Unchained - DEX in the City: KelpDAO vs. LayerZero: Who Is Liable When a DeFi Protocol Is Hacked?

Episode Date: April 24, 2026

A $300M bridge exploit is forcing the question DeFi has been avoiding: when users lose money, who is actually responsible — the protocol, the infrastructure provider, or both? Thanks to our sponsors! *⁠ As Bitcoin's application layer, Citrea gives you access to the first trust-minimized BTC on a fully programmable platform and a native stablecoin for Bitcoin, ctUSD.  You can now participate in Bitcoin capital markets with lending, privacy, payments, Bitcoin yield, trading and predictions. You get expanded Bitcoin utility without sacrificing its security.  ⁠Citrea mainnet is live. Put your BTC to work at ⁠⁠citrea.xyz/unchained.⁠  *⁠ Nexo is the premier digital wealth platform. Receive interest on your crypto, borrow against it without selling, and trade a range of assets. Now available in the U.S with 30 days of exclusive privileges. Get started at http://nexo.com/unchained A $300 million bridge exploit at Kelp DAO has put DeFi's most uncomfortable question back on the table: when users lose money, who is actually responsible?  Katherine, Jessi, and Vy dig into the Kelp and Layer Zero finger-pointing and ask whether the industry's core values — permissionlessness, open composability — have become its greatest vulnerability.  Then: the Ninth Circuit heard oral arguments on prediction markets last week, and the panel's pointed questions signal the case is headed to the Supreme Court sooner than most expect.  Finally: American Express just solved three of agentic commerce's hardest problems — identity, mandate, and accountability — with a product that's live today. The crypto industry, which should be leading this race, is watching from the sidelines. Hosts: ⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠Katherine Kirkpatrick Bos⁠⁠, General Counsel at StarkWare. Previously held senior legal roles across DeFi and centralized exchanges. ⁠⁠⁠⁠⁠⁠⁠⁠⁠Jessi Brooks⁠⁠⁠⁠⁠⁠⁠⁠⁠, General Counsel at Ribbit Capital ⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠TuongVy Le⁠⁠⁠, General Counsel at Veda Learn more about your ad choices. Visit megaphone.fm/adchoices

Welcome to Decks in the City, where the wallets are cold and the takes are hot. Before we get into the good stuff, here's a word from our sponsors. Bitcoin changed how money works. Citraean changes how Bitcoin scales. With a trust minimized BTC and a native stable coin, CTUSD, Citraia enables Bitcoin capital markets with lending, privacy, Bitcoin yield, and more. Get started at sitrea.x.yZ slash unchained. Hi, all, and we're back.

Before we get going, remember, we're lawyers, but we're not your lawyers, so nothing you hear on decks in the city is legal or financial advice, and it doesn't create an attorney-client relationship. For the fine print, as always, check unchained crypto.com. And first, we have Jesse, Web3 prosecutor turned Web3 protector at Rivet Capital. And V, from the SEC to Web3. And I'm your host, Catherine, KK, fluent in TradFi, and conversant in deep tech over at Starkware. And today we're really excited to be joining you from beautiful Phoenix, Arizona at the Eve Wells Summit. Women are dominant here. There's about 99% women. So we keep joking with the one man, or not one man, a handful of men. It's like three men.

This is what it feels like to be a woman at a crypto conference. And also, like everyone's just complimenting each other on their clothes and being so attentive. Nice. Oh my God. It's also really hot, but I like that. Yeah. Yeah. Frank, Frank Scheparo made that joke on stage. and I actually loved it because he was spot on. Like the women are so nice to each other. It's giving me all the feels. So we have so much to cover today. And we're going to talk a little bit more later on about Paris blockchain.

All three of us were speaking on a, if I say, if I may say, standing room panel, a standing room only panel at Paris blockchain. But before we get to that, we got to get to the good stuff immediately, good stuff, or should I say, bad stuff. It seems like every day there is. is another massive exploit. This is a problem. So we're going to talk about this. B, tell us more about the latest. Well, so I guess the good news is there has not been a hack yet today. But as we all know, there's just been like a run of attacks in Defy lately.

You know, just I think in the last few weeks, there was resolve, which involved manipulation of a minting function and then drift, which was Oracle manipulation, and now Kelp Dow, which involved a bridge exploit. And at the heart of everything, just like questionable security configurations and operational practices all around, right? So needless to say, this is really, really bad. I think I saw somewhere that there's been more lost in these kinds of attacks in the last three months than like all of 2025 combined. And so I think, yeah, we actually be 2025, right? Yeah.

