402 Payment Required: a New Way for AI Agents to Pay, with Nemil Dalal, Dev Platform Lead @ Coinbase

Hello and welcome back to the Cognitive Revolution. Today my guest is Namil Dalal, Developer Platform Lead at Coinbase, who recently co-created the X402 protocol, a new open standard designed to enable both humans and AI agents to transact seamlessly on the internet with cryptocurrency. The protocol takes its name from the HTTP 402 payment required status code. Which, believe it or not, has existed since the early days of the internet, but has never been properly implemented due to the limitations of traditional payment systems. As we've covered in a number of recent episodes, AI agents are now quickly becoming more capable and reliable, such that it's no longer crazy to think about giving them modest budgets to spend on access to information, APIs, or MCPs, or perhaps even to hire other agents or human collaborators to help them with When needed, however, because traditional payment systems are designed for human users and because they balance security-related trade-offs very differently, they're often very difficult for AI agents to use and also generally priced in a way that makes micropayments uneconomical. X402 seeks to fix that in a way that's both easy for sellers to implement and for buyers to use. Anyone who wishes to charge for access to an online resource. can, by adding the x402 library to their application, return a 402 response that specifies the details of the payment required for access. The user, which could be an AI agent equipped with a crypto wallet, can then make the necessary payment to the specified address and then re-request the resource with proof of payment attached. Now, it's worth noting that while I have used Coinbase to send international payments with the USDC stablecoin, I'm not very knowledgeable about the crypto space in general. So, wanting to properly understand the fundamentals of this system, I asked Namil, who previously ran USDC at Coinbase, to give me a primer on stablecoins in general. We discussed their benefits to retail users, who we are all trusting to keep the stablecoins stable, and how those actors, including Coinbase, benefit from providing these services. For some, this may be remedial content, but for me, it was very helpful. From there, we get into more forward-looking possibilities. Among other things, I was really interested to learn how crypto wallets, which can be programmed to require multi-signature approval for certain types of transactions, seem really very useful for protecting AI agents from adversarial attacks, to which they do remain very vulnerable, and more generally for keeping humans in the loop when it matters. Beyond that, we also consider what are as yet unanswered questions. Could this be the technology that finally creates a viable alternative to the advertising model? What sorts of reputation systems might we need, and how might they work, especially considering the fact that agents are so easy and cheap to create and delete? Should AI agents have to be capitalized to ensure accountability, as Tyler Cowen has suggested? Is there any way to anticipate the dynamics that might arise when millions of AI agents have wallets, especially considering all the scheming and other bad behaviors that we've seen from the latest generation of models? And what are the prospects for AI finally putting the smart in smart contracts and greasing the wheels of even local commerce by using AI models to provide low-cost resolution for offline disputes? All of these are fascinating possibilities that we should all be watching closely as the agent economy continues to accelerate. And big picture, this discussion is also a powerful reminder of the contingency of the technologies that have enabled the AI wave. Most famously, the development of GPUs, aka graphics processing units, was driven for years by the unique demands of video game rendering engines. Here again, cryptocurrency, which was originally motivated by a desire to create digital money that not even nation-states could control, might prove to be the perfect enabling technology that allows AIs to participate in the economy. I see this both as a reason to be humble about our ability to predict where all this is going, but also as a challenge to technologists and technology investors to invest now in the development of potentially complementary technologies that will enable the monitoring, supervision, and control of the AI agent economy. I'll definitely be returning to that theme in future episodes, but for now, I hope you enjoy this primer on cryptocurrency and exploration of how AI agents might get their first access to economic participation. With Namil Dalal from Coinbase.

Namil Dalal, developer platform lead at Coinbase. Welcome to the Cognitive Revolution.

It's fantastic to be here. Thanks for having me.

Yeah, I'm excited for this conversation. I honestly have a lot to learn. Regular listeners know that I'm totally obsessed with the AI technology wave and study it from all angles. But the same is not true of the crypto technology wave. So I'll probably ask some kind of remedial or sort of basic questions to get started. And with that, introduction and helpful orientation from you, then I know we've got some envisioning to do of what the intersection of the AI and crypto tech waves might look like, and specifically a protocol that you have recently put out to try to make this more concrete and really start to enable it. So maybe just for starters, because this is going to be kind of a foundational technology for the purposes of the vision that we'll be sketching out, Can you tell me where we are on stablecoins today? I think everybody knows that Bitcoin goes up and down and it can be crazy. But then there's this whole domain of stablecoins, which I've used a bit actually through Coinbase to pay international contractors from time to time. And it does have sort of a magical feel to it where I can just drop some dollars into the Coinbase account, convert them to USDC, send them seemingly anywhere in the world. On the other side of that, I'm not entirely sure exactly what happens, And I'm not entirely sure how it works, but maybe you can kind of give me a little stablecoin 101, which I think not only I, but probably a number of listeners would really benefit from.

I would love to. And I actually used to run USDC at Coinbase, um, when it was a first venture bet, just a few tens of millions in market cap and helped grow it to a billion. So really excited to talk about it. So maybe step one is like in the last decade, we basically went from zero stablecoin usage in the world to about $250 billion today. There are a number of different players in the market. Coinbase is one of the really big ones with Circle, which is USDC. And if you think about the blockchain, there's a lot of benefits of the blockchain. One is it's globally available. Anyone can use it. 2 is fees are often a lot lower than traditional financial system rails. 3 is it's often instantaneous. It's a matter of seconds to be able to send this anywhere in the world. So these are all the benefits of the blockchains. We've known about this for a very long time, you know, since Bitcoin, a lot of these benefits started coming together. But But what was always hard was that there wasn't stability. And the power of something like the US dollar, the euro, any of these other fiat currencies is that it has that stability. People buy things in it. They go to the supermarket, they go to a drugstore, they can buy things in it. So the power of a stablecoin is it combines all the benefits of a blockchain, but it has the stability of the US dollar. And so you asked some like more detailed questions about this, and I'll share maybe a few areas I think that are interesting for folks to know about. One is that like, how does it get its stability? And typically the way this works is what the issuer does, in this case, Circle and Coinbase, what they do is they take dollars that they're getting, they put it into a bank account, they put it into treasuries, basically some type of short-dated government paper is what they put it into, or a bank account. And then what they do is they issue a token 1-to-1 against it. So the way it gets its stability is that you have those tokens in a bank account, you're issuing it 1-to-1. So if I have $10, somebody gives me $10, I can create 10 USDC tokens and I can give it to that person. user, and then they can go anything they want with it. And if they ever want to come back and get the underlying fiat back again, I go ahead and change that back. And so that's the power of this thing. Now, like, think about, like, this is the vision 10 years ago. The thing that's changing these days, or maybe I'll just spotlight one or two quick things that are changing. One is just blockchains are getting faster and cheaper. So Base nearly had 1,000 transactions per second. Settlement times are a matter of like less, like 500 milliseconds today, um, is the plan that what Base is putting together. And then fees are basically cents or less than that. So they could be sub-cent at different moments in time. And so all these things suddenly mean that, you know, blockchain before was good for all these things. Now it's essentially exceptional. And so the vision that we really think about is like, can you make sending money as simple as sending a text message? And that's the vision with stablecoins and where we're headed.

So that's a great start. A couple questions for somebody, again, who doesn't understand all this deeply in a technological sense. Obviously, one of the big ideas of Bitcoin originally was it's sort of trust. You don't need to trust anyone in particular, right? You're just trusting this sort of decentralised network, and somebody would have to take over the majority of the network in order to launch an attack. And over all these years, it's been robust to that. sort of thing. When we have a stablecoin, am I trusting someone?

Who am I trusting?

