Kalshi Cooked, Bitget Hacked + Agentic Bank Runs

Transcript

Oh, North Korea's in here. We're going to shut this off. 387 million. >> Tarun is the 10th most famous Delaware resident. >> Consensus is a big boy. >> I never swear on the chopping block. Hey, this looks cooked. >> Agentic bank runs. >> Your estate is paying for dialogue. Homoeconomicus is back, baby. >> Not a dividend. >> It's a tale of two. >> Now your losses are on someone else's balance sheet. >> Generally speaking, airdrops are kind of pointless. Anyways, unders trading firms who are very involved. Polic is the ultimate DeFi protocols are the antidote to this problem.

>> Hello everybody. Welcome to Chopping Block. Every couple weeks, the four of us get together and give the industry insiders perspective on the crypto topics of the day. Quick intros. First, you got Tarun, the gig of brain and grand puba of Gauntlet. >> Yo, >> then we've got Robert, crypto connoisseur and Zar of Superstate. >> Good morning everybody. >> I am Steve, the head hype man at Dragonfly. But we're early stage investors in crypto, but I want to caveat that nothing we say here is investment advice, legal advice, or even life advice. Lisy chopping block.xyz for more disclosures. So, we are together here in Amsterdam in studio and you know, we usually try to make sure we have four people in the show, but uh we have this very interesting futuristic self-s served podcast.

>> It's an automated robot booth. >> Yeah, it's quite awesome. Very touch. >> This is how AI is going to get all our information from us. >> They're going to make the walk into the pod talking. Yeah, that's smart. That's smart. Well, but it's also like look, I mean, the jobs have already been displaced. Here we are, and we had to talk to Well, we did talk to one human, but technically we wouldn't have to talk to a human. >> It could have been a humanoid robot, like functionally like one day. >> That one day, one day um and so we're here in Amsterdam. Um it's been an interesting week in crypto markets and it's unfortunately been a week of multiple hacks. So, let's talk about the big one. Um Bitgat in which you know, Kavette, we are we are investors. Bit is uh one of the biggest exchanges in the

world, top five in any given day with respect to derivatives volumes. Uh they got hacked about $387 million that was hacked from Bitcat. Uh very likely attributable to the Lazarus group which is a part of North Korea. Bit it seems to have mostly affected their hot wallets/warm wallets. There was some kind of infrastructure linked hack. The details aren't fully clear yet, but uh they stole a bunch of different assets out of these hot wallets. Now, Bit thankfully they are very profitable company. They're quite solvent. They had 460 million set aside in a insurance fund/user protection fund. So, that was enough to fully pay out all the claims. They've already begun withdrawals and so people are starting to get their money out. So,

thankfully it's not a contagion. Everything's okay. The amount of money is a lot less than the amount of profit they made last year. So, they're in good shape and and the the industry is solid. However, it's brought up a lot of conversations about moneyaundering and specifically because the DPRK, which of course took control of these funds and then had to try to get them out, used a bunch of different strategies to try to move these assets into assets that couldn't be frozen. So, the first thing they did, of course, was move the stable coins out of stables and tried to move everything they could onto Ethereum such that it couldn't be interdicted by by uh you know, freezes and so on. Uh but the other thing they did is they ended up using Thor chain. So >> not the first time they have used Thor

chain. >> Correct. Thorchain appears prominently in many of these hacks. And Thor chain of course is a crosschain spot trading protocol and it allows you to trade. Let's say you're sitting in a protocol like for example Arbitum that we've arbitum has been known to freeze itself when their assets hacked. I think Taye, friend of the show, she made the joke that the uh DPRK moved faster to get out of arbitrum than they did to get out of USDC because USDC >> takes USDC at least 6 hours. >> Yeah, they it takes a while. They generally need like a court order or law enforcement uh approved instructions to freeze something whereas arbitum obviously froze a hack in the past. Not to say they would freeze for this, but

there was some fear from the DPRK presumably that like, oh, Arbitum has done this before and we need to get assets off arbitum as fast as possible. So, the criticism has been of Thor Chain that Thor Chain basically allows anybody to move any money and Thor chain volume spikes anytime there's a massive hack like this. Uh, and so many people pointing fingers at them and saying, "Hey, there's a little bit of a decentralization theater going on in ThorChain in doing this." The other player is near. We're in investors and holders of near. >> So are robot ventures.

>> Okay. Yeah. And near also has this thing called near intense which is a crosschain trading protocol and in principle similar intentbased trading to whatchain does. However, near has a protocol called shield which sits at the validator layer which basically looks for anomalies as well as looking for reports of hacks and um you know kind of analysis type signals and so on. uh and will refuse to trade with people or sometimes even stop people who are um demonstrated to be high risk and therefore NAR does not want to facilitate that kind of activity and they end up the freezing uh I think 500k worth of assets as well as refusing to trade with I think over 100 million that was trying to get routed through near

intense. So the conversation that has arisen is is this decentralized? Is this good? Is this bad? Is store chain liable? Is is near not li you know how do we think about this? And I'm sure you guys have seen some of these conversations. How do you guys react to the crosschain movement story here? >> Well, I think the interesting thing for Thor Chain is when they got hacked, they shut down >> I think for like a month. >> Yeah. So it's sort of like a but to their credit I mean they are realistically if they had more validators it would actually be very hard to shut down. They just like have had some historical hacks that have made people lose confidence and they've kind of such a small validator set that it's

not the it's not that hard to revert. I think the near model is probably the best model of like balancing being as clo you know more decentralized but providing some censorship when it's assets that there's kind of a valid reason for stopping and I don't I I think there's obviously everyone in crypto you know if you've been around for long enough you you you 10 years ago you're like I want everything always fully decentralized and over time you realize there's some trade-off for that. And I think the main thing is as long as there exists alternatives where you could use the decentralized thing, but maybe you have to pay a worse cost, like worse

liquidity, it's hard to move larger size, whatever, that seems to be the trade-off people have been kind of as the anti- North Korea kind of thing. And that that's sort of what you kind of could argue you you basically see with with Near versus Thor Chain, where Thor Chain is just like a bit of a free-for-all. I think I think if Thor Chain was credibly decentralized and neutral and didn't just like shut down last time, it would make more sense, but their history is not exactly favorable. >> Yeah, >> this argument >> my my mental framing is that you can have the two extremes and you should have the two extremes and nothing in the middle. Namely, you should have completely

neutral, censorship resistant infrastructure that is completely blind to what's happening on top of it, right? In the same way that like the internet itself, you know, doesn't control the packets that are sent over it, right? It doesn't censor them. I think that's a great place to potentially be. I also think the other end of the spectrum is something that there is the ability to censor transactions very deliberately. You're actively trying to scan a blacklist. You're act you're actively trying to prevent fraud, laundering, everything bad. I think the middle ground doesn't work and is in nobody's favor. And so I think you're seeing this in policy arguments. I think you're seeing this, you know.

>> Wait, so what's the what's the middle ground? Well, the middle grounds are things where it's like I don't know, we could shut down. We could revoke it, but there's an active decision that like there's not necessarily even a policy force. Like Thor Chain probably exists in the middle. They did shut down previously. They're not shutting down here. It's actually a decision. >> But to be clear, they shut down when they got hacked. >> Yes. >> So, you know, this is like, you know, it's it's it's a bad money flowing through. >> But I think that's his point. The point is like if you're really decentralized, a you can't shut down.