So like we need to talk about it. But so, you know, before I jump into like what happened and sort of the interesting legal questions, I think these things raise, I wanted to say like, you know, we all follow this stuff pretty closely on X and social media. And one thing I actually really hate is like every time something like this happens, happens and people call out things like the poor operational security practices or, you know, governance design or whatever. They're accused of like grave dancing. And I don't really understand like what that word means. And, you know, it actually reminds me so much of whenever there's like

a mass shooting and people are like, you're not allowed to talk about gun control, only thoughts and prayers. It's like, well, where's the accountability? Right. How are we ever going to improve things if we're only allowed to give thoughts and prayers. And like what about the victims, right? They deserve to know what happened and why and what these teams are going to do about it, what the industry is going to do about it to make sure that it doesn't happen again. So I think accountability is really important and we can't have that unless we're allowed to ask questions. And so we shouldn't be silencing each other every time. So now that happens. Can I jump in there, V, because I want you to explain exactly what happened, but also I just thought that was really, really beautiful, and I completely agree.

And sort of the school shooting thing is something that has been on my brain for a while. And I sort of hate to bring that up as a comparison because it's so devastating when it comes to the specifics of that. However, when you think about how DPRK is using this money, it can lead to death and harm at the physical level as well. So I don't want to minimize that either. I think another thing that relates to sort of the analogy of when we talk about gun violence and things like that is there are a lot of hacks. And sometimes it's like, oh, another hack. And then we sort of move on to the next like Twitter conversation because it's hard as humans to sit in a place where there's just constant threat and cybersecurity risk all the time. But what I worry about is like the more hacks there are, the less we're paying attention.

and we're just sort of assuming, oh, they're all the same. Like, if you have good cybersecurity, then you're fine. Because as you detailed out when you were just quickly going through it, like, they're all different and unique. And just having like a smart contract audit or a CISO that understands one part of your platform is not going to be enough. Yeah. And it's not even about smart contract like exploits anymore, right? So it's like you can have a million audits.

Like that wasn't that wasn't the cause of like any of these recent attacks. Right. Now it's about like all the dependencies around oracles and bridges and collateral. And multi-sigs. Yeah. And like more like multi-sig like configurations or operational security practices. So it's like it sort of feels like you're just playing whackamol. Like, you know, North Korea and other illicit actors are just going to keep coming up with new ways,

probably greatly aided by AI to exploit vulnerabilities. And so like we can talk about that a little bit more. but I did want to just jump into like what helped in Kelp Dow just because it was the most recent one. And I think like the largest one this year, right? So last weekend, almost 300 million of restaked Eath was trained after attackers compromised infrastructure tied to Kelp Dow's layer zero bridge. And a bridge is just what allows you to move crypto from one blockchain to another. And because Kelp was running a one-of-one verifier setup, the attacker was easily able to forge a cross-chain. message, mint restaked eat, and then use it as collateral on AVE to borrow or withdraw

real assets with no backing. So that's what turned this from just like a single bridge exploit into kind of like it felt like a system wide like crisis, right? The full impact of which we actually still don't know. So like gosh, there's so much we could discuss here. I think it just raises so many hard questions, a lot of interesting questions. But I just wanted to focus on two things in particular from a legal perspective where I'm really curious to hear what you guys think. So the first thing, you know, like any good lawyer, the first thing I thought was like, like who's liable for this, right? Who's responsible for the losses here? So kelp and layer zero are pointing fingers at each other. And so I think this is like a really uncomfortable one

for the space. But so I think the question is if a protocol's, one of the protocols default configurations effectively leads teams towards something like a one-of-one verifier setup, like the fact that that's an option at all. And apparently 47% of teams chose this option. And that ends up being the single point of failure in an attack like this. Is it layer zero's fault for even offering it or potentially recommending it if that's what they did? Can you really blame a team for going with that option? Right. It's like if a car maker sold a car where airbags were optional or something like that. Right. So I think, you know, on the one hand, maybe Defi is supposed to be about choice and maybe teams should be able to

choose their configurations and maybe they should bear the responsibility of that choice. But on the other hand, I do think defaults matter and like the options that you give actually shapes behavior. right? So like if you have a large portion of the ecosystem, in this case, like all of layer zero's users, if you have something like 47% of them going with this one of one, it starts to look less like an individualized choice and more like actually standard architecture or even like the industry norm, right? So I think what courts will eventually have to wrestle with is when is it like, you know, like not enough to say, oh, we just provided the options or we provided the tools, right? So KK, you used to be a litigator, right?

So I was so curious to see how you would view this. If we were just talking about litigation over like some regular consumer product or something, like who would be liable in this situation? It's a great question because, look, a couple things. Obviously the finger pointing when all these exploits that happen, you know, every time an exploit happens, there's not a lot of accountability. There's a lot of finger pointing. I love that meme where you have the superheroes.