And is there like a concept of a fractional reserve that's behind that, or is it literally one-to-one? And how do I know as an end user when I have a USDC that there really is the ability to redeem it? Because I think a lot of times, as I said, I have done these transactions and all of the surface-level user features I have experienced, but I'm not quite sure exactly what is underlying that and providing that sense of stability.

Yeah. So for stablecoins, unlike Bitcoin, you absolutely have to trust someone, and that is one big difference. And the person you're trusting is an issuer. In this case, it'd be Circle and Coinbase are the issuers for it. And you have to trust that they're taking the money that they they put the funds in a place where they claim to put it in, right? And this is some of the risks that come with stablecoins. And the way Circle and Coinbase get around that is what they do is they share a transparency report. Uh, so not even get around, but they actually like share transparency and show a radical degree of candor in terms of where those funds are going and where they're held. So with that degree of transparency and trust, and by the way, Congress and others are in the process of exploring ways to legislate stablecoins to make them even more trusted all around the world. But absolutely, you need a form of trust in it. And it's typically on the issuer that you have to be able to make sure that that is the case. The other main thing is that you have to make sure that the keys for the mint and burn of the smart contracts themselves are safely held and someone can't just infinitely create USDC coin without having the underlying reserves. So typically, it's fiat-level transparency, and then it's smart contract security that you need to be able to do to be able to support this network.

And then how do you guys— like, when I go onto Coinbase and bring my dollars and convert them, and then I make a payment, and that payment may go to anywhere in the world, right? And then on the other end of this, my counterparty, who also— actually, I don't even necessarily know if they're working with a Coinbase account or maybe some other wallet or whatever, provider or enabling technology. But if they're in India, for example, then they're going to cash out into local currency, right? So they're going to get rupees at the end of the day. How are you doing this with no fees? Obviously, there are significant fees if I try to do that through the traditional banking system. And even if I imagine sort of saying, well, okay, we're just going to do this as sort of a public good because we want to enable the ecosystem or whatever, there would still be currency risk, right? And it just feels like there could be some risk of buildup or currency movement. Such that like in being the market maker doesn't sound like a great place to be if you're making no fees on it. So I'm just a little confused as to like, who's taking on that risk that seems like it can't be totally eliminated. And why are they taking on that risk?

Totally. So with any product like this, you need market makers and you need cash in and cash out points. And both of those take money. There's compliance infrastructure that's involved with it. There's physical branch locations. Um, there's money movement as you're going back and forth between different currencies to keep the books healthy in a financial exchange. So absolutely all of these parties need to be able to make money as part of this. The secret of stablecoins is that the money is often made on the yield itself. So when I take $10 and I convert that for the stablecoin, that money is going into some form of short-gated treasury or a bank account, and there is yield earned on the other side of that. So for example, today the stablecoin market cap is about $250 billion, almost exclusively US dollar-based stablecoins. There is a 3 to 4, 4.5% yield that is being made on those funds. And that is a lot of different economics that can go around to be able to support an entire ecosystem that lets people do it. Now, candidly, this exists in the traditional financial world and the banks take all that and still charge the fees, right? And the idea of this methodology is you can make the fees from that, but then you can offer a very cheap or free service that anyone can do. And so let me just tell you the story of like how a user might approach this. As they come to Coinbase, they want to be able to convert $10 in their bank account into stablecoins. So what they do is they hook up their bank account to Coinbase. Coinbase will pay some ACH fees for that. So it might be a few cents in the US for ACH, takes 2 to 3 days to move over. Those $10 show up at Coinbase, and then suddenly you get 10 stablecoins, 10 USDC tokens in your wallet. Now you can send them anywhere in the world. They don't have to be a Coinbase user. They, all they need to do is have a crypto wallet on the other side. So let's say it's someone in India, like you said, that person in India will suddenly have 10 USDC. Now there's a few different ways this could go. One is they decide they wanna keep the USDC. They wanna put it into a DeFi protocol. They wanna earn some yield on it. They wanna spend it for some goods. And slowly we're starting to see this happen more and more, but it's still early days. Not every merchant that you go to on, on the local street in India, and by the way, I'm Indian, so this is a cause near and dear to my heart. You can't just go spend that at local stores. But what you can do is you can sell it and there's a market maker on the other side. And that market maker is collecting the USDC, deciding whether to keep in USDC or actually go back to Coinbase or Circle and convert it back to dollar. Right. And then on my side, what I'm doing is converting that, probably paying some type of fee to do that. In this case, in India, like probably I would have to pay some type of fee. It might be half a percent or a percent, or it might be a few cents depending on the dollar amount, on the rupee amount. And then suddenly I have my rupees. I can go spend it wherever I want. So that's the system as it is today. But I think rapidly what we're moving to is a world that's closed loop where I have a stablecoin. A stablecoin on-chain is better than a fiat stablecoin or a fiat rupee. Like a rupee on-chain is better in that it can be sent around the world, it can be cheap, and it can be used with this entire decentralized finance infrastructure. And I don't know if your audience is aware of that, but basically there are things like Yield protocols, ability to trade, just basically a lot of the traditional financial services people are used to, and there are ways to store it, et cetera. And so these are all the things that you can do once you have that on-chain rupee or that on-chain dollar.

Yeah, I mean, it's just, it's a, it's a lot for me to wrap my head around as someone who hasn't done it much, but I hear you on the fees on the cash outside. Makes sense. Would it be fair to say that we are in sort of an early, almost like the sort of subsidized era of this from a Coinbase perspective? I mean, famously, Uber and Lyft were burning tons of cash in order to scale their operation. And it's a joke at this point that the sort of millennial lifestyle was subsidized by venture capital for a long time. Are we in a similar period with crypto where Coinbase is just willing to take you know, a customer acquisition initial loss to get people onto the platform and the sort of the chain lifestyle in the first place?

No, like this is already a big business. Like that's the part that's crazy is that when you earn 4% yield, so there's $250 billion of stablecoins. That means the people who issue stablecoins are making $10 billion plus a year already. The Uber example is an example where they were making very little and they were using VC to subsidize that. But the reality is these are already great businesses. Like these are fantastic businesses. And, but they're like, if you have less margin than a bank, you can do a lot of amazing things with that. And so what I would say is like, they're making good money. And as a result, there's all these different things that you can do with these stablecoins. Does that make sense? Yeah.

Although I also get yield on my USDC. I haven't held USDC much. But I do notice when I deposit dollars into Coinbase and then I do that conversion, all of a sudden I am getting yield.

Right.

So I guess there's obviously just a spread between what Coinbase is getting and what I'm getting.

Yeah, that's exactly right. So like Coinbase is making some money. There's also value to Coinbase for USDC continuing to grow. And that's one of the reasons when you have it on our platform, as an example, we offer you yield directly on top of it. But again, if you took that off platform and just had it in your wallet for a year, You would be not earning yield on that, and Coinbase would be earning yield, and Circle would be earning yield together. And that means essentially that, you know, like, you can either move it to Coinbase or move it somewhere else where you earn that yield, or you can hold it off and Coinbase is still earning the yield, very similar to, you know, how a bank might earn that yield.

Yeah, interesting.

Okay, and just kind of—

And then, yeah, one other thing I'd say is that blockchain That's another really interesting thing is that like, why does it cost so much to send a wire transfer in a world where an SMS around the world is free, right? A lot of what we are sending is really just information. It's not necessarily the money that is always moving because there are fiat, like there are on and off ramps and various other ways to send things. So what I would just say is that what the blockchain has been really amazing at is suddenly make this a much more commodity and open market for anyone to move. So the first version of blockchains, I remember in 2017, is about like $2 to $5, right? And then gas fees spike got to $100, $200 to be able to basically make an on-chain transaction. But today with Base, with others, it's like a matter of a cent, 2 cents to be able to send this. And that's not with subsidies or anything. That is the power of the blockchain, right? The blockchain allows you to have this like open market. Anyone can be able to send transactions and the technology is a lot, lot better than what it was for traditional banking rails. And I think like, I've worked back earlier in my career for these banks. A lot of this is like 1950s, 1960s, 1970s infrastructure that they are running on. And it's really, really hard to convince a ton of banks to change their new technology stack. The power of the blockchain, just like the internet was, you have this old system, you have a new system that's built with a whole new world in mind. And as a result, everything is so much cheaper and faster as a result of that. So I mean, that's the other thing that people really surprised by is the fees are so low. And that is not a function of like subsidies. It is a function of the technology architecture and the fact that just in general, networks have gotten so much better over the last 30, 40 years, but it just hasn't flowed through to finance yet.