>> You can't shut down. >> So there's no shutting down in a response hack. And then >> I mean, okay, I think I think this is a little bit simplified, right? Like >> I'm deliberately simplifying. >> Yeah. >> For a philosophical discussion. >> Sure. Sure. But I think well look, we can ground the discussion in like the realities of what we know now it's like to run a network, which is that okay, the validators sit in a Discord or they monitor a Twitter channel or whatever, right? >> I would say Ethereum is on the end of it can't shut down. It doesn't shut down. There's no if there was this activity occurring on the L1 itself, there's no conversation to be had about like who slash what slashh how to shut down the network.

>> I mean, why did they use tornado on Ethereum, right? Like there's a reason. >> For sure. For sure. For sure. Yeah. So, there's there's practical there's sort of impracticalities of like, hey, we're not going to be able to shut down. I think for most networks besides Ethereum, it's a shutting down is obviously not normal behavior. And like in the case of Thor chain, something was so I don't know what the details were, but it was so fundamentally broken that they had to bring the whole thing down. >> And let's use Arbitum as a very similar example. Arbitum just a few months ago, there was a similar issue, right? >> Yeah. >> Stolen funds. Arbitum said, "Hey, we haven't done this before. We are going to seize those stolen funds." The network itself took those transactions, put it into the governance bucket or whatever it was.

>> Yeah. Right. >> Cuz you think that is the middle ground. >> Yes. and that that should not be happening. >> Correct. That that middle ground is when a couple humans are saying like should we do this, should we not? What will the community think? Right? How will people react? They balance the moral trade-offs of it and they like pick an outcome. Either you should I don't think you should be picking at all. Either you shouldn't be able to pick or you should be always picking for prevention. >> Okay. Interesting. I I think I Okay, so I think Near is actually in that middle ground.

the the one end is not near. The one end is a centralized exchange. In the centralized exchange, you know who everybody is. Everything kind of you know everybody can exit the exchange at your discretion. Every withdrawal is subject to your approval and there is no self-custody, right? That is centralized exchange land and we kind of know what guarantees are to be expected when you use a centralized exchange. Other side is unis swap. There is no universe in which unis swap can decide, oh, North Korea's in here. We're going to shut this off. Correct. Right. That literally not even possible. completely that is on the there's no discretion >> there's no discretion right exactly so I think everything that we're talking about is somewhere in the middle and there different points on the middle you know okay Nero maybe is here and Thor chain maybe is over here and my claim is

that if you're going to be in the middle you should do so in a way that and that that basically picks a set of values that are that are positive for the industry and on some level like it is useful to pick values that are potentially even distinct from what other people are choosing Right? It's a little bit like you know we think of states as laboratories of democracy. They can all do different things and people can decide where to go. In the same way you know near might make different decisions than some other protocol that decides to be a little more aggressive or a little bit less aggressive. And that ultimately it does create choice among users of like which kind of protocols do you want to use and support.

That said obviously it's going to be true that North Korea is going to choose the most permissive protocol at any given time >> always. um pretty much yeah that that subject to liquidity and and the ability to actually move the money there. So I see both sides like I do think look if you're going to say that something like near because many people took this tac is to say why isn't near like unis swap near should just be look if you're not going to be a centralized exchange just be unis swap and close your eyes and never look at anything and my view is like look it we the enforcement of property rights is not just to say that every state transition is valid that is one form of property rights that you can guarantee

is everything that's a valid state transition, I'm going to allow it to pass. Right? It's sort of the most basic form of property rights, which is state shows up, you have this purse, therefore it's yours, right? Okay. Another form of property rights is to actually look at what happened and have police show up and interview you and say, hey, who is did you steal that purse from that woman? If so, like you don't get to keep that purse. That is a fuller conception of property rights, right? And so I think any robust conception of property rights that that does not belong to North Korea, even if it's a valid state transition, it doesn't belong to you. And there are times when it's very easy to adjudicate using anomaly detection and AI to be able to say, hey, that was sitting in this protocol 10 minutes ago

and now it's all living on your balance. You should be able to detect that programmatically without having to say code is law. This is not an EVM state transition that we approve of. And I think you can do that in a way that is genuinely neutral and programmatic in the sense that we think blockchain is supposed to be. >> Yeah. So I think one kind of complication here in your mental model for property rights is you know one of the reasons people really use near intense is the crosschain aspect right and the crosschain world you inherently are >> taking two different systems of record that have two different notions of what property rights in the analogy are.

>> Yeah. >> And so inevitably you have this discretion of having some number of operators. it's not like fully anonymous in the same way when you have to translate do this translation between those two and so I think that's why we only see this debate you know you don't see it within a chain as much of a debate right it's like there it's just kind of like okay I used unis swap I swapped from USDC to ETH and I'm the attacker holding a bunch of ETH but >> hey are the ETH LPS responsible no because like they facilitate a transaction that they have no control over yeah >> by being in the pool whereas in the crosschain thing it's like a lot more.

It's a lot different, right? Intents have a solver, there's an RFQ, there's someone who's full >> signing of a message. >> Yeah. Yeah. Yeah. There's like >> based on some parameters and algorithm and >> and somehow that is the that there's something about that line that's like very where it feels like that that where you draw that line is like where all the fights are. >> Yeah. Yeah. Agreed. And like I I agree with you. If near intents were all on one chain, this wouldn't be a discussion because obviously it would be so simple that you know there would be no room for any kind of arbitrary injection of blacklisting or you know whatever anomaly detection or >> but to to take your analogy further I think um a funny version of the world would be you know a lot of countries have

allow you to import property rights from other countries or other parts of the world but only in a limited way in like free ports. or these like places that you don't get taxed or you don't they don't recognize or care what is in that property. And there's some sense in which it would be interesting if like in crypto there's some notion of like opt-in or opt out to those rights and you have to pay more but you get better liquidity for giving you guarantees. But I think that's hib's point which is have a spectrum of control verse lack of control and >> but that the user >> theoretically the user would price them differently for yeah so even in the case of arbitrum I'm actually very supportive of arbitrum having done the arbitrary

state transition quote unquote to like >> wasn't arbitrary >> sure sure yes >> and the community was as well and I think if arbitum had done that in 2023 three, the response from the community would have been much more hostile. >> Totally. Totally. And like the values of crypto have clearly changed. That's true. >> 22 23 post FTX people would have been like a little, you know what I mean? Like >> maybe but like the the whole idea of like, okay, all these blockchains are laboratories of democracy. They each have different expectations and rules. And I think a lot of people will choose like, you know what, if something catastrophic happens on arbitrum, I want to be on the L2 that will do something about it. and the L2 that won't do something about it has demonstrated that they won't do something about it. I

don't want to be there. So, it's a little bit also like look, you know, when the Fed bails out the banking system, it's like, okay, people look at it as immoral hazard, this and that, blah, blah, blah. But people, a lot of people say like, well, look, I want to be in a country that bails out the property, you know, the the banks as opposed to the country that doesn't bail out the banks because those people live in countries with a failed financial system and that's really bad. So I think we will see more of this over time is that people will start to choose chains based on their commitment to different kinds of property rights relative to just we always follow the longest chain rule.