I mean, I retweeted the other day when, you know, because it was like stated in defy right now. And I was like, every time there's an exploit, finger pointing. Now, part of that finger pointing is obviously a, you know, kind of cover your ass mechanism from the liability perspective. Now we're all going to be realistic whenever there's an exploit. Who is working alongside the comm team? It's the lawyers to make sure. And we've all seen a history of litigation where one party effectively gets sued or loses a lawsuit because of contemporaneous communications where they take ownership for an incident. In this situation, it's actually really messy because the first thing you have to ask about litigation is who had the duty and then didn't rise to the level of that duty.

it's often unclear in Defy as to who had a duty if a duty existed. The other thing that's that's really notable is we haven't yet seen a massive rash of lawsuits tied to negligence. Okay. And for those non-lawyers, obviously negligence is a concept that exists in law. And it exists in pretty much every legal agreement that we've ever worked on where people may not be liable for negligence, but they're liable for gross negligence, which is a higher standard, which means, like, they're a hot mess effectively. But I'm wondering if we're going to see more and more of an environment where plaintiffs are suing pretty much everyone involved. Let's be honest. Like, it's Kelp and it's layer zero that are going to get sued if anybody gets sued or if anybody gets sued. You've seen these complaints. It's like 10 defendants, right?

Everybody. Yeah. Plaintiffs attorneys throw everything they can at everyone they can, frankly. It's actually a very strategic tactic. But then the second question that you have to ask is, well, who was negligent? Meaning who had the duty to do X and didn't do it? And I think that's entirely unclear here. Yeah. So, you know, speaking of like litigation and and accountability, I do think one dynamic that we're seeing, we're going to see play out more is, you know, I talk a lot about how Defi is sort of moving past

it's DGen phase and more and more just everyday retail users are starting to touch DFI. Right. Like, and I made the joke on our panel in Paris. Like, if my mom loses $15, she's calling the SEC. So we have, I think, you know, as an industry, we have to be mindful of that, that we're dealing with a potentially very different user base now where we do have to be very careful. We do have to be very responsible if only because I think lawsuits and, accountability are going to become more and more common.

Defy and crypto, like, love to think that we're recreating the wheel, right? But, like, a lot of these lessons have been learned in very recent industries that have been built up. I mean, you just need to go back to fintech. Like, think about this, like, the banking structure, right? Like, a community bank, relying on, like, a SaaS provider, relying on someone who's, like, the custodian of funds, right? And we saw that happen when these neobank.

that aren't quite banks or sort of banks that are actually like banking as a service blowing up. And what it turned out to be was the money either being accounted for by the bass, which is a banking as a service, was not correctly tailored to what the bank thought, right? And it's sort of very, very similar. It's all about ledgers and accountability, which we thought we were building in a better way. And what it has come down to you, because there has been a lot of liability litigation coming out of those, is that if you have the relationship with the customer, then you are liable. Because as a defy user, or if your mom is a defy user, it should not be on her if she is on a certain like front end interacting with that front end

to know everything that is happening underneath, particularly because it might not be written out in the terms of service. There might not be terms of service in defy. you don't know who's changing it, you don't know who's updating it, let alone like your sub-processors might change all the time. So it should not be on the responsibility of the consumer every single time to know the entire stack underlying whatever you're transacting in, whether it be a bank or a fintech or defy. And if defy wants to get serious here, they need to understand that they have a relationship with their customer. And they should be taking that seriously and understand that liability comes with it, but also some form of the current legal structure has to come with that.

Now, it might not be exactly one for one, but like, I have friends in the DOJ still prosecuting crypto-related cases. And when you ask them, like, not all of them, obviously, like there's more nuance, particularly like with the tornado cash case, for example, but when you ask them, like, 99% of their cases are dealing with hacks, what do you do about defy? Like, who's responsible? they're like, most of them say, me, what is DFI? You can always find someone to send a subpoena to. Yeah. Well, the other thing, I just want to really quickly mention. So Austin Campbell, who is also an unchained, well, Bits and Bips host, put something in the lawyer chat that in many ways it was obvious, but it really resonated with me. He noted that, and I should also say, we're all in these

telegram crypto G.C. Crypto lawyer chats. And they're quite fascinating when things like this happen because everyone has a different perspective and you get like a brain trust spectrum of things like liability. Austin's point was,

I think the problem is the whole space is coming face to face with the reality that the rules they wanted to obey like decentralization and true permissionless creates a reality that, you know, effectively maximally empowers bad actors and criminals. I was just so that's so and I'll also mention Austin has said this publicly so I the chat to chat rules no like I so that was the other the other issue I was going to get into is like I think this attack in particular because of the like contagion I think this one in particular was like a devastating example of how the things that make crypto special right so like open permissionless composable interoperable systems can also be its greatest vulnerabilities, right? So that's why you're seeing all this contagion,

and that's why it's still ongoing. Like, we don't yet know how this is going to resolve. So I think it's a really important question that we have to grapple with as an industry. And, like, you know, we have to ask ourselves, like, are we, should we start thinking more carefully about constraints, right? So I think one, one thing that we conflate a lot in crypto is decentralization and permissionlessness, right? Because they're not the same thing. And even just talking about permissionlessness, people usually assume you mean like KYC.