Hey, we'll continue our interview in a moment after a word from our sponsors. Yeah, so maybe 2 follow-ups on that. One is on somewhat of a different services profile, and the other just in terms of the underlying technology that's enabling low transaction costs, which is maybe a good transition into the new 402 payment required X402 protocol. On the services side, there is some difference. Maybe you can kind of compare and contrast what I get, because part of what I do get if I send through the traditional system is I get some recourse, right, that I can sort of say, the money didn't get there, or somebody defrauded me and it wasn't actually me that made this payment, right? There's various, obviously, fraud vectors that I can sort of engage the traditional banking system on. And my understanding is some nontrivial fraction of the fees associated with those transactions are basically to cover the fraud that actually gets through, and then also to pay for the services of remediating fraud that people are trying all the time. Whereas my general sense is that there's just much less of that on the crypto side. And so my trade is like, I am paying a lot less in fees, but I also have just qualitatively less in terms of redress available to me if something goes wrong. Complicate or clarify that picture for me.

Yeah, absolutely. So what I would say is that there are exchanges around the world. They have forms of KYC, as an example. They have other forms of checks and balances that they have on their system. So absolutely, there are some basic form of that, but I think the core of your message is correct, which is that unlike more traditional systems, this is not a reversible system by default. So blockchains, when you send a transaction, they end up on the other side and that's it. There's no recourse to be able to necessarily pull them back. by default. But there's some interesting things that have been baked into this. So for example, for stablecoins in general, a lot of them have to work with national governments and partners. And so for example, if there's a legal case associated with USDC that enforcement agencies reaching out, that is absolutely something that is baked into the protocol where there is the ability to be able to hold those funds. Like Coinbase, like any other exchange, Circle and others, they all need to follow, for example, sanctions law that the US has and other international jurisdictions. These are absolutely things that can be done, number one. So there is some form of recourse, but you're absolutely right. It isn't the like press a button and suddenly you can reverse the transaction. But that's also the reason why it's so fast, right? Because like if you can press a button and reverse a transaction, the other side can't really trust that they've gotten the money yet, right? And this is why for ACH, you might have to wait several days to be able to like access your funds, right? While with blockchain, it might be like 2 seconds and suddenly you have access to your funds. And then the last thing I'd just say is that blockchains can absolutely add reversibility in more and more complex ways into the stack if they want to do it. Just like one layer on top, which we say like, here's the foundational layer. It's not reversible, but actually this layer on top, you can always, you can wait for 2 days and the sender can pull the money back. These are all things that people can build. And in fact, people have built these. They're not like super popular yet, but I expect over time they are going to become more popular. And then the last thing is like, So this broader point of like, that's one of the reasons for the fees. Yes, that is part of the reason for the fees, but I don't think it justifies all the fees. And I think that's probably like a big shift in terms of like what this is, is that you're right. That part, like if you added a better system there, probably their fees would be a little bit higher than they are today. But compared to, let's say, a wire, which people pay $20, $30 for, I can send a trillion dollars or 10 cents for less than a cent. today on the blockchain. And that is something that's a really new feature, I would say, of blockchains compared to the traditional financial system.

Yeah.

So maybe my last question before we get to the X402 is, can you maybe just sketch out a little bit like how those transaction costs on blockchain have fallen so dramatically? This is, you know, something that in some sense, even a totally crypto ignorant, AI obsessed crowd will have some sympathy for because we're used to paying for compute. So my general sense, very high level, is like, it used to be that the transactions were in some sense backed by this proof of work concept, which is basically proof of compute. And a lot of the compute requirements have basically been taken out and replaced by other schemes. And obviously we're seeing a Dramatic. The compute is not taken out, but certainly many efficiencies have been found on the AI side as well. So what's kind of a very 101 story of how the transaction costs have gone from having spiked to $100 or whatever to now being so low?

So maybe just the overall message for your listeners is that just like there's a Moore's Law, there's a form of that just for bandwidth. Bandwidth gets better and better over time. There's a form of that for blockchains, and that We've had the Bitcoin was launched in 2009. We had 16 years now to be able to refine and innovate on ways to be able to make it cheaper to be able to do that. So the biggest shift that's happened is there's more and more block space available. And I would say like, there are just different examples that I can give you is like in the Ethereum ecosystem. So what Ethereum did was back in 2017, if you wanted to launch a transaction, I put it on the blockchain and I put it on the Ethereum layer 1 blockchain. But today what's happening is that you have these layer 2 blockchains that are on top and you even have layer 3s and they're taking all this information. They're letting people make those transfers and then they summarize it and they put it onto the layer 1 blockchain. So they use a lot less amount of storage and storage has also gone up on the layer 1 side as well. And so as a result, the amount of block space has just gone up a ton. So I think this is not that dissimilar actually from the traditional banking system where Banks don't settle every single transaction individually. They might settle the net over time of the different transactions. And I would just say like, this is not exactly the same, but I liken it to a model where the layer 2s are moving a ton of different transactions on them. And over every periodically, what they're writing, something called, we call it Ethereum, the beacon chain. They're writing to that as one way in which to do it. There are lots of other ways to scale, and I won't go through all of them, but another example could be is that Bitcoin was very, very, focused on decentralization. If you centralize that a little bit more, let's say Bitcoin, the vision was that everyone could run a node. But if you centralize that a little bit more and say, hey, thousands of people or tens of thousands of people around the world can run nodes, but not every human is able to run a node on a phone, as an example. If that's the case, suddenly the amount of transaction capability for any node goes up substantially. If I can run it on AWS or any of the other cloud providers, that's another way in which this goes up. So if I look forward, I think that like the TPS on Base just hit almost 1,000 TPS just last week. My expectation is that this is going to keep growing substantially. And in fact, like everyone around the world is going to be doing transactions on the blockchain. And it's not just financial transactions, but social networks are going to live on this, all these other things, just like the internet. And so that's the vision that we see like going forward. And that's exactly how we've gotten so far. So in 15 years, we've done this much. Now it's going to be really exciting to see what it can do in the next 15 years for this.

So it's a mix of both efficiency advances at just kind of the raw algorithmic level, and then also some design trade-offs, basically, where some combination of those is delivering much higher scalability with some compromises in terms of the original vision of Bitcoin specifically.

Yeah. And so Bitcoin was one network. And by the way, I'm saying this holistically for a lot of different networks. So different networks have chosen different paths along that journey of like, do I want to make a design trade-off? Do I just want to work on, you know, this other factor of it? Um, but absolutely there's a mixture of different approaches that the entire blockchain industry is taking. Cool.

Well, let's get to X402. This is the new protocol that is meant to First of all, I love the fact that there has been this HTTP code. I didn't know this, but everybody knows 404, everybody knows probably 401. Not too many people, including myself, probably have heard of 402. So introduce the X402 protocol, like what it's meant to enable and a little bit about how it works.