>> By that logic though do you think that users and people will accrete to the chains with the strongest property rights the strongest probability of a roll back a fix a seizing of the Lazarus assets whatever >> not always. I mean, it's a little bit like look, you know, people think of Delaware as having this very robust, predictable kind of set of >> and then they decided they hated Alon Musk and then they shot themselves in the foot by violating their standards and going in a different direction and now all the corporations are abandoning Delaware. Sorry, Nevada. They're going to Texas.

>> Okay, it's it's okay. You know, uh you know, my my state is famous for other things like Hunter Biden, so I'm not exactly sure. A lot of great experts. is the 10th most famous Delaware resident. >> Wait, where are we getting that? >> No. No. I I I don't actually even know who famous Delawareians are. >> It's like Biden's family. Really? >> The Biden family. That's my turn is number Hunters. You got to be number two now. >> Yeah. >> Okay. Interesting. >> Well, okay. So one thing I do think is like you know as we start getting more you know the privacy stuff I think you're seeing a lot more to another aspect of near intense that thorchain didn't have at all was the shielded pool aspect of like being able to do transfers of shielded pools. I think

chain was trying to but it somehow I've never seen any volume on their their usage of it >> but the private there's sort of a question this is sort of maybe again more of a philosophical question but like I have different countries with different property rights and so maybe I have like the Switzerland that gives me the private banking but won't bail me out and then I have the US which not private but I get bailed out. not private meaning like the government sees every transaction over $10,000. they see, you know, like KYC every, you know, right?

>> Yeah. >> And those two worlds interfacing, right, have always in the human world had tons of problems like right of like and so inherently it seems like you have to have that happen there. But the real question to me is like the blockchain world should be all about because it's open. The key thing is like open market so anyone can join and leave. There should be like a pricing difference that is like a premium that you pay for one side or the other which is gets to this point of like will everyone go to the one that gets bailed out or will everyone go to the one that has some other property but they might not they get hacked it's private funds you you're done right like a lot of the Zcash incidents in the past have sort of had that flavor. So I'm just sort of curious

how you think what do you think you why have we not priced that because it hasn't really been priced as like a fear factor except in this like arbitrum case as like a very exceptional thing but if I look at governments around the world deposit bank deposits implicitly pricing this premium for like different places so how when do you think that will ever happen here >> yes I think it's a very complex calculus okay so you know originally laid it out not in the terms of like a blockchain network But in terms of like DEX to sex, right, like Uniswap to Coinbase is kind of the spectrum, which is I think totally correct from like a trading perspective.

I think it's also like Ethereum to I don't know what's the most centralized chain we can think of. >> Do you want to >> uh let's not, you know, let's not make enemies. Let's not make enemies. >> But Ethereum to the most centralized chain that's the most easily rolled back right? >> I think you can go too far, right? I think a lot of consumers are angry at Coinbase because they're like, "Oh, you're overly protective and censoring transactions and like my account got frozen when it never should have." >> And you can go too far in the protection like you can become like >> patronizing in a paternalistic >> paternalistic like centralized system, right? That's what people hate about the existing banking system, right?

>> And so I don't think the optimal point on the curve is like all the way towards protection. I think is probably like I'm making up a number of 80% of the way there. And maybe the optimal state for the user is one in which it's not a KYC chain, right? It's low friction but very high protection. >> But I would argue that the only way a user can realize that optimal point is actually keeping their money on all the different chains and like rebalancing to stick to 80% versus >> way that maximizes the risk that you get screwed. Well, but I think binary, but the property rights thing is like each chain has their own >> rule on that. So you if you wanted to really have >> you're most likely to lose money if you

have money on every chain >> and that's true. And I do think users have like a nonutility nonlinear utility function of like ever losing any money is like extremely >> stressful. Yeah. Once you've been rugged. >> So that then this gets back to this question of like what's the pricing premium you should expect? I think there is already an implicit pricing premium in that like if you have money on some like janky ass, you know, kind of longtail chain. There's less liquidity. It's harder to move your money out. Bridging takes longer. That is a cost that is not doesn't show up in your DeFi llama or in your or not your >> Sure. Sure. Sure. But being the new Calonia of L2s is not exactly like even at the top tier like let's talk the

Zcash versus Arbitum versus Ethereum. Those are all established ecosystems. They have economies. They have GDPs effectively. You know, they're not like >> Zcash isn't really I mean because Zcash is, you know, UTXO coin like it's obviously not going to be like you're not going to stable asset. Yeah. >> You're getting paid for providing liquidity to people who are using the shielded pool. That is a form of economic activity there, right? Effectively. >> Sure. Sure. But how how are you going to get the apples to apples of like the sort of >> Well, that's the Swiss bank. That's your Swiss Switzerland, right? Like imagine there was like no other economy.

>> Well, you'd have to take the same asset. So take USDC. USDC sitting on arbitum versus sitting on, you know, some privacy preserving protocol like some Aztec type thing versus sitting on Ethereum. >> Yeah, maybe we could if Aztec, which I know they just relaunched their whole thing. So that would probably be the >> Let's do rail gun. Take rail gun. >> Rail gun, right? There's not that much of a premium. So I will say one thing about ETH. There is a premium, right? where like ETH that's been within X hops of tornado cash. >> I think this is wrong. I think this is a wrong analysis. I think there is a premium and you would realize the premium if you actually tried to sell in the moment, right? It's a little bit like how you you know you hold bonds and you're kind of like, oh well, I'm going

to mark this to par price because like that's what whatever I'm allowed to do that for my regulators, but if you try to sell it, you're not going to get par, right? In the same way if >> those have to be some pretty funky bonds. >> Yes. Uh well, this was what happened with treasuries, right? No, I mean the treasury market is so deep and liquid. >> No, no, no, no. But I'm talking about when the banks when like SV when SVB what was it? What's the term? >> That was that was a different problem that the problem was they bought all of these bonds that were >> held at cost. Yes. >> And they were allowed to hold them at cost even though the marketing to maturity.

>> The market price was like 93. >> Exactly. Exactly. So that's my point. My point is that if you were distressed and you had to sell a bunch of USDC sitting in rail gun, you would not get part. My point is more actually just like if you look at the premiums right now like the differences they're not that big relative to the real world tract as they shouldn't be because they're arbitrageable within like an hour. >> If USDC was 101 on any of these chains that would be arbitrageed in like two seconds. >> No, no, but we we know that sometimes the shielded USDC does have the have a much higher >> not because like not there's not much >> otherwise anyone is going to go just like mint USDC.

>> Yeah, you can arbitrage this stuff too fast. No, no, no. But but the shielded one, everyone has knows they take this risk cuz like do you remember USDC froze some some uh there is that's what I'm saying like if you look try to analyze this there's something fundamentally different about crypto premiums for this >> this and like I think that's the thing that >> but okay D does not trade at a premium to USDC >> that is a sensor resist sensor >> sometimes it does in hacks actually because the hackers exit to die true >> so it it.