But that's not actually the only kind of permissioning that you can do, right? Like a protocol can restrict what kinds of assets or collateral it will allow or it can impose rate limits, right? So, you know, for example, a rule that says a new user can't just come into AVE and deposit and borrow $300 million in one shot. Like, I don't know.

That kind of permissioning actually sounds like a good idea right now, I think. So I think these are tough questions that we're going to be answering in the near term. And like you said, I think Defi is facing a reckoning right now. And if we really do want to be the financial system of the future, I mean, like, if we even want to survive as an industry and a technology, we need to really seriously think about the trade-offs between crypto's core values, leg permissionlessness, and keeping users safe. So I really defy's at an inflection point. That's a hot take me because, well, and I think I think you're right, meaning that we need to think about these things. It shouldn't be a hot take, but it is a hot take.

And I think the problem we're facing right now is there is a core part of the community that doesn't want to talk about this because there is a palpable fear that if we talk about this, regulators and legislators will seize on this and mandate it across the board. And that's overreact or overregulated. Exactly. But I think the problem here is we're facing two scenarios where regulators and legislators could very well overreact to the absolute chaos of exploits. Or we can have a reasoned conversation that comes along with distinguishing the different

kinds of defy that we're talking about. You know, it's not a one-size-fits-all approach. which like not every DEPI protocol should have circuit breakers or, you know, et cetera, et cetera, all kinds of mechanisms. But in my opinion, as Jesse and I have talked at length before, there are certain things that projects may want to consider that do not reintermediate or all of a sudden make permissionless DFI permission. Y'all know.

We has left the station, honestly, on this. Like, the regulators already have a need on the bones to overreact or maybe just react in some way whether defy goes on Twitter as you're right we should take responsibility he's not going to have an impact here at all I guess my like further hot take on your hot take let's just layer them like a bass system here is I just feel like there are certain people that are in conversations that we're in who would you've been saying that like defy's at an inflection point for years now, like which of these hacks is going to be like enough for us, right? Because my worry is in a week, we're going to forget about it. And like with clarity, who knows where that's going, I can't

believe I even brought up clarity today. Like people really don't want to have this conversation while that is pending, right? And it has been pending for so long now, right? So how do we balance like really being builders in the space while simultaneously working in an ecosystem that doesn't want to point at like their thumbs at themselves and say, I am the problem. I need to fix. You have to do better. Yeah. I say the best way to get regulated is to fail to regulate yourself. Right. So we need to do better across the board. I love a hot take. I mean, we were talking about these at dinner last night. V, you had a great one. What was your hot take last night? I can't remember what it was. Oh, that jeans are really uncomfortable. My hot take is. I didn't agree with that because I love

that. That was a hot take. My hot take is that Taco Bell's canteen a chicken is delicious. So I don't think it's a hot take at all. Not controversial. That was very much. Tacco Bell catering days. And pretty much everyone has to stop working after that lunch. I mean, it's delicious. So good. it's so good. So yet again, if Taco Bell or Diet Coke wants to sponsor this podcast, I'm not going to say no. So I'm just going to put that out there. Like this is not a Baja podcast. So shifting gears to something maybe slightly less depressing, which is good, but intellectually, equally as challenging. So it wouldn't be a month at index in the city if we didn't talk about prediction markets. And why are we talking about predictions and prediction markets now? Well,

because something major happens every five seconds with prediction markets. And what I'm talking about is the latest development is that last week, the Ninth Circuit grilled two prediction markets and a futures commission market in FCM. And it didn't seem to go very well for the prediction markets. And just a refresher to everyone, because I know that this is complicated, but the kind of derivatives environment, the DCMs or the designated contract markets are the exchanges, are the prediction markets themselves. There's also FCMs, futures commissions markets, which are they're kind of like intermediaries. And then DCOs are the clearinghouses. So the DCMs, everybody wants a DCM. DCMs are so hot right now. But they're all litigating all over the country.