Yeah. So let me just start actually at the start, which is that we launched something called AgentKit. And for a lot of your listeners, you have an AI agent, but you want it to be able to do things on chain. And so what it needs is basically a wallet. So we built that last year. We built that. We had thousands of developers start using that and embracing it. We had teams like Virtuals that launched on AI personalities that you could tip, that could send money around the world, et cetera. And they're all using AgentKit. And when folks were using that, they were like, look, wouldn't it be amazing If with this wallet, this AI agent could actually go pay someone or something. So for example, I could go buy a good, uh, or I could go to an API. I could access the data that the API is providing. I could, um, use, send a text message, for example. These are all things that AI agents can't do because they don't have access to money. Right. And in a world where AI is essentially becoming a proxy for humans, that sounds crazy. And so like, it was very clear to me, to our entire team, that AI agents needed to be able to access money and to be able to go do things with it. And that's where X.402 comes in. So X.402, in the early days of the internet, the vision was to add credit cards to it. And at the time, it was really, really difficult to do. You needed Visa, MasterCard, the others to play ball back in the '80s and the '90s. And as a result, this never happened. And the internet had a standard where if I wanted to access a webpage, I didn't need to like look at an ad to do that. Right. That's the model that we're all used to today. But what it could do is say, hey, uh, Namil, you're looking at this website before you, we can send you the page. You actually have to pay 10 cents to be able to do it. And I would pay the 10 cents some way, right? I get this 402 status code and say, oh, I didn't realize that I had to pay. Okay. 10 cents. Awesome. Send the money along and suddenly I have access to the webpage. So the internet had this conception that they wanted to do this, but they could not do it. And we decided to say, look at that. And today we said, look, the use case is actually a little different. Like we kind of gotten around the fact that 402 didn't exist by just using credit cards, but credit cards are not ideal for AI agents. You need things like permissioning. You need to sign off if they want to do it. And AI agents will want to access like lots of different things. You know, it's not just like a subscription where I sign up for it and I can pay my money and then use it all the time. AI agent may go here for one query, it may go here for another query, and it needs to pull all that together. So we said, look, this is the perfect moment to be able to bring this back. And the blockchain is the perfect mechanism because it allows you to move stablecoins around the world, lets you move other Bitcoin, it can let you move NFTs, all these other things. And so that's essentially where we put it together was X402 is a standard. It makes it super easy for any website to be able to accept money and then to be able to return a result based on it. And so simple, 2 simple examples. I have a storefront and AI wants to buy from that. I can add xFero to, and suddenly they send me some crypto tokens, USDC, let's say 20 USDC, and I send a good to an address that the AI agent has told me. That's one example. Another example is I have an API and I want to be able to provide that API to AI agents. So one example you can think about is there's some financial data and I want to let that AI agent trade the stock market or trade on the blockchain. That's an example where I can sell that data for a fee. Another example could be like Twilio, where I want to send a text message or use, you know, some other form of API. So I'll, I'll pause there, but really, like, ultimately xFlow2 is an open standard. Anyone can build on top of it. We just, like, incepted it into the world, but it makes it super easy for anyone who has an API. to be able to accept money in the form of crypto tokens and then offer a response back as a result of that. And we think this is really valuable for humans and micropayments and all that stuff. But I think in terms of where this moment is today, I think it's especially powerful in a world of MCPs and AI agents that want to go on, like, online and be able to do things. Because, by the way, a lot of times AI agents are not watching ads. So one of the big changes in the web that's going to happen is the entire monetization model that It's like my point of view is like the entire monetization model is going to change going forward.

Hey, we'll continue our interview in a moment after a word from our sponsors. Yeah, I want to get deeper into how it works. Maybe just taking one step back, like it sounds pretty nice from a user standpoint as well to, you know, from a human user standpoint to envision an internet where it's like Okay. And especially if you're talking like display sidebar ads, whatever, those rates are not super great, right? And we all kind of know that it's not super easy to— you just plug in Google Display Ads or whatever, and it's not like the cash is just immediately rolling in. And I think a lot of people, it seems to me, it feels intuitively appealing that I would like to have an option for publishers to say, you can buy our content on a micropayments basis. Obviously everybody's got subscriptions all over the place, but that's getting tiresome and people are obviously overloaded with subscriptions and it's kind of run its course. It feels like in some ways now I go to CNN and all of a sudden they want me to pay $30 a year and I'm like, I don't think I need another subscription to CNN, but I might pay the 1 to 10 cents for an article of interest. Why hasn't this already happened? Because we've had sort of the blockchain capability for at least a while now, right? What is it that has blocked that from materialising for human users, leaving aside that we now have a new class of AI agent users?

Yeah. So first I'll just say, for micropayments, traditional methods are really challenging, if not impossible. Right. It's very hard to be able to charge $0.01 with a credit card. Yeah, you can't.

But I mean, with the crypto technology that we have, yeah, I still have never had the experience where I've gone to a website and they've said, You can pay, you know, 1 Satoshi or whatever for, uh, for access to this article.

Yeah. So, um, by the way, there are some sites that are starting to offer this. Zora, um, they're more crypto native today, but Zora, Farcaster is another one. These are examples where you can do things for just a few cents on these apps, but you're absolutely right. And I would say like the biggest shifts have been, one is like early crypto didn't have stablecoins. That was like one issue. 2 was that fees were super, super high. So like I'd have to pay on Bitcoin, you know, $5, $10, $20, $30 to be able to make transfers. And only in the last like 6 months to 12 months has that really, that price really nosedived because of things like layer 2s, Solana and others have started really cutting down on the fees that are charged for it. So previously, I think that like with $20 fees or $10 fees or even $2 fees, it really impacted the types of micropayments you could do. And the weird thing was that I think in 2015, 2016, it was possible to do really like small transactions, but then the blockchain became super popular. And so fees started spiking. And that's essentially the battle that we're fighting right now is to make it easier and easier and lower and lower fees. So my instinct is that like now it's going to become increasingly possible, but it required us to be able to come up with these layer 2 technologies with these cheaper blockchains that just wasn't possible even 2 years ago or 3 years ago.

Yeah, that's interesting. I mean, in a way that may very much parallel the AI story, where often people are like, if this AI is so good, why isn't it in use everywhere? And a lot of times my answer is like, well, it wasn't useful at all until 2 years ago, and even then it wasn't multimodal yet, and the context window was 8,000 tokens and whatever. And so depending on exactly where you want to put the thresholds for real usability have only, you know, fairly recently been crossed. And so it sounds like there's kind of a similar story there.

It's the thing I really enjoy about seeing the AI and crypto journeys, like, together has been they're both, like, early technologies that are, like, now starting to get mass market. And we're finally seeing, like, the mass market capabilities. I think crypto has been around a little longer in its, you know, in its form. But I would say that, like, only in the last few years have we really figured out ways to make this really simple and easy. That, by the way, that's one other big challenge in crypto is usability. Like, the usability of it, like, with like addresses and things like that, was historically really, really difficult. Only now are we figuring that out. And I see the same thing in AI, is that like 6 months from now, it looks unrecognizable in many ways from what it was 6 months ago because there's just so much innovation going on. So absolutely, it's like, I think this is the thing where like the moment is finally starting to be now, at least on the crypto side.

Yeah, for AI, I mean, the, the whole concept of Prompt engineering. I'm old enough to remember when prompt engineering was a thing, and there was a time when it was really sort of a high art. And we've had some of the early pioneers of the most creative prompt engineering strategies as guests on the show. And now, especially just the last 10 days with Claude 4, I'm like, this thing genuinely feels like talking to a human in many ways. Like, when I'm asking it to develop software for me, I don't have to lead it by the hand. I don't have to think super hard about exactly what I want back from it. I kind of give it notes like I would give to a human developer, and it is increasingly capable of taking those notes and running with them. And it really is an amazing thing. So let's go deeper into how this actually works, because we do have a lot of people, I think, especially who are going to be building agents, and they might be building agents totally bespoke with a framework or coding their own or having Claude do the coding. They might be building something that's a little bit more of a structured workflow slash agent that might live in a no-code environment. Who knows what they might be doing, right? There's all kinds of form factors coming online. So maybe just take me through the cycle of the protocol. What happens in one sort of successful loop? What does the payment receiving side have to do? What does the agent builder have to do to equip their agent to do this? What is required to actually make use of all this?