>> I mean, you're just not going to be able to sustainably see that much of a price differential given how fast >> actually it doesn't. They used to mint through the facility and get one, >> but you know, they didn't buy it on a deck. >> I'm trying to remember like what hack it was where it went up to like 125 because like the hacker attack >> a long time ago. >> Must have been a long time ago. >> Two years ago. >> Really? >> Yeah. >> 125. That's crazy. >> Yeah. But it was like a very >> 10-minute spike. But it it was like it was like only because of this one person took the entire stack went immediately and meant to die. >> But the reason I'm bringing this up is >> somehow this gets to this philosophical question, right? Robert's talking about

this being a very binary thing. There's no premium. You can't price this difference. But then in practice, you observe people pricing it as a premium, but then they're also not differentiating it as a premium as much as in Tradfi. And the question is if all these Tradfi assets are coming to crypto, I have the same tokenized stock on imagine there was like a shielded one like rail gun tokenized stock, you know, normal ET tokeniz stock, arbitrum tokenized stock, base tokenized stock, right? They all kind of have slightly different. Do you think that you get this like large spread in premiums like the way you do with an ADR and the underlying or do you think it actually >> No, it should be a small spread. The spread you will see through different versions. I think it would be

convenience based. How easy is >> versus security based? >> Correct. I I think >> I can see that. >> And that's one of the reasons why an ADR is better than the offshore original stock is because it's so much easier to trade, >> right? >> Yeah. >> But the difference also like if you actually had different >> ADR, it trades at such a crazy premium and Korea >> or just like the DRAM ETF. >> Yeah. It's like the convenience layer of letting more people access it. more people are bid >> or at least more dollars >> more dollars >> more scarcity for there's like a fixed supply that is not expanding in response >> so so I I think the one interesting thing is like people should actually study these premium in crypto and try to imply whether

>> implicitly the market is voting on this spectrum thing or binary thing because I think the data actually might you you might have had enough of these attacks now >> to actually separate what people actually view these things as like right infer which one of the two worlds >> well okay interesting let Let's take the other hack that happened this week. Actually, just happened very recently, which is that apparently consensus was hacked in their staking node, their validator. So, this was on Ethereum. Consensus uh I don't know exactly how much stake they manage, but it's fairly significant. They were something like 5% of Lido is delegated to consensus validators. So, they massively uh they basically all at once exited their stake and are rotating the key. So, there's a

lot of ETH now exiting in the exec. didn't really seem to have that dramatic of an effect despite the surface area potentially. Of course, that's because when you run a validator, you can't actually do that much like steal the money. >> Yeah. They don't own the money. The money is delegated to them. So, they're sort of pointing their lasers at this guy is going to validate for me, but the end assets that go that are pointed at the validator are owned by users. So, those users didn't have their keys compromised. Consensus had their validator compromised. Um, so apparently they were able to scoop up supposedly it was just like less than an ETH of unclaimed rewards that were sitting in the validator. Yeah. But because the reward because the validators were offline, there's potentially I I don't understand if this is correct, but uh

supposedly there's up to like 200 ETH in total that is in foregone rewards that otherwise might, you know, consensus might have made had they not had to exit all the stake. That's of course across a lot of ETH that's pointed at them. So, um >> I mean it's it's basically like either you view it as opportunity cost or implicitly like getting slashed if you're like >> I think it's opportunity cost. I think it's like your yield is zero for a little while. >> Yeah. Yeah. But you can also >> slash you can lose your principle. >> Yeah. Yeah. That's true. That's true. >> Right. Slashing like nobody So nobody's lost their principal here. >> No one's lost their principle. But I think like to the end user who's like I was thinking of this asset as like being >> Well, the whole risk-free rate conversation has reared its ugly head. I

think in that like you know the last time this happened I think was kiln which was a very small niche player it was like a decentralized kind of thing consensus is one of the biggest validators right this is infura that you know >> well actually I think during the kil act though kil was a very large portion of salana at that time >> oh is that right >> yeah it was like 3 or 4% of so so I think it that one was more complicated because I think on ETH maybe they're not as big but the other networks they're >> yeah but like you know consensus is a big boy so them getting compromised and again you didn't lose principle. There was no double signing or anything that would cause slashing. But does this impugn the concept of a risk-free rate?

>> Yes. >> Yes, >> of course it does. >> Okay, tell us. >> Only because consensus is probably the most well-known name, not for staking, but within the Ethereum ecosystem, right? I think to your point about the rewards disappear for even a temporary amount of time, right? I think we think of staking as the risk-free rate, but it's really the best possible rate, not the worst possible rate >> or the lowest volatility rate, but this added volatility. >> Correct. This adds volatility to it and it adds, I think, a risk premium in general. Knowing that there are extra surface areas of risk, I think calls into question, is this risk-free?

Assets were not stolen, right? But I think the validator as a piece of the risk surface area hasn't fully been explored. I think people have aggressively explored bridges and you know smart contract risk and all of that. I think this is one of the first you know major validator risks but I don't think >> for Ethereum other networks have other networks have had way more issues. >> Yeah. But, you know, I do think it calls into question what is riskless. We don't know if >> Hey, is the 10-year US Treasury bond right now risk?

>> No, there's CD. >> No, I mean there's CDS that trades on the US, right? That's not zero, right? So like >> wait, what do what do they trade at right now? >> I haven't looked in 10 years. I'd actually I'll pull this up while you guys talk, but it's not >> I mean I mean the real question to me is like this reminds me of like in 2021 there were like a ton of relatively large relative to like the mint and burn fee for tether uh spikes in Tether price mainly because of like the US price and they created this weird arbitrage where like tether would go beyond would stay unarged for a while because you were getting slightly better carry in US and at that time people were treating them at are. And so there was there were all these people I remember

these like Tradfi people who like this is their first time ever getting into crypto and they were like buying US as if it was like $1 always and they were just like oh like can I just buy some CDS on Tether DPEG so that like my US is covered and like there was a time where people were really interested >> people were shorting tether like a lot of >> Exactly. Exactly. There's a lot of I mean it was a slow wid >> but the question is what will that happen here? >> Okay. So I have looked this up while you guys were discussing it. USA CDS is 35 basis points. So the true risk-free rate if you actually wanted to insure a treasury is 35 basis points less than you're getting on treasuries or T bills.

Right. >> How much was lost here actually in basis? If we knew the amount that consensus we could take the 200 divided by their total stack and get get >> probably a lot less than 35 basis points. A whole lot less. Right. like way less. >> So, but you know that's cuz we also presume that the risk-free rate doesn't have any risk in staking. I think if you mentally add 35 basis points of required return, what do you get? You probably get less people staking and >> we definitely get less people looping. The looping would go down quite a bit. >> I mean, but this wouldn't even this would blow out someone who's looping with it.

>> It wouldn't blow them out, but they would definitely be taking >> if you lost all of your rewards for a year. It would. Yeah, sure. Sure. >> Right. If you lost all of your rewards for a month, it might depending on >> or at least you would just lower your leverage ratio, right? >> Yeah. Yeah. Yeah. I guess my thought is like look, when you're staking on average, okay, so consensus had this issue, there's some I don't know how long exactly this is going to take for them to rotate. Let's say >> probably days. >> Uh I think the I think the unstaking is like 4 days and then restaking takes a while. Like entering staking takes longer. >> No, there's seven days or something. >> I don't know exactly. Um >> but it's slightly different for the LSTs because the LSTs guarantee you some

immediate and then they manage that >> comp the LST is the right mental model which is that look if you're staking there's this happened to one staking provider right and they have I don't know let's say 5% I don't know if that's exactly right but let's say 5%. So that means that like as a you're a staker on Ethereum, you should assume that like once every four years maybe there will be one incident. >> The frontier models I think are going to expose a lot more risks. >> You think R's going to go up? >> I think we've had roughly zero risks and now we're in the nonzero territory. So I wouldn't say it's once every four years.