States are suing the markets. The markets are suing the states. And the architecture of the legal in legal structure in the United States, and I'm obviously simplifying this, is you have the first level of courts, okay, then you have the second level, the circuit decisions or the ninth circuit decision, and then you go all the way up to the Supreme Court. So the ninth circuit decision is more notable because the higher you go with the courts, the more impactful this is on a precedent basis. And this ninth circuit decision is also notable because if we have the Ninth Circuit rule against the prediction markets and another circuit rule for the prediction. prediction markets, that makes it way more likely that the Supreme Court is going to take on this

issue. So the Ninth Circuit heard this, and I will say the other interesting thing is all three judges on this three judge panel were actually appointed by Trump. But they sounded openly skeptical of the CFTC's federal preemption argument. In fact, one of them was like really grilling the prediction markets and seemed like this was definitely not happening. In my opinion, After reviewing this, I actually think neither Nevada or the CFTC hit this out of the park. I mean, if I were Nevada, I think they really could have leaned into some stronger arguments. However, the judges kind of came into this square one with a highly skeptical position, mainly surrounding as to whether this is gambling. Because this argument was specific to kind of the sports betting questions.

there's also litigation specific to the political event contracts. The last thing I'm going to say before I open it up to V&Jesse is, why don't we have an answer? Some people have asked me like, well, what did the Ninth Circuit do? So I wanted to explain the process here. Like, when you have a kind of a circuit appellate argument like this, it actually takes a really long time to get the decision. And the Ninth Circuit is actually by far the largest federal appeals court in the country. it handles thousands of cases. So it's going to take a few months, like probably 60 to 120 days to actually get the opinion. Sometimes these things can even take up to like six to 18 months,

which is crazy. And so we have to wait for a while to see if the prediction markets are going to win. And while that's happening, we're going to keep tracking the litigation all over the country and see how these oral arguments go in similar cases. But either way, we're going to the Supreme Court, 27, 2028 see you there. Just to add on the legal process explainer for a second, you know, as someone who worked in these kinds of courts and I think maybe both of y'all clerked as well,

when judges asked the kind of pointed, obviously, you know, bias in that they have their decision made up kind of questions, it is 90% likely that they already know how they're going to rule. And that's actually really really, important because you might all be wondering, like, why are we talking about a freaking oral argument? It's just a conversation happening in courts. But the reason why this is really indicative of what

could come out is that when you hear an argument this skewed one way, it usually suggests that the judges pretty much know what they were on a rule. And like the other side, which was Kalshi here, really needs to convince them otherwise. Yeah. And Judge Nelson, the most skeptical one of the three, I do have to mention he had some like real intense, quotable moments. My personal favorite, least favorite in that I support the CFTC's federal preemption argument was this can't be a serious argument. And he was referring to this very specific rule, 4011, an interpretation, which Nevada was arguing that the CFTC has this rule that does not allow gambling contracts. So this has been a really controversial part of the arguments. If you're doing an oral argument, that's about the worst thing that a judge could say to you.

But to the point about oral arguments, I think in general, the assumption is that they don't really make a difference, right? Yeah. But it's interesting to hear. They're so fun to nerd out on, though, you know? KKK, I can listen to you talk about prediction markets like all day. Oh, my God. Let's just do a shout-up for KK, because she always promotes us a lot. KK went on a panel at Eve wealth yesterday to talk about prediction markets.

And her explanation for the CFDC structure was so clear and entertaining. And I'm like, I spent a lot of time learning about the space, obviously not even 10% of what KKS meant. And I'm like, oh, that's really interesting. I learned a lot. So call her up when you need some explainers. You guys are nice. Honestly, at this point, oh, by the way, we're going to, I'm also on the prediction markets panel at consensus. So that one's going to be a lot of fun because they actually.

named the panel, our prediction market's gambling. I was like, well, let's just get right to it. Yeah, yeah. You have to have fun on these panels. I think the three of us shared that, look, you're at conferences, the audience is zoning out. Like, you need to keep it entertaining. Like, you need to make jokes. Sometimes I make jokes and I laugh at myself because no one else laughs. Let's own it, right? Can we talk about for a sigh for a second? Can you share? Okay. We have to think Yes. Yeah. Yeah. Okay. Let's just do a lay of the land here. And then I'll let KK jump in with commentary because I'm still very torn about the whole thing, although it was beautiful. So as Catherine mentioned, we were at Paris Blockchain Week, all of us. And Catherine and I went to a dinner the first night for a number of people, or probably a few hundred people there, where Paris Blockchain Week, the organizers rented out Versailles.