Yeah, so let's go through maybe an example from end to end of how this works. I think one that's near and dear to my heart is inference. If you're in AI, you need inference to be able to work. One of the fun use cases that we talk about is an AI agent can have a wallet and it can pay for its own inference as a way in which to be able to do this. Let's start at the start. You have the buyer, which is let's say an AI agent that wants more inference, and it goes to an API and it wants to access inference. So what it does is say, hey, I would like to access inference, send it my way, right? That's essentially like the request that it composes together and it sends it and it gets a 402. And the 402 has a few sets of details. Hey, the inference costs this much, put it together on this network, for example, the Base blockchain or Solana blockchain or any of the other ones, and send me like a transaction as part of that. And so what the purchaser does, in this case, the AI agent says, okay, awesome. I know how much I need to pay for it. What I do is I take my crypto wallet, which they can use, and you know, you're, you can use AgentKit to be able to hook up an AI with a wallet. And so I have that wallet and I say, okay, awesome. It said I need 5 USDC on Base to be able to do this. And I'm, so what I do is I sign a transaction. I don't need to broadcast it. All I need to do is sign that transaction and send that payload back. So I respond back to it and say, awesome, same request. And by the way, here's the 5 or 10 USDC, whatever that amount is, in a signed transaction. Now on the inference, on the inference provider side, this is where it gets really interesting is it says, okay, awesome. It sent me this payload. It looks like it's a well-formed transaction. Number one. Two is it looks like it's good for the money. It actually has the money, which is awesome to see. And hey, I'm going to actually wait to submit it because I actually don't know if their request is going to use all $10 that they've sent me of the inference. Maybe I'll use more, maybe I'll use less. Let me go do the work. So it does the work and then it has a decision to make. It says, awesome, this looks great. I submit the transaction and suddenly now I have the $10 myself, like they've arrived on the blockchain to me. And then I can send back the funds that I have. But there's some, like, as you can imagine, there's a ton of different edge cases and happy to talk through any of these if they're interesting. One is that if I overpay, I can always overpay, get some type of token that allows me to keep that as basically like a prepaid card that lets me keep spending that money until it's exhausted. Or let's say I've underpaid. And that is actually, I need more money to be able to do that. I can respond without the response that the buyer is requesting. I can say, actually, you need to send me more money. I thought it was $10, but actually it's $15. Go ahead and like make a new transaction and send me that. And again, here there's like a bunch of different edge cases where maybe I spend that money on inference and actually it should have paid me more. And so I'm out of that money and things like that. And so again, there are a lot of like smart things that we can build on top, but I'll pause there. Like, that's the core of it. It's super simple, which is that, hey, request x402. Awesome. I'll put together a blockchain transaction, sign it, but don't need to submit it or anything. The other side receives it, validates it, verifies it, does the work, submits the blockchain transaction, and then returns me whatever the inference capabilities that I need. And so that's it. It's like basically like request and then a response and then you're done.

And part of what you guys have released is like middleware software for the payment receiving side, right? That allows them to I saw sort of a one-line drop-in of like, yeah, this one line of code will sort of check for the payload, validate, yada, yada, yada. And if it fails, kind of kick back this 402. Is it really that simple? Or like, how simple is it really to start accepting money this way?

Totally. So first of all, x402 is just a standard. Um, it's like any other form of internet standard. It's out there. Anyone can use it. They can modify it, adjust it. We're super excited to see what the world does with it. And we're also seeing like tons of companies reaching out to us. That's number one. But what we realized was like, it's not as simple as just throwing out a standard. You have to make it super easy for people to do. And you actually have to come up with use cases, like the inference example that we were talking about to make this super easy as well. So what we did was we built some middleware. We started with like Node.js, which is one of the most popular JavaScript. It's one of the most popular languages in the world. We said, okay, let's make a super simple middleware where I can put it into my web server. and my— in front of any API call I already have, I can be able to put that in front of it. And it's one line, basically one line of JavaScript that I include, and suddenly that's it. It works. But that wasn't like— I don't think that's even simple enough because you still need to manage the blockchain infrastructure and things like that. And that's where Coinbase Developer Platform comes in. And that's the team that I lead. But we basically made it so that like we can spin up a wallet for you. We can automatically convert the funds you receive into dollars in your bank account if that's where you wanted them to end up with, rather than having them on the blockchain. And in general, with standards, I think like the key takeaway for us is like, we got to make it super easy. We got to support people, you know, to be able to do that. And that's essentially what we do. We have a Telegram group where anyone, anyone in the AI community who wants to come to that, we'd be super excited to get them spun up. But we have a Telegram group, they can like check it out. And for anyone who's interested in more details, they're all available at x402.org. So you can see both the white paper, you can see the diagram that I went through of like the request response. You can also see the details of just exactly like how the middleware works and download it as well.

And then if I'm on the agent builder side, does this sort of become another tool or even another sort of MCP type construct that I give to the agent? So I would say, hey, you can call this inference provider, you can call this search API, you can call this weather service, what have you. Any of these things might return to me a 402. And then I also have my payment tool, which, having received the 402, would then— my AI would know to call this payment MCP and send the money and then loop back.

Totally. By the way, we talk about MCP. Imagine all your listeners know it well, but MCP is one example of how we're rewriting the web to make it easy for AI. The web was made for humans. And so MCP, one example of that, we want to make it easy to be able to discover the different APIs and what you can do with it. That's what MCP was. But even MCP is missing the payments layer, right? And that's essentially what X402 adds on top of MCP. So we see a future world where every MCP server is basically hooked up with X402. And when there's a paid API, they can basically do that. And so some websites like, I don't know, the New York Times, maybe someday will charge you when you're crawling the New York Times. If you're an AI crawler and you want to be able to do that, it may charge you. You know, that's the way in which it like enforces it and does it. You can't access the webpage if you're an AI agent unless you pay, as an example. And then absolutely is that you brought up this idea of discovery, right? Which is that like, how do I discover all the x402 APIs that are out there and use them? And today the way we started was we just have a page on x402.org. So any of your users can check it out. I think it's one of the first links on the page and you can see the ecosystem. We're starting initially with the crypto community. We think there's an inference of the 2 key areas that we're focusing on. In the crypto community, if you want a crypto API, there's a number of different teams like Nainar and others that have integrated it. Then there's inference providers that we're also exploring and talking with to be able to add their capabilities. But you can see discovery of all the different things that are out there. But over time, where this gets even more exciting is probably have an MCP server, which is like, here are all the xfer2 endpoints that you can access. Here's a description of them. Here's why it matters. And then the AI agent can decide, do I want to go buy from this one or this one? And maybe they're competing inference providers, et cetera. But again, we just launched this just a few weeks ago. So this is just the start. And with like any standard, you know, you start at one layer, you start getting people excited about certain use cases. And then over time you build from there.

I guess, One thing that seems like it could be still a challenge, but maybe also could be a huge unlock, is the inference providers. Certainly your Anthropics and OpenAIs and presumably Google, although I haven't hit this myself yet, are starting to do more of a KYC type of paradigm. And so I wonder, is that a barrier or is that something that this can sort of help with in a way? Because you've got lots of people who may struggle to meet the KYC requirements. And maybe that's— yeah, I just don't really know what to expect as the intersection, because the notion is generally just like, the models are getting really powerful. We're going to want to know who's using them and for what. They're starting to implement all these account-level monitoring type of things. Would they be able to do a version of that with these protocols driving the the transfer of funds, or is there an incompatibility or maybe like a missing extra component to the system if you want to do that?