I would almost say it's let's remove three years ago. And >> actually it is crazy. has happened almost exactly four years to the merge. We're within like a couple weeks of the 2022 merge. >> Four year cycle. >> Four year cours. >> But even so, okay, say say that it's this year, right? Okay. So this yearce you have a 5% chance of having your rewards kicked out for let's say a week or a couple weeks, right? >> If you basically multiply that by 5% like that's actually not a material effect.

>> I think you have to anticipate a worse outcome. I think you have to anticipate there's a 0.1% chance that like a validator could somehow lose assets or be slashed maliciously on purpose by an attacker. >> But even if you are double dipsish, right? Like two weeks. Let's just say it's like two week I was rounding up to two weeks saying >> for that 5% >> for the 5%. Yeah. >> That's like one pips for the whole stake. >> Yeah. Exactly. That's nothing. That's nothing. >> No, but it it's like I think it is. >> But you you just said one bip multiplying the two together. >> Yeah. >> You think one bip is not that crazy? >> It's not that crazy. Yeah.

>> Yeah. But >> it's a whole lot less than the 35 basis points of CDS on the US government. >> Exactly. Exactly. Now, I mean to be clear, like if this is something that was relatively mild, obviously a lot worse things could happen. Like if there is actual slashing, somebody takes control over a cluster and then starts double signing, then you could >> settle how would you settle a CDS on this >> because you can't settle it on Ethereum. Actually, >> kind of annoying to settle on Ethereum. How would you settle a fictional Ethereum staking CDS? >> Yeah. Yeah. >> Like OTC contract, whatever. Fine. But I'm kind of I kind of imagine you have two.

>> Well, it wouldn't be a credit it wouldn't be a credit default swap. I think you just have a total return swap on the staking, right? Then it's very easy. It's trivial to settle, right? >> Yeah, that's fair. But it would be kind of funny if you had two POS chains. Imagine like all the Salana ones were settled on. >> You could have a slashing CDS. >> Yeah. Yeah, that one actually did. People have done that. Yeah, that exists. That exists. >> But slashing is easier cuz it's like ver you can verify in the chain later. >> Here you might not be able to verify this claiming thing in the chain because it's like done on the validator side and then pass. >> The easiest way to hedge this would be a total return swap where it doesn't even matter. You don't have to even observe

the cash flows. You just pass them forward. >> Honestly, consensus should just start offering >> new business model. I mean pay back pay back the off by selling them. >> I agree. By the way, I I know we're joking about this, but I I actually want to drill in for a second. I actually agree with you that a staking provider could offer effectively a fixed staking yield that's below the current yield and them taking that risk and ensuring it. >> Right. Consensus has enough resources that they could offer that with confidence and they could say, >> "Hey, we will pay 3% no matter what." I am pretty sure that a lot of these DATs have staking insurance >> but I sort of assume their staking insurance is not from the provider

because they assume it's from Yeah, it's exactly providing it. >> I forget. I know that at least one startup has told me that they're trying to offer this. >> I guess Realm Realm offers this like DeFi insurance for loopers. So like I'm if they're already doing that, I don't see why they wouldn't do that. >> Somebody's probably doing something. But like doesn't Coinbase have staking insurance >> next week on the chopping Next week on the top, we will have more data. >> We will know what we're talking about. >> Yeah. Yeah. Yeah. They do have the omnibus Lloyd's policy, but I thought that was only for custody. I don't know if that Yeah. I don't know if that incorporates taking these are all good questions where we should get guests.

>> Yeah. This would be >> Let's get next week on the >> Okay. Okay. >> But it is kind of fun to think about because I think people for a while everyone was like, "No, no, no, no, no. This could never happen." And I think AI has like changed people's >> No, I think people are underpricing. Would you okay if you had your finger in the wind guess what the true relative to the US now we have the US as a marker and we have this like one basis point lower bound where do you where in between 1 and 35 do you place this >> or you could place it higher >> are we talking specifically about Ethereum validation only >> okay let's do Ethereum and Salana just just so to get an idea of the spread >> I think Ethereum validation I would put

it probably at about five basis points Okay. >> I have no idea how to This feels like a very Robert question. >> No, no, no. But but at least think relative to the US. Let's just take US way lower than US. Way lower than the US because one is a >> cryptographic open- source system. >> Yes. >> That like Yeah. Like we don't really know all the risks and like a frontier model is going to know it better than we will. But like >> also if there's like a catastrophic bug, Ethereum would roll back. You know, if like if like twothirds of the state got slashed >> totally for some like random thing that someone found in like some code obscure code path.

>> We gota get luben on. >> You got to tell us what they think. >> I like if if >> we were just saying that Ethereum is the L1. >> Yeah. Like the real it's very clear the constituents of Ethereum who matter the most are validators. Like Ethereum cares a lot about stakers. Yeah. >> Like abnormally high about stakers relative to users even. So if there was something that went totally wrong that was clearly not user's fault, they would roll back and fix it. >> Well, here's where I agree and disagree. I think they care about validation and stakers the most because they care the most about transaction finality and security. Like that is the root, right?

It's not >> are the Ethereum stakers making money. It's are transactions secure. I think if transactions were corrupted. You think if transactions were not corrupted, but a bunch of people got slashed, they wouldn't do anything? >> I think that's I'm in agreeing agreement. >> You are agreeing with me, correct? That they would do something. >> No, that they wouldn't. >> Oh, I think they would. I think they absolutely would just to protect stakers from an unfair slashing event. >> I don't know. I mean, if it was an opportunity cost staking event, I don't think they would. I think if it was like slashing usually loss of principle, >> if it was significant loss of principle, then yes.

>> Yeah. I think they would totally roll it back and like you can very likely roll it back just on the beacon chain without affecting transactions happening on you know without affecting the state. >> I think if the state was corrupted in any way the odds of some sort of >> all I will say is this is all a great reason that that EIP should not have gotten in anyway which got removed today actually officially officially the inflation >> adjustment. Okay. Uh well actually so speaking of things being removed >> switching gears a little bit. There was a big story about Kulchi friend of the show. We we love love Tom's best friend.

Tom's Tom is a good friend of Khi as we all know. >> Tom's not here. So at least we're reducing the amount of swearing on. >> That's that's true. >> I never swear on the chopping block. >> That's true. We did establish that. We did establish that. >> Data science proves it. >> Yes. Very cool. Um okay. So uh story about Khi. So Khi they offer a purpose product. They were one of the first to get approved by the CFTC. They were the Yeah. Right. They were the first they used that position of privilege. >> Yeah. So, um, crypto allstar named ICO beast was tweeting about Koli and how amazing their volumes are and their market share is. And, uh, this, uh, autistic Swiss quant named Benny uh, replied to him in a fairly >> self-description.