It was the B. Not like the Versailles restaurant, pretty much Versailles. And we were greeted in black tie attire, which is very different than what most of us normally wear and also what crypto people wear. And we showed up and we pretty much had like a private escort with people bowing to us as we walked into Versailles. Like I honestly can't even really even make this up. And then a private tour through Versailles.

as there were people dressed in period outfits, playing period music as we walked through. I couldn't believe that I made it that long in high heels. But then we finally get to the end and they give us champagne and real champagne. That has a label of, you know, the Paris Blockchain Week label created for them. And you're in this huge room right next to the Hall of Mirrors, which I did not know much about Versailles other than the history associated with it. that is like a private room that they then turn into a place for you to watch a private show of fireworks. So it was really lovely to be there. Okay, we wish B had come as well. And it was lovely to experience this. It felt like a freaking out of world experience at some point. I just

could not believe where I was. I just was sort of in shock about the whole thing. It felt a little bit like disconcerting in the way that, you know, so much is happening in DFI that we're talking about and like people are losing their money, but also we're sort of celebrating the success of it as well in this way that was beautiful, but also a little over the top. And Paris Blockchain Week was such a beautiful event. So I don't want to like take away from the organizers of the amount of time and effort they put in because it really, really was beautiful. But I'm like still digesting it. It was a great conference. I have to say, it was incredibly well attended.

It was fantastic to engage with a lot of the European market, like a lot of European regulators, really fascinating conversations. I thought that the organizers, like, look, really well-organized conference. And that goes a long way with me. Like, I've been to one too many conferences where things are running like 45 minutes behind. Which will not be named. Those conferences will not be named.

I mean, the event was absolutely, it was very brief. Bridgeton. I mean, it was very... What was the right word? That's better. It was Bridgeton. It was pretty insane. I was like, guys, my friends in my Chicago suburb are really not going to believe that this happened. So, you know, so look, shout out to the organization, the organizers of this absolutely incredible event. I think, look, the environment that Jesse's describing really sums up with how we're feeling about crypto right now, right? Like, one day I feel like, we're the top of the world. Like, look at all these institutions that are, you know, like climbing over themselves to hire crypto people and look at what we're solving and what

we're achieving in our TPU and our cost savings. And then on the other side, you have the exploits. You have the, you know, offshore prediction market, death markets, although to be clear, not all prediction markets have anything to do with crypto. In fact, most of them have nothing to do with crypto. I need to shout that out specifically. But look, good and bad, all mixed in together. and I think that that we felt that at the Versailles event. There was also a police escort to Versailles, which was crazy. And they also did a really good job with personal security. There was like a heavy police presence.

Obviously, we've seen the wrench attacks in France in particular. So that was also an interesting kind of safety dynamic. Shout out to the organizers for that. So moving on. Oh, by the way, we have to send the Unchained Team at least one picture of the things that we're talking about, like pictures or it didn't happen. So moving on to our AI topic of the week. This is so crazy to me because when we first started doing the pod, I was kind of like,

oh, there's not that much going on with AI that touches on-chain behavior or activity. Now I feel like we have 7,000 different topics to choose from, and they're exponentially growing. And when we talk about hacks and exploits, like the rise of autonomous agents, exploiting smart contracts and doing all sorts of other things. Terrifying. We'll say that for another episode. So, Jesse, turning to you. Tell us about the latest.

I love that you did that buildup. Before I dive into it and do my rant of the day, let's go maybe hear from our sponsors for a second. Bitcoin changed how money works. Satrea changes how Bitcoin scales. Citrae uses Bitcoin as both the settlement and data availability layer. As Bitcoin's application layer, Citraea enables the first trust-minimized BTC on a fully programmable platform, and a native stable coin for Bitcoin, CTUSD.

Satrea offers Bitcoin capital markets with lending, privacy, payments, Bitcoin yield, trading, and predictions. Citraea expands Bitcoin's utility without sacrificing its security. Satrea Mainnet is live. Get started at Satrea. Xyz slash Unchained. And we're back. So, Jesse, tell us all about the AI News of the Week. So the suspense has fully been built.

So I'm just going to dive in here. And it's actually the AI Crypto crossover news of the week and just gives us another opportunity to look at ourselves in the mirror and make sure that we're holding ourselves accountable. So, you know, this week, Amex just launched agentic commerce experiences. And here's what that means. So your AI agent now, if you're using an Amex, can book you a United Flight, charge your Amex, and if the agent books the wrong date, Amex covers it.

So covers the screw up by the agent. And they do this with something called agent purchase protection. So this is new. And we were talking about liability before, right? Like no payment provider has ever said, if a robot screws up on our rails, that's on us. And just like compare that with the like conversation we were having earlier of all. all the defy protocols being Superman, Spider-Man, Superman, I'm obviously not into those, but they're like pointing at each other. So for Amex, like this isn't just a press release.