I think that's independent in that like Claude, OpenAI, anyone can add, that's their decision in terms of do I want to gate, you know, certain functionality or features behind KYC or not. So I would say that that's independent of this, but where I think where Xfer2 gets interesting is that like, I think it's the lowest level of like rails. Right? So you can imagine things on top of it, for example, like reputation. There could be a reputation system that's built on top of this. And so, for example, if I don't have the money, could I use that reputation as essentially a credit score for an AI agent so that they, they want to pay for some API, they don't have the funds, but I'm basically floating their credit. And there should be a world of like, basically like a reputation system on top of that that allows you to be able to do that. So that's something which doesn't exist today. And X4 too, again, it's just the payment rails. But as you know, like, just like with credit cards, just like with other things, adding reputation in addition to a credit card network is a really powerful thing. And the same here is that we absolutely see a world where, like, over time, people will start adding those primitives in. And by the way, the blockchain in general, there have been a lot of different attempts at building really powerful primitives around reputation. I think it's just a matter of time where, like, people have reputations, but also AI agents have reputations. To be able to do things. And so they might get credit at an API provider, they might get access to something just because of their reputation. They've done good things in the past, they haven't done things in the wrong way, they maybe haven't scraped a website in the wrong way. And so as a result, they have higher reputation and higher capabilities as a result of it.

Yeah, I'm really interested in this reputation for AI agents notion and even just the provenance, whose agent is this? On whose behalf is this acting? And it strikes me as a kind of challenging problem because try as I might, I haven't figured out a way to draw a real clean box around an AI agent. Sometimes if you construct them very simply, then it's pretty clear what the agent is. But an example that I kind of keep chewing on is another past guest company, Augment, which is doing AI coding assistance but focused on large enterprise codebases, put out a coding agent that used in turn a smart MCP. And the smart MCP was provided by another open source developer, and they used that MCP to do the thinking. So it kind of maps onto the schematic that you had outlined before where you have an AI, but it might want to go get more inference. That's essentially the coding agent is going out and saying, I need help planning all the code changes I'm going to make. So it has a planning tool presented to it in the form of an MCP. There's going to be some inference there, presumably. This thing happens to be open source, but it could obviously be closed source or proprietary in any number of ways. So it might have to pay for that in order to get the thing on the other end to do the planning for it and then get its plan back. But now it starts to become like a very weird thing where it's like, what exactly is the agent, right? Is my thing the agent? Is my thing plus that thing the agent? Do you have any extra clarity on this? Because obviously funds and who controls what money and who's responsible for what is important and possibly could be clarifying for this sort of thing. Do you have any sense for how we can get crisp on this? Because for me right now, I am expecting to see just like a total smear. And the sort of forms that I see people outlining to me feel like more sort of ends of a spectrum, where the sort of middle of my agent, but also a smart MCP, kind of blurs where is the intelligence coming from and who's ultimately responsible for what. And if the thing gives me a bad plan or acts maliciously toward my agent, who's really at fault for that? I just feel like we have so many unanswered questions. But I wonder if your study of all this from the financial side of it has given you any rules of thumb you'd like to see become best practices or any sort of clarity, really.

So, like you, I'm seeing this and then marvelling at it, right? The behaviour that's coming together. I think one model I just historically have seen, and I'll give you the reason why, is that just generalist and specialist AIs. So like AIs that are like custom made for a certain purpose. And so like one example we've seen is like web crawling. And then there are other AIs that are generalists that want to tap into those specialist AIs. And so this is where I come in, is that we've seen at Coinbase Developer Platform multiple examples of AIs paying different AIs. And to a lot of people that like breaks their brain, probably in the AI community it's not as surprising, but like, like most people in the world, it's like, oh, I thought that AI was a thing. And to me, there might be multiple layers behind the scenes exactly like you're saying. And I might interact with one and really I'm just talking to one person, like one AI, but the Behind the scenes, it's hitting MCP, which is really like forms of data, maybe. It might hit other AI agents who are super specialized. And so where XRO2 and Coinbase Developer Platform come in is we think that like you need a financial layer to incentivize that, right? That like if one AI is asking another AI, then like it probably is not going to, you know, it's not going to get charged by an ad. There's some form of value transfer that needs to happen there. And again, if it's the same company's AI. It's super simple, but if it's multiple companies and multiple different agents that are out there and more specialized and more generalized, um, in different context windows, and they all want to interact together, we think that the thing that's missing is really just money, right? Some form of financial exchange that can go on. So that's like one part of the solution I feel is that you have some form of money movement around that allows you to be able to talk to generalists and specialist AI agents, get the information from them. Maybe it's like some MCP servers are free, but other parts of the MCP server, just like, you know, APIs are going to charge you money and you pay for it. So that's like one whole set of like conjectures. The only other thing I'd just tell you is that like the blockchain is a perfect identity layer. And I think this is like for, like for anyone in AI building, I think this is an area where like, I always love bringing the AI and crypto communities together, but an example is that like the blockchain, like If you have an NFT, that essentially is a form of identity, right? So like, think of it as like creating a user, like on the blockchain, I can create an NFT. And by default, that user is empty. There's nothing in it. But then if I start doing things, let's say I'm that AI agent and I, I like correctly deliver medical advice as an example. A user who's seen that information can like tag my NFT and say, hey, this is great. It's really good medical information. And they gave me the right thing and I confirmed that it was the right thing. And other people can do the same thing. So think about like a passport where you're getting stamps, right? So like you have, like the AI agent has an NFT, it's getting stamps more and more. And maybe all the stamps are on medical information and suddenly over time it's seen as a credible source for that reputation. So again, this is not to do with X402, but in general, what we're seeing in crypto is a lot of innovation on the identity stack. And again, like the blockchain, it's like, think about it as like an open source credit bureau. Right, where that information is available. Anyone can add information to it, and all you're looking at is like, who are the people adding information? Do I trust them? And are they adding a lot of positive signals? If they are, I'll maybe give more weight to it. And hey, they did this one thing bad. It's kind of like a review system. You know, they've done this one thing bad, or they did mistaken advice this one time, but 999 times out of the thousand, they did it right. So to me, like the two things I see is like, blockchain is amazing for the payments layer. I see it also as amazing for that reputation system. And this is all possible today. None of this is sci-fi. I think the sci-fi part is how do we bring them together and which use cases do we start at? So again, if there are entrepreneurs listening today, I think this is a really interesting area of innovation and exploration.

When I set up a wallet, presumably I'm not going to give my AI agent my main wallet with the bulk of my funds. Because we've got both much harder, if not impossible, to reverse transactions, and we've got all sorts of hard to predict and easily hijackable behaviour on the part of the AIs. So I guess my first instinct would just be like, limit my downside risk by keeping the deposit into the AI's wallet small. And as long as this is microtransactions and it's like paying for information here or there that it needs to accomplish tasks and it's all very small, then whatever. If I drop $10 into the AI's wallet and somebody prompt injects it into sending that full $10 to the wrong direction, I can just live with that. What more do we have in terms of keeping— because this also intersects somewhat with another idea that has been floating around, and I mostly associate this with Tyler Cowen, who's been saying that He expects we're going to need to have AI agents somehow capitalised, that there's going to need to be like a financial stake so that there can be some redress if the agents go wrong, that they're not just totally flying around willy-nilly, but they've got to put basically some skin in the game for the agent, or obviously the parties that create the agent. But presumably that would need to be a lot bigger than the $10. So I feel comfortable with, like, all right, I can drop $10 into this agent or even $100, but I'm not going to put major funds into it today because I don't really trust its judgment that much. And I also know that they're not adversarially robust. But then we've got this other idea of maybe they really do need to be capitalised in order to be able to go around and create the right incentives for people to have the right guardrails and people to have recourse if they do mess things up. So, yeah, I guess 2 questions are like, what are the best practices? How do I keep myself safe when I do this? And does this extend to that sort of capitalised agent concept, or would you put that in a whole different bucket?