>> Correct. I'm not I'm not imputing that on him. That is in his profile. >> But, do you also personally believe that he's an autistic Swissant? >> I would believe it. I would believe it. I I also think >> well let me finish the story. Let me let me give let me give the frame. Do you think autism is now just like a >> aspirational? I think there's an aspirational. That was what I was >> right. That's true. That is true. That is true. Um so, uh he replied to ICO saying, "Oh, you can't believe that's real." Right? And then ICO was like, "Oh, course. Why are you crying, you little [ __ ] Like, come on, dude. What What's the matter? You're just a little hater." Yeah, exactly. And Benny was like, "Look, if you really want to come at me like this, I'm I'm You really You

don't think that we know that this is fake?" And ICO was like, "Yeah, whatever." And so he was like, "All right, you know what? I'm dropping something. Everyone stand by." And so he dropped this gigantic thread where he basically demonstrated that KCHI was massively inflating their per volume, which I think to anybody who's been looking at the pers on Khi is not that big of a surprise given they have 3 million of open interest and hundreds of millions of dollars of volume in any given day. >> Also, is it's not totally public yet, right? Isn't it like private beta still? >> Is it? >> I I forgot when I don't know if if that changed in that, but >> I don't think that's true. I don't think that's true. At least it's on the front page of call sheets. You can certainly go there if you're if you're using it.

>> But but but even when it was private beta, the numbers looked even weird like >> Right. Right. Right. So so uh Benny not only showed that information but also showed that um basically there were incentives for some market makers to be able to uh basically wash trade. Didn't directly accuse them of wash trading but the effective equivalent of wash trading um in order to make money. >> Incentives such that the market makers >> were incentivized would be following their incentives to wash trade. Uh and then second showed that there was lots of $5,500 that were getting traded repeatedly back and forth all day long visible in the public order books of Kchi's Perth volumes. This led to a very dramatic pylon. Uh I think the entire internet got insensed about this whole uh uh you know kind of autistic guy

showing up this uh $40 billion company. Um this led eventually to them walking it back. I see apologizing publicly. It's like hey it may have come in a little hot. Wait, wait, wait, wait. The ICO beast apology and the retorts were 100% written by Ki's GC. >> Oh, 100%. >> There was no like you look at the like word choice, the jump in. >> Yeah. Yeah. All of a sudden he comes back wearing a suit, you know, sort of. Yeah, exactly. With a bunch of um u what's it called? >> Michael. It was it was it was like it had phrases like I may have misspoken. I was like, okay, come on.

>> My personal opinion. >> Whereas like previously it was like [ __ ] you. You you idiot. >> Yeah, exactly. Well, um this all this drama ended with the CFTC announcing or not announcing, sorry, with it being reported by the Wall Street Journal this whole drama. Then the Wall Street Journal reported that the CFTC was investigating Kolshi and some of their volume claims. And then lastly, I think it was yesterday that Khi announced that they were ending their liquidity incentives program around all of this involved, which presumably should lead to an end of this behavior and maybe not as much volume happening on Koshi Purps. The first thing I would like to say though is that exchanges in general, derivatives exchanges and equity exchanges, right, are not far into the concept of having

market maker incentives, right? When you look at Kali's Tradfy competitors, they all very formally have fee structures to incentivize market makers. They have rebates, right? Rebates are a died in the wool part of traditional exchanges. They are not on their own bad in any sense of the word. I think the question is how did call sheets go so horribly arry. >> Yeah. >> Because this is not new. They didn't invent rebates, right? Every exchange has rebates to incentivize market makers to provide liquidity. Kshi's somehow jumped the shark. So, so I didn't really read theirs, but the pl like I didn't I have no clue how they actually did it, but

one one place in traditional derivatives like CME derivatives where it can go wrong is when it's like usually they try to do some type of like praa incentive, so it's like based on like percentage of volume that you did within some interval. There's a fixed pie and like you did 20%, you get 20% of that fixed pie. If the market's not competitive and there's only like one or two people, they effectively can implicitly collude to try to like maximize their >> each get 50%. >> Yeah, exactly. >> And so in Khi's case, I could imagine that they just copied like some CME thing and but the CME thing is because it's so much more competitive where there's like always hundreds, thousands potentially of market makers.

>> I I think that is a very charitable reading. I'm just I'm just trying to say that's like the most if I were to give them the most, oh, we didn't know we were doing this. That's the only way I could see it happen. >> Candidly, I don't really fault Kali that much for just being like, "Okay, let's like throw some incentives on this thing. We have a lot of money. Let's just like try to win a market." That said, I think they obviously knew that these numbers were implausible. >> How could they not? >> Right. The most junior analyst inside the company looking at whatever graphana they had would have been like, "Hey, this looks cooked."

>> Yeah. Like we have 3 million of open interest and 500 million of volume. Like there's nobody trading here except market makers. >> No. Correct. I'm sure everyone inside that company, >> right? And so it's always a spectrum, right? It's always a spectrum. Like you know, Coinbase International very famously also has a lot of volume relative to very small amount of open interest, which tells you that it's a lot of market maker incentives and not a lot of organic activity. This is the oldest thing in the book, right? We saw this a lot with perexes that were just farming incentives and you know, we we understand this in crypto. But the question is okay when do you decide we are over incentivizing a market that has no demand and basically we are giving the market a false illusion of traction where we don't have it and second are

you advertising that publicly as a way of trying to get people excited about fundraising or you know whatever your next round or blah blah blah >> fundraising for yes so like that is where this steps out of okay this is fair game this is just like part of you know how market markets get off the ground 100% it's like okay if you 3 million of open interest. Do you need this much incentives? And if not, are you scaling it in such a way to create a false impression of usage? >> Again, I haven't dug into the exact incentive structure to run has not either. Yeah, my like two second read and I should have prepared more for this show by going into it a little bit, but my read of it was that it was just a basis point rebate or not base point whatever it was like

>> it was a rebate on each dollar of volume right? >> Yeah. >> Knowing that taker side and all of that was like a positive fee, right? But but also if they don't split it, like if you don't like make people compete for the rebate, you give fixed rebates and then then you just get the crypto the like correct >> the crypto normal wash. >> No. Correct. But I don't think it was like a fixed dollar rebate for the market. It was a >> it should have been so I believe the the structure was that um certain makers were getting.3 bips y >> on making and takers would pay.3 bips. >> Yeah. It should it should have been like >> trade would net to zero but certain market makers were earning equity based on how much volume they were doing and that was probably the incentive.