We've heard so many of these like, injudic commerce and we're all sort of like exhausted with it. But Delta, Expedia, Hilton, Stripe, PayPal, they're all partners here building with them and making this like an actual you can do this now product. So why are we talking about this on the crypto pod, right? Isn't this just another Apple pay with a chapbot, which also sounds fun and interesting, but we need to look closer here. So with this launch, Amex answered the three hardest questions in Agenda Commerce, and I'm sure we'll talk about them again on the pod, but I just want to lay them out here because I think Amex did really interesting things in all three sectors. So first, the issue of identity. You know, we've talked about that on the pod. There's lots of way that's

crypto solving this, but the question really is, is how do you know the agent is who it says who it is? So, Amex built something called agent registration. So every time an AI agent gets its own tokenized credential before it transacts. Second is a mandate, which is slightly harder. So how do you constrain what an agent can do? Book me a flight is not like, book me the most expensive first class flight to the Seychelles, right? Amex built something called intelligence and that has programmable guardrails on it. Does that sound familiar, right?

Does that seem to be something that crypto can do? But then the third is accountability. And that's where I think we should just like pause and pay attention here. So when the agent screws up, who pays? And this is the one that matters most and crypto has obviously not figured out. So Amex says, we will pay if you use our rails and the agent screws up. And that is what is going to unlock adoption here for agented commerce, the liability and the accountability, right?

So especially because the laws in this space are so uncertain. I don't want to like get into a legal deep dive here about it. But just earlier on this pod, you heard a lot about how these questions are just very open-ended. And they really very much are in commerce because a lot of the existing laws right now are built for credit cards and built for a human swiping a credit card, for example. rather than an agent being the customer, essentially. And, you know, Amex isn't alone, visa, plenty of other platforms are doing similar things in different ways. Tadfi is racing to build this world where the agent is the consumer.

It's not a human that you have to transact it. So where is crypto here? Obviously, there's a lot of people building and trying to fix these problems, but they're just not closing the gap quickly enough, in my opinion. And what's so ironic to me here and what I want to discuss with you guys is like blockchain, we should be leading here. This is our time to shine. We're at the family barbecue and finally AI should be paying attention to us, right? This is the conversation we should be winning because blockchain is Agenda Commerce's natural habitat. There's smart contract permissions. There's composable settlements. There's micro payments, right? Without the intermediaries having to make these

decisions. If I were building a payment system for scratch for AI agents, like, I think most people would pick blockchain or something like it nine times out of 10. So the opportunity is still here, but we need to act now because what Amex built is real and it solves the problem of today. But in my mind, it's like just the guardrail. It's not the architecture. And that's what crypto and blockchain can bring to this. Because, you know, with guardrails, everything lives inside. the AMX system. The liability framework is one that AMX can change it anytime, right? They, like, are accountable for until they're not. And so crypto and like, let's turn around and look at the mirror behind me and look at ourselves, right? You know, get your shit together, essentially.

Now is the time. Stop tolerating the hacks that we talked about and your customers getting rug pulled because crypto's real opportunity is not just to like do it like MX is doing it, but build something that's programmable and durable, regardless of like whether the terms of service change with time. So it would be the actual architecture for agentic commerce. And the building blocks exist. And there are protocols for agent wallets, on chain micropayments, programmable spend limits. But to get back to the composability conversation of before, the strength of composability can also be a big weakness in agentic commerce because even if we have one platform that's

doing it and building it, quote unquote, right, or with the, you know, identity, mandate, accountability built in? Is that going to be enough? Especially, like, we talk about the speed of blockchain. Think about the speed of blockchain with agents transacting on it, right? It's like at a different level. And so we need to be thinking about that future now and not shoot ourselves in the book because the opportunity is really right now. Well, and these systems should be cryptography based. Like, they should be on chain. So it's, it's a, it's a match made in heaven. And Jesse, it's such an interesting point because I feel like if you're not already, like, get on AI and get on it now and, you know, like work. Like, this is what I've been doing lately. Like, I've been pulling

late nights, like making sure I am conversant, not conversant, but proficient on all the AI tools. Because I feel like, frankly, it will make me become a better GC. But it's also something that's going to change pretty much every industry. And I think a lot of GCs in particular, lawyers are not doing that. But there's parts of crypto that are doing that. Like, get on board. The other interesting thing, it's, it's, I used to say this particularly a few years ago. It was like 20 years ago, all the Harvard graduates, all they wanted to do was work at Goldman Sachs. Ten years ago, all they wanted to do was work at Google. Now, all they want to do is work in crypto. Like, I used to say that all the time. now cross out crypto and put an AI, which makes sense, but we are at the forefront of emerging

technology. Emerging technology is the best solution to improve and accommodate and make other emerging technology safe. So I would love to see the foremost cryptographers, you know, ensure that they're doing exactly what Jesse said, like considering how we can do this, but make it better, make it faster, make it safer. I just wonder if we're going to capitalize on that opportunity as an industry. B, any thoughts? Are you also pulling late nights on the AI front? But not doing that in particular?