Yeah, I think it does extend to the capitalised agent. So let me take a step back. You mentioned, hey, I might want to put $10, $100. I'll be upfront, I think people are going to put millions, if not billions of dollars into AI agents over time. And just to give you a really simple example, I want to be able to trade on chain. Like, I want to trade on the blockchain. I want to trade the stock market. That's a lot of access to funds that you're giving an AI agent. And we're seeing that people want to be able to say, I have an AI agent. It's a trader agent. And we had like hundreds of people at our hackathon that were basically like exploring ways to build basically like a wealth manager, right? Like I can, I give it money and it can go make more money, you know, from this. And like, it sounds like science fiction, but the flip side is like, there's tons of quant funds on Wall Street. You know, you and I probably both know that like, they do this already. So the question is like, how do we get to that world eventually? So to your question, absolutely. One way is segregation, right? That's the way you defined it as like, can I put a small amount of money aside, only have it access that? But I think like beyond that, what you need is some form of permissioning, which is that like, if it's more than a certain dollar amount, please like escalate it and I have to like sign off. for the ability to do that. And if it's, you know, like, I'll let you do a certain degree of transactions, and then the more and more that you do that are right, then I'll give you a little bit more and more credit line, essentially, for that AI agent to do. So the 2 things we've been seeing is basically segregation, and then some form of credit limiting process. And it's, it's not just credit limiting, but it basically is like a sign off, you know, like if the agent, like, is doing the right thing or not. And so like, let's say the AI agent wanted to do a trade. I can look at that trade and say, does this make sense? Does it not? What's the rationale for it? Okay, sounds good. Go ahead and do it and then give that permission. So that's part one. To your question on the Tyler Cowen thing, I would say like crypto has a form of this already. It's called slashing, which is that like, so all, many of the newer blockchains are all something called proof of stake, where I have different transactions. I put some money at risk. So I put my $1 million, $10 million, $10. And I earn some yield from putting that money at risk. But if I do the wrong thing, I can get slashed by the rest of the community. And so I lose that money. So this is exactly the same model that you can have for this is that the AI agent has that crypto, right? By the way, the cool thing is these are all primitives built on the blockchain. So the AI agent has that money, it puts it into escrow. And by putting it into escrow, it can do certain transactions that it might not be able to do. But then if it does the wrong thing, suddenly that those funds are slashed and taken away. And absolutely, I think the blockchain is a really great element in which this, and again, all of this requires like movement of money, which is what X402 is. And then you can build all these intelligent primitives on top, like reputation, slashing, et cetera.

Is that slashing ultimately a social process? It's a matter of like consensus in the rest of the community that this entity, whether it be human or AI or otherwise, did wrong. And so we're taking your funds. Like, it sounds almost like a jury, a trial type function.

Yeah. So on blockchain, that is the case, but it doesn't have to be the same for what AI, like what mechanisms you do for AI. So what I just say is like, that's one inspiration, but it could be like a centralized party, a single centralized party is the way we could architect this for AI. We could do it as like multiple different parties and they could all be trusted, right? It doesn't have to be just a random. juror of folks that are out there. On the blockchain, the way it works is like, it is the community. And again, different networks have it different ways. But there might be like different leaders that are elected through some decentralized governance vote. And as a result, they're the ones who get to decide whether something's right or wrong. And if they decide wrong and it's ultimately shown to be erroneous and they all get slashed is like a way that it's done in decentralized systems. But absolutely, like we could design it in any way that we wanted to for AI. And a lot of times what we see with these systems, like in crypto, it's like they always start decentralized. And then there's some like trade-offs of like how decentralized there are. We've also seen like we talked about stablecoins earlier and that's a centralized form and it could be a centralized form of jury. Like there might be a company or a person who like attests to is this the right action or is it not? And people trust them so much because they've done such a good job and they haven't like gone against that, that social goodwill for a very long period of time that they're allowed to keep doing that. So I wouldn't say that you have to follow what is in the blockchain, but the cool thing is. all the primitives to allow that slashing are possible because like blockchains today, like they have a form of like smart contracts and think about this as a legal contract, but it's a legal contract that the computer follows to the letter. Right. And so however it's written in code, that's what it follows. So if that's the like slashing mechanism, whatever that is, you can guarantee that it'll be followed, which by the way, like in a world where every country has a different legal system, where we may not agree on the laws and things like that. The power of these smart contracts is they're the same. They enforce the same the world over. Right. And so it could be a really powerful primitive for this.

Yeah. Well, let's come back to smart contracts in just a second. In terms of escalation, you had mentioned creating rules to govern my AI agent's payments. If it's above a certain amount, then escalate that to me. That obviously makes a lot of sense. I assume that would be something that would be done on the wallet side in sort of a hardwired way. I don't just want to be prompting my AI agent, hey, by the way, in general, if it's more than whatever, you should come to me, right? I think I, at least for now, want to have a much firmer boundary for it. So presumably I can set that up in a Coinbase-powered wallet, no problem, right? Do I have that understood correctly for starters?

Yeah. The power of— again, this is a smart contract. It's just a form of what are the rules to let me do something. The way we would design this in code would be the AI agent would be able to spend any dollar. Let's say it's up to $10 they can spend without getting permission. If it's more than $10, if they try to do it, it fails by default. But if they want to do more than $10, the way it works is like, oh, bring your human and they have to approve a certain thing using their private key, right? So like the AI agent might have its own wallet, but the human has their own wallet and these 2 wallets, like both of them have to say yes. And then only are the funds spent. Now, the reason you want to do it, like you want to do it not only at the AI level, like in a perfect world, the AI agent is prompting you, right? So it knows a stock, right? And the AI agent is saying, hey, Namil, I would like to spend this, but I need your approval. So probably you want to tell the AI agent, hey, but for more than $10, let me know. But you don't want to trust that on its own, right? There could be hallucination. There could be other reasons why the AI agent may use that. And so you have it at the wallet level as well. And so the wallet, basically it's what we call a 2-of-2 multisig is what we call it. Think of it, most people know it kind of as a safe deposit box. I have a key, the bank has a key, you know, in this case, the AI agent has a key. I have a key and the 2 together are required to be able to open it. Same way. That's the way we'd architect this.

Yeah.

Okay, that makes sense. How about more on a systemic level? One of the most fascinating bits of research that we've covered on this show in recent months is a paper that looked at different models and their ability to learn to coordinate with each other as agents over multiple generations, or their failure to learn to coordinate with each other. The result was, and I don't think the result actually matters too much, but Just to show that nobody really knows what's going to happen in some of these dynamic systems that we're starting to develop. Claude was able— this is a prior version of Claude— was able to play this donor game effectively with itself in a way that multiplied its funds over the generations of the agent and kind of settled into this cooperative equilibrium. And the core mechanic there is that one Claude could donate to another Claude, and if it did, then the receiving Claude would get twice as much. And so the Claudes were able to do that with each other and kind of not only take the prosocial action, but also punish defectors and kind of get into this equilibrium where the majority of Claudes are doing the donation and their collective wealth is rising. GPT and Gemini could not do that. Interestingly, of course, my sort of all angles are like, all possibilities are probably going to become real outlook is like, that sounds amazing. Could be amazing. It could also— the flip side of cooperation is collusion. So I could imagine AI agents cooperating with each other in extremely prosocial ways that I would approve of. I could also imagine them colluding with each other in all sorts of ways that might be really problematic. What do we have, if anything, at a systemic layer? I mean, I know obviously the original vision of Bitcoin was like, nobody can shut it down, but I'm a little bit nervous about the notion of creating a vast sea of autonomous AI agents on a foundational technology that nobody can shut down. And I think about episodes like the famous flash crash, which I don't think is necessarily a great analogy for what the AI agents might get up to. But the key point is, when the flash crash happens, there is somebody with a breaker switch that can say, hey, this whole system is going totally haywire. We're going to shut it down, look at what's going on, And we'll kind of come back online when we have clarity for people so this doesn't run away too far from everybody and make just a total mess. Do we have anything like that in the sort of crypto-enabled AI agent economy? And if not, is anybody working on that? Because that seems like it's probably going to be something we will need or we will want in some way, shape, or form.