>> Okay. So it wasn't the per trade fee it was probably the >> basically the per trade fee would allow you to net. >> Yeah. They should have paid anything for that liquidity. They >> in parallel were paying equity for bringing volume to the exchange. >> Correct. Correct. And so basically you have some market makers crossing each other kind of doing these nonsense trying to get trades. Right. Exactly. And the thing is like look, Khi obviously has protection. >> Equity probably looked closer to Terrun's fixed value >> inact. If they said, "Hey, we're going to give half a percent of the equity in KI." >> Exactly. And that equity, >> this is where most of the perplex farming that's gone wrong has happened where like they give their tokenact exactly. And so like I think the um the the other thing too is like almost

certainly Coul Street has detection for like oh there's a bunch of $5,000 lots that are exact same size that are going every day. any anomaly detection protocol. >> I can't fathom that they didn't all know like it's it should have been impossible. anyone inside the company with any access to the data any internal dashboard would have. I mean another interesting thing was just sort of like there because I guess the vitrial was so high. There was also a lot of analysis of some of the prediction markets where there were market maker incentives like the longtail ones that like they lose to poly market like the political ones and whatever >> and um there was also some weird stuff although it's like a little bit hard like the per one is like more just

blatantly obvious in the prediction mark one you have to be a little more careful because like well if I hold the thing to maturity the real payoff is one minus something. So if they're oscillating between 80 cents and 20 cents, then maybe there there's like some economic reason like they're quoting around resolution but >> but there was also some stuff in the prediction market side that looked very sus as well. >> Yeah. I mean, look, there's always going to be some weird stuff going on and like I don't think it's the job of an exchange to parse every single example, right? >> If you're if you're regulated exchange, it is your job.

>> Well, no, you're never going to get fraud or >> No, you're never going to get it be zero, but it is your liability, right? like decentralization. >> Yeah, of course. Of course. And but there's some understanding from regulators that like some stuff is going to get through. You make a best effort. >> Yeah. But not something that's 98% of your volume. >> Exactly. Exactly. 98% of your volume is fake. Like and so you know their defense they published a blog post before they cancelled the program and in their blog post they said this is not fake because people are actually taking and if these guy like these guys are losing money doing this but it's genuine activity because they're losing money and somebody else is making money and that person they >> same two users that are each making and losing the same 55%

>> obviously they don't disclose exactly how all this >> I know. Yeah, that's so farical. I mean >> I mean I don't know but that's obviously their defense. We don't we don't have the data to be able to parse it but the CFDC almost certainly will. >> Well, what's the guy guy's name? Mibo Bino. What? The analyst. The autistic one. >> Oh, Benny. Benny. >> Benny. >> Benny has a >> I mean it is it is in his bio. >> It is in his bio. It is in his bio. So, credit to him. >> If Benny has the data, we all have the data. >> Yeah. I will say hell hath no fury like an autistic scorned uh Twitter [ __ ] poster. Well, I also do think Khu's marketing of it was a little like I feel like I was getting I was constantly

getting sponsored tweets that were like call she's hit X billion dollars of volume. I'm like >> you if you did that you wouldn't have to pay to advertise. They're clearly getting that volume. I mean that's true. It is weird to >> No, no, no. Why are they Yeah. Yeah. Yeah. That that was the part where I was like something is weird about this. >> Yeah. actually that >> like I didn't even look at it but the sponsored ads from these like Koshi influencers I there that were like paid ads I was like something is weird about this >> well I think they were just trying to position it as a mainstream competitor to all of the other exchanges right because I do think >> but when their volume is higher than these established exchanges but their is 100th >> well no but they weren't uh they weren't

advertising per right they were advertising the volume on the core exchange >> no no but they were actually advertising the oh the Bitcoin per that we have has like >> they were advertising the per it was like mentioned in and I was always like why are you advertising like shouldn't this product fund itself like why do you need to if it's really has that much volume and oi people come >> I mean they will be IPOing supposedly next year so or targeting IPO anyway and uh if so maybe it's pre-marketing >> yeah but it's it's like if you think about exchanges they don't have to advertise their volume they advertise like oh our fees went down >> yeah fees are down something something is weird to advertising for volume like that's something I don't Yeah, that is a little that is a little odd if if they're in fact doing that.

Um, okay. So, last story that we want to cover. There was a post quoting a slide by Apollo that showed that Apollo believes that it may be true very soon that muse style or Hermes agent style AI agents may start to cause agentic bank runs. So agentic bankr runs of course implying that well two things I think there was a little bit overloaded term but one version of agentic bankr run >> all you have to say is thorston the head economist at Apollo very good marketer >> okay well >> that's the number one thing I took the marketing of that I mean >> beautiful okay so let me describe what bank run is the assumption is that okay a lot of people are sitting with their cash in JP Morgan savings account or something which means they're getting

paid basis points even though right now the risk zero or they're getting paid effectively zero >> you know what I get paid on my JP Morgan deposit account as an individual that I've had with them for 20 years. >> What's that? >> One basis point. >> You can pay one on savings. >> Savings account. >> Checking. >> On checking on savings, it's like 20 bips or something. >> Okay. >> Yeah, but why the [ __ ] do I have to separate them? >> Well, no, I mean that's like such a that's such antiquated >> boomer question. No, so like I think uh so the idea of an agentic bank run is two two things I think which were overloaded together. one is that your agent, you know, your Muse or your Hermes agent, whatever, >> switch you to the better product.

>> It will switch you to, you know, let's put you into treasuries and let's basically remove all the cash sitting in your sweep account. Uh, if you're sitting in like, you know, fidelity, >> it's too difficult for you as a human to move money out of checking into tea bills. >> Yeah. >> Or back again when you need to pay your credit card bill or whatever. It would take too many hours of your day. An agent can super optimize >> optimize away your laziness, >> right? Laziness based business model. Nobody should have money in a checking account anymore. The one basis point you're earning is insufficient. >> Right.

>> AI agents will optimize your life and your portfolio. >> Yes. So the second aspect though of the agenic bank run is like actually because that's not a bank run. That's just like okay well now they need >> rebalance to a better product. Yeah. >> Well yeah. Well well that too but also like okay if there is a bank run being experienced by JP Morgan they just need to raise their rates and they'll keep the money. Yes. >> And then that's just okay. Higher cost of capital, lower return on equity, bank. >> I think all of this is just like basically they're learning about DeFi because it's like, oh, someone changed the [ __ ] rate in a lending protocol by 10 basis points. All the loopers like exited from one protocol, moved to this one. It's the same thing. We've already

had cryptos actually already used to this. People moving between exchanges, same thing. I think it's just more the trady people are just like not are afraid >> for consumers in for consumers >> but even consumers some of the new DeFi wallets will like auto loop between protocols and you don't even like touch it >> but I mean I mean for uh the f the finance world with respect to consumers has always assumed sticky deposits and basically non-responsiveness to >> the entire product is built on the laziness of consumers right the reason >> why rates are zero is because they have actual hard evidence that their depositors will not leave. It is too burdensome. Right? This was I think I mean I worked at a bank in 2007 when

like it all just fell off the cliff and rates went to zero. Zerp for the first time in history. Right? The first time in history rates went to zero. The entire conversation was centered around what is the consumer response when rates rebound. And inside the banking system, nobody knew what the stickiness of deposits looked like and the oligopoulist competition of banks to change deposit rates looked like when rates rebounded. And guess what happened? We had never in the history of the world had 0% interest rates. Rates went to zero and on the deposit side, they never went back up again. They never bounced back up.