I'm like way behind. No, I actually think this could be a good conversation now and for future pods. Like maybe we could all share like the one thing that we're doing an AI right now, whether it's crypto or related to our jobs as GCs in case there are lawyers listening, of like, that has been helpful for us. And, you know, it can be small little things you can do on AI and whatever platform you're using. So that could be really useful. And, like, Keke, I'd love to hear, like, what's the one coolest thing you've tried or, like, worked on over the past week or so on your late nights?

Well, our paper, Jesse, obviously. Jesse and I. But that's not written by AI. It is not written by AI, to be clear. It is a paper on AI. You know, and I get furious because lawyers for many years have loved the M-Dash guys. Okay, it doesn't mean it was written by AI. I know.

Stop thinking that an M-Dash is written by AI. I don't know what to do without my F-Dash. I've been using the M-Dash since I was 25 years old, guys. Okay, so anyway. So, Jesse and I are finalizing a paper. V is separately finalizing a paper. So just be prepared, listeners, get some shilling about our papers. So many papers.

God, we're such nerds that I've lost. Totally coincidental. I didn't even know you guys were working on this. And I don't think I mentioned mine, but like the timing just kind of... I love AI use. When you get our paper, put it into AI and ask for the big takeaways. That's one of my favorite things is processing large volumes of information, frankly. It's maybe my favorite go-to use case. I also love AI for like light hacks, which, you know, I could go on and on about those. But this brings us to our favorite portion of decks in the city are good. news. So this week we actually have two little nuggets and neither of them are traditional crypto good news. I'm going to be honest. So Jesse, why don't you share your fun little robotics crypto or robotics good news and then I'll wrap with my, uh, all the feels good news.

So like many of us, I get a lot of like newsletters in the morning and I get like obviously crypto ones, which have been a lot of hacks lately. But I get an AI one specifically about like good news and robotics essentially. And there's so many good ones. So, I'm going to throw one in here. So essentially, if you've ever been to China or have seen pictures of it, you know that a lot of buildings there, especially in the main cities, are really, really tall. Like many, many more stories than most buildings in the United States and many other countries. And so what happened was, unfortunately, one of these buildings got, like, destroyed by a fire. And the elderly people who lived in the top area of the building couldn't get up to go retrieve their,

belongings that they had had there for so, so long. So there are these robotics. I don't want to call them legs, but if you look at them, they sort of look like that. But it helped them go up these dozens and dozens of stairs in order to be able to retrieve their belongings that meant everything to them. So for all the like scariness of AI and like, let me freak you out about adjunctic commerce today, there's also such amazing use cases out there. And then there's another crazy one of there's a video which everyone should, you know, Google and I'll send it around of robots chasing wild boars out of like civilian neighborhoods in Europe. Like that was, it's slightly scary like I robot, but also like an amazing use case. And we're not even talking about the robot marathon. The robots won a half

marathon. I don't know. This all freaks me out. I've watched too many. I grew up watching too many like the robots took over the world movies, but there's so many good use cases. So the last thing I want to mention for our good news of this week is I just want to shout out. As I mentioned before, the three of us are at a conference, the Eve Wealth Conference. It's fantastic. It brings a lot of women in crypto together. And there was just, and there are men here too. But there was a really fantastic panel yesterday or fireside. Actually, there were two. One quick, this is not the one, but one was with the basest from Good Charlotte, Paul, who was actually awesome, who was also a blockchain consultant talking about, you know, creative initiatives and artists in in blockchain. So

shout out to Paul. You made all of my teenage dreams come true when I met you and got to do this like because I went to the warp tour a million years ago. But the, the fireside chat I'm talking about is so Chris John Carlo, who is an ex-CFTC commissioner, he's really has a fantastic reputation as being one of the most innovation oriented or sorry, CFTC chairs in CFTC. history, frankly, really innovation-oriented. He's been in private practice. He actually interviewed his daughter, who is an oncology nurse practitioner who has done some really interesting research on women's health. And they talked about a slew of things, including things like intergenerational finance. And I loved it. I loved it because it was certainly non-traditional programming at this

kind of conference. But it also showed how everyone here, we should be talking to our friends, to our family, to our children about innovation in all of its forms. So I talk to my kids about crypto. I want them to grow up understanding crypto and AI. It doesn't mean they have to live on a screen, but I want them to be conversant and I want them to be armed with the tools to navigate this ever-changing ecosystem. And this is also particularly important to educate all aspects across the socioeconomic spectrum. Knowledge is power. I love to see their bond and their relationship.

So more of that content, please. And with that note, I will wrap us for this week. Thanks so much for joining us. We'll see you next week on Dex in the City.