

43:50And they did their, they did their product reveal today. I mean, they've described what their product is about for quite a while now, but they actually had hardware in hand and did an Apple-esque reveal, which is kind of fun and cool. But the industrial design, I think, really caught our eyes. Yeah, I forget who all, I mean, I think maybe, oh yeah. And yeah, what, what popped out to you? Well, okay, so first of all, I say, I actually somehow missed it today. I didn't realize when Jesse was gone. So I want to see the video later, but so I want to hear from you guys, but I did get sort of a private reveal yesterday and I'm blown away. And like, you know, Jesse and Eric are obviously very smart and, you know, I know Jesse has a long history of photography. So of course, I was excited, you know, to back them, but it was one thing to hear them talk about. It was another thing to see the finished product and the guy they brought on that did hardware clearly is like a stud. I mean, I was literally, you know, because one question I've always had for them is like, you know, I believe this is going to be very popular. I think obviously Apple was super well positioned. I think obviously Apple is still very well positioned for it.

44:51Going after the consumer market versus enterprise is very challenging, but I do have to say, like, I see more of Jesse's vision now. It was truly like a beautiful piece of wood, like real wood embossed with their cool logo. Like the box, like it was, it wasn't like a tech product. No, it looks like something you want to display in your living room. Like it's like from the 70s, like a speaker or something. It feels natural. Like these guys, the biggest thing I came away thinking is like this team now, um, it's craftsman. Like the way to get really remarkable products is to really give a fuck, like really care. And then also have taste and these guys do. And yeah, I see the vision for like, and for those of you who don't know in case we haven't talked about, imagine a box you plug in at your house, you can run local agents there. There's secure elements on the chip. They showed me the first motherboards. That's so cool. And they showed me the secure elements. There's a honey badger. Yeah. The honey badger on the PCB. Yeah. And they showed me there's two different secure. It's beautiful. And also how quickly they got this stuff done and they were mentioning that obviously

45:53AI is accelerating a lot of the process, but like they got a lot of this stuff like manufactured and like the first run of the motherboards and like only order of months, which would have normally taken, I'm guessing like years in a past life. Um, so yeah, I'm really excited. What are you saying? I mean, so it's hard to run. Yeah. You buy it. You as a consumer, you buy it, put it in your home, plug it in, and it gives you privacy. I know they put a lot of attention and detail in the security as well, like protecting your problems injection attacks, which is very difficult problem to solve. Um, and then, um, that's, that's the value proposition. Do you, do you use it basically like a web server to access almost like a, like a chat GPT or a cloud like interface directly from that as a web server, and then that has its own kind of GPU inside? Yeah. I don't know if you like SSH into it or it's a web server or not. Yeah. I mean, I, I'm not sure I'm all the standards, but I think the ultimate thing is you have on your phone that connects, I don't know if it's SSH or not. But the one key thing I would call out is, um, and actually, we curious to get feedback

46:54from the audience. Right now, it's a very capable machine that you can run, you know, a lot of good models on. Um, but the idea is not just that you do everything locally. The idea is you can sort of put a barrier between what you want to be stored locally for your memory and everything and, you know, outside calls. So you could still use this thing to call an instance of cloud or an instance of code acts or something else. And then they're really trying to be, I guess, the protective membrane between those. Um, I do think, though, and I'm, I'm kind of like maybe pushing them a little bit, um, curious y'all's take, um, you know, they, they kind of want to go more mass market. I'm telling them, you know, it'd be nice to even consider starting with like a super premium product with like max out the RAM, like let it be run the juiciest, beefiest models possible, um, charge a higher. I think the product they revealed that a 32 gigabyte. 64. Is 64? Yeah. I mean, you know, uh, you know, did they show you the light too? Yeah. Yeah, light on the front. It's sick. Yeah, it's so well done. I mean, like, yeah, people listening to this are probably like, yeah, it has a light

47:54in it. But you just have to see it. There's a wood. It's a very warm, welcoming light. Um, I forget exactly how they described it, but they discovered it in really what's like a natural kind of aesthetic, right? Yeah. It's anti-blue light. It was like kind of red. Uh, yeah, A plus on the industrial design. Unbelievable. It is super easy to use, like plug it in and it works. Um, I don't know yet, but hopefully that's the case. And, um, I can't wait to use it. I mean, I'm also interested in whatever we can do with pre orders, if you need more spend, do they announce chips here at price point? Well, I have some ideas. I don't know if that's supposed to be public, but I, again, I told them if you could do more expensive version to get it out faster for the first run, uh, says you can reserve it for $10. Yeah. I'm talking about that. That should be changed. I don't know. But, um, Vora.io is a website, V-O-R-A.io. Intributed to like a mesh L-O-M. So that's where the ultimate vision is, right? Uh-huh. So I think sort of phase zero as you these beautiful machines, um, but something that, you know, we've jammed a lot on over the years is like once you have that sort of infrastructure,

48:57now what can you build on top? If only the people working on the mesh L-O-M would want to add Bitcoin, and if only the people working on the Vora would want to add Bitcoin. Yeah. I didn't. If only. Um, which is one benefit of Bitcoiners like immersing themselves in the AI world. Yeah. But I don't know. I mean, Eric was telling me they're kind of leaning towards solar right now. So I was, uh, Z, Z-Cash. Z-Cash. Anyway, all right. Please, um, this is the shortest episode we've ever done. I'm actually super happy we did it. We have us. We come over to come. We're open to come. Yeah. We probably didn't have a huge live audience given the odd timing, but hopefully people will accept that later. Yeah. And, um, we can replay it tomorrow. Yeah. Yeah. Exactly. We're in a back. Yeah. Anyway, good shot, guys. Yeah. Yeah. It's fun. We'll drop it. Till next time.