>> And I think everybody was surprised by just how sticky deposits were, right? We've seen this, you know, recently in the Clarity Act with all these banks, you know, arguing that like stable coins are going to make everyone leave their banks. No, they're not. Like banks have been paying zero. >> Yeah. Like why isn't Mercury making everyone leave their bank? >> Yeah. Correct. >> Or column or dot dot dot. Right. >> Yeah. >> Right. So the sensationalist part of this research piece that I like is the idea that yes, consumers are incredibly lazy. They won't do this on their own. That is true. The entire profitability of the banking sector and the net interest margin is based off of this reality, right? Fed's raising rates, amazing. They're like profit margin just

went up because they're not passing that through on the deposit side, right? Asset side benefit, liability side isn't hurt by this. I do think it's true that humans are lazy, right? And I do think that agents will solve this for people. The thing that I'm not worried about from this agentic bank run is that this will be gradual, right? Like the speed at which your average American is going to be handing over their like portfolio and banking information to their agent to say like please handle this for me is going to be much slower than people anticipate. I think the average 50-year-old in America is terrified of this. I actually think another reason

it's gradual is all this these trillions of dollars that are going into estates from like boomers to their kids that money will just be like flowing incrementally and like okay yes the kids might be more techsavvy and like go do this type of stuff but the the retirees are not really >> well it's going to be it's going to be it's going to be gradual and I think everyone's going to have a huge amount of time to react and to adjust it's not going to look like a bank run which implies sudden Right? This is going to be a very gradual tectonic shift. >> Well, here's what I would argue as well.

Like, so I recently got set up with like if you if you take, you know, column or mercury, >> all of these or Arabore for that matter, all these new generation banks, they have MCP hooks, they have APIs, you can literally connect them up to your agent and give like very fine grain permissioning. >> JP Morgan is going to fight that to the end of their days because they know exactly what it's going to entail. >> Yeah. Which means that like >> they fought Plaid, >> right? >> Which was read only for like 10 years. >> Totally. Totally. >> Oh, you can't even get readon access to your account.

>> Totally. And so I think it wasn't Amazon that blocked Muse. Yes. >> From being able to buy on Amazon. So like I think you're going to see this. >> Amazon is incentivized for Muse to be shop shopping spree. >> No, no, no. Absolutely not. Absolutely not. Because Amazon makes most of their money from advertising. >> Yeah. And not most, but a large >> I think it's most No, no, I think it's most most of their margin. >> Oh, most. Yeah. What's their margin? I think gross revenue is not it's high. >> That's like super super low margin when >> I was not as familiar with Amazon's margin. >> Yeah. Exactly. So like Amazon I mean you go to you search for anything on Amazon the first [ __ ] 16 things are are sponsored.

>> I don't shop on Amazon so I don't know. >> Okay. Well I'll tell you there there first 16 are sponsored links. There's an auction going on for the thing that you're searching for. They're no longer a search engine. They're purely >> it's Adwords for products. >> It's Adwords for products. And as a result the positioning is how they make all their money from the merchant. To be fair, this is true for like all these companies. Door Dash, Instacart, Uber Eats, like they all are doing a lot of them all have the same. >> Yeah. But it's nowhere near the same as Amazon would expect. They're sorting their business model. >> I think that's why Door Dash Amazon's business model breaks. Door Dash is not. >> But this is why Door Dash is like pivoting. We're going to be like agent friendly whatever.

>> Right. Which which tells you the opposite is that like the advertising is a small part of their business. Fulfillment like they $12 to every >> Yeah, exactly. They have they have significant search charges that they throw on top of these things. Whereas Amazon has the opposite. Amazon is like dirt cheap. It's like we'll get this to you tomorrow morning. You know, everything is going to be free and bundled with Prime. And the way they make that work is they're charging the merchants via placement in the search results, which is a beautiful business model. >> So, I will tell you one thing. Tell us two things. The the one my the only task I've really done in Muse because I felt more confident about giving it some account information >> is to basically go through all of my historical health insurance >> uh co-pays and figure out if like there

ever any doctors or dentists cuz like >> if you have the account information to my insurance company, the only thing you really do is file claims. You can't really like get money out of them directly, right? So, so I was like, "All right, let's let's give it some uh some my health insurance and see if it can find out if I had any any doctors or dentists like overcharge me." And I didn't pay attention and look, I found like almost like $1,700 of like overcharges over 10 years of where like the dentist would something that was supposed to be zero copay they charge me >> and stuff. Are you >> getting the actual money back?

>> So, and then I just told agent, "Go file all the claims. Great. >> I've got 200 bucks back so far. >> Really? >> Yeah. >> Amazing. I think no, but again, insurance is a lazy business like this all like it has a similar I think >> I was at a dinner a few nights ago with a bunch of like hedge fund people and they were all talking about their like short basket of the companies that rely on consumer laziness >> and you know knowing that like muse will start cancelling subscriptions that are hard to cancel and all of that that are just like sticky and like people are lazy about like AOL is still a business now it's private or whatever But like people still have AOL subscriptions from 1994.

>> What? >> Oh yeah. >> What is the product now? >> I mean it's like consulting. >> Your estate is paying for dialup. >> Yeah. No, I mean it's there's so many legacy payments that are like still wired up today. >> Yeah, that makes sense. Well, the thing I was going to say was that I do think we will also see actual agentic bank runs in the sense of like SVB style bank runs where if you remember SVB it was like >> these all are going to just look like the crypto things. I think we've already if you know all the stuff in crypto that's happened, you're kind of going to be like, "Oh, Tradfi is just learning this 5 years later with less cyber security% with less security."

>> But the difference though is >> they're going to be like, "Oh, there's a new yield farming thing out there. >> Everyone the hot ball of money." I guess this is like if you assume that the Asian thing takes off and everyone in America is using it, >> the whole world becomes the hot ball of money. Instead of like a $10 billion hot ball of money, it's a $3 trillion hot ball of money that just moves around. >> But not not just that. I think the other thing about agents that is different is that agents are not very diverse relative to each other compared to humans. So >> you're telling me you're telling me yield farmers were very diverse in their strategies. There's a lot of crowding.

There's a lot of crowding. >> Well, but I mean you look at a lot of these protocols and like there's a bunch of money that still sitting there. Yes. >> That's just like it's like did these guys not get the memo that like this thing got hacked through? >> I I think you're also under people how many people lose their keys? >> Sure. Sure. Sure. LA last night um some people and I were like looking at old protocols to look at this and somebody discovered that compound v1 still has millions of dollars in it. >> Yeah. Yeah. There's always stuff like >> in a protocol that in a protocol that at its peak was like $30 million of TVL >> which is clearly lost keys.

>> Yeah. >> Yeah. Yeah. >> Yeah. No. So typically obviously there's going to be some stranded money and and whatnot, but I do think like the the ways in which humans, you know, SVB was weird because it happened so fast and it was one subculture that coordinated around oh hey, you know, kind of a bad uh whatever earnings call or whatever it was. Uh so like let's all get out of this thing and that cascaded really quickly. It seems like AIS will do that even faster and even a small risk of like oh well you know maybe there's a way crypto is the same thing all the yield farmers are basically already doing agents there's manual rules based agents >> right right yes so like something new happens someone tweets about it people take off like

>> but not in the banking system that's my point consumers don't >> I think I think this is why understanding the crypto market especially when people are down on it is actually very useful because all these AI things just basically make everything look like >> the rationality assumption suddenly got a lot crypto >> homoeconomicus is everyone's everyone >> economist is back okay beautiful all right I think uh that is it on time we got to get going but thank you everybody uh hope you're all safe and sound out there and we'll be back next week see you all