Yeah. So what I would say is, first of all, crypto and AI is At its most nascent stage. So like before we launched AgentKit, I don't even think there are many like AI agents with a wallet, like maybe in a handful of like AI agents in the world had a wallet. So really we're at the early stages of this. I would say a lot of things we talked about are the things you need. You need some form of reputation, which makes it much, you know, like for an AI agent to do the wrong thing, there's like social or impact financially to that AI agent to be able to do it. So that's like one part. 2 is you have humans in the loop is a really, another really powerful primitive as part of this. The thing you mentioned, the flash crash, I think often where that would come in would be like the networks themselves. So for example, in this case, it might be the Uniswap, which is a decentralized exchange, or AirDrop is another decentralized exchange. These are all examples where they might have something where if it's more than this many transactions on this book in one block, we will want, you know, basically circuit breakers. I think that, you know, you mentioned the flash crash and I'm like, Yeah. That's basically a way to build a circuit breaker into the system as part of this. And so my sense is that like, as these things come together, we're going to be coming up with many forms of enforcement, like mechanisms. And probably the easiest one is just a human, right? Human in the loop. But at some point, I think like these AI agents get so good that humans may not want to be in the loop. It'll be super like, kind of like those pop-up notifications you get where you just get too many of them. And I think humans may kind of abdicate a little bit of that responsibility, especially for lower dollars and things like that. So I think in that world, I think like reputation to me And having some other form built into the protocols themselves is going to be super, super critical. But like, I think that like all of this, like we dream about, um, but it's just like until the amount of volume of AI agents doing things on the blockchain go up, I just don't see that happening. And I think the 2 things I see are just like, one is payments seems obvious. Like agentic commerce is clear. People already are asking their agents to buy things. So X4O2 is a really powerful thing for that. And I think the second thing I see is just what we talked about on reputation. Is that the blockchain is a perfect place to build reputation. It's an open, like an open access, anyone can do it. And once you write it, you can't just change it. Right. So it's not like a centralized repo. So for all these reasons, I think it's a really good reputation layer. And I think those are probably the 2 starting points for this. Gotcha.

Okay. Yeah, I think I'm going to want some of those circuit breakers at some point.

I agree.

We're not quite there yet, but it's funny. I think the models are getting there. The adoption, all the dynamics are not there. But I don't think we need anything really, honestly, all that much more powerful than your Claude 4s, Gemini 2.5s, and your O3s to really create a very dynamic agent economy. It's maybe a half generation away, but it's really not very far away. Honestly, I've been feeling over the last 10 days that Claude 4 is smarter than me. I have to admit it. And maybe O3 too, I don't know. But it is really starting to feel like Boy, these things, they have it where it counts, and they don't have all the affordances, and we haven't created what you have created, but they haven't been adopted. But when they are, it seems like it could go really fast. One thing I've been looking out for a long time, and I haven't really seen, maybe you have seen, or maybe you would know why it's not happening yet, or could point me to something, but I've had this notion that maybe AI puts the smart in smart contracts. Where what I mean by that is obviously what you can put into code, in fully explicit code, is limited, right? And that creates sort of a very sort of brittle nature to smart contracts. It doesn't feel like you're going in front of a human judge, right? It's like a much more sort of mechanistic, deterministic, and ultimately it can't really resolve disputes between people. On the other hand, AIs are getting there where you might say our smart contract resolution mechanism is that we will call Claude or we will call Gemini 2.5 and we will submit the evidence to it and we will let it decide what is right or wrong. I think about just like local services. I've got a 100-year-old house and when something needs to get fixed on it, it's not easy to find somebody necessarily. The reliability factor there is not great. But it seems like we're getting to where something like that could be possible, and I just haven't seen anybody come forward and say, here's my arbitration on the blockchain powered by AI system yet. Are you aware of anything that I'm missing? And what do you think is the sort of prospect for that kind of future?

I haven't seen anything. And I think the exciting thing would be if someone listening to this podcast went and did it, because it's all possible today. What you're saying is just like a form of multisig, right? Where one of the people who holds the multisig is the AI agent and they may make the decision. So as an example, it could be a system that's like 2 of 3. So let's say there's 2 humans in the loop. They don't agree on it. So only one of them agrees that, you know, the funds, for example, should be released. And another AI agent is there and decides to be able to like weigh in, make a decision, and they're the second one. And that's it. decision made. I'll give you another example that I really loved earlier in my career. So I used to work, I worked for a few months at the Gates Foundation and we worked on something called weather index crop insurance. And so the way this stuff works is that if the rainfall is less than a certain amount, you automatically pay out. The question is, how do you figure that out? Right. And that typically is some form of like Google that needs to like align and make a decision on whether the funds should be released. So the cool thing is you could pay for like this growing season. You have $1,000, money's there, investors are putting the money in, and the money goes to the investors if the rainfall is more than a certain amount. Um, and if it's less than a certain amount, then it goes to the farmer themselves. This is a really great example where AI can be plugged in. All AI has to do is say, like, on this date, the AI should weigh in and, like, review the precipitation of the course of the last 6 months as part of that, and then unlock the funds and let it be on its way. And so all this is possible. Like none of this like is required. And I'll say one other thing for folks on the AI side is that there's this interesting thing in crypto where oracles is the hard problem. So blockchains can, and smart contracts can only access what's on the blockchain. And there's a lot of things in the world, like weather information that's not on the blockchain. And so they use something called oracles to do it. You're essentially like, the idea you came up with is like, let's just use AI as that oracle, right? And like, we should trust the AI. Hopefully the AI doesn't hallucinate and things like that. If we believe in that AI and we think it's like a pretty high quality, then that's a really great example where suddenly the AI is the oracle. It can decide how to vote using its key whenever it needs to be pulled in. So I would just say it's like, I haven't seen it and it would be amazing. This is super, like relatively simple to do. So if anyone here wants to reach out to me on Twitter or anywhere else, super excited to like give them details and help them get set up if they wanted to do this on the AI side. But basically you need an AI you trust and you need some type of problem where you want it to weigh in. And then suddenly the AI can absolutely do that.

Do you also need— it strikes me that AIs are very gullible. I'm a close follower of the AI bad behaviour research, and much of that research is done by telling the AI that nobody will read your private scratchpad, so you can think about what you want to do there, and then you can take your action here. And then, of course, the researchers are reading the private scratchpads, and that's where they're seeing some of the evidence of the bad behaviour, but just more generally, right? They're easy to trick. And so a hard thing from the perspective of the AI would be knowing what constitutes authentic evidence, what is made up. Is that really a picture of your roof? Was that picture taken now? Was it taken a month ago? You're saying it's not fixed, but I can't tell if that picture was taken today or a month ago. So it seems like we may also need some sort of provenance technology, like on-device signing of photos, or in your weather example, it would be like some sort of indication that a sensor in the field is actually in the field where it's reported to be. So we can trust these raw numbers, and then the AI can sort of provide the analysis on top of those raw inputs. But we still need to have some— the AI needs confidence that it's actually dealing with real evidence that it can reliably reason over. I think we have a little bit of that, but not much. How would you characterise the state of how much we can demonstrate to AI that the evidence we're submitting to it is in fact authentic?

Yeah, and I think there are 2 parts, right? One is the evidence we're submitting, is it authentic? And the second part is, is AI hallucinating or is it being goaded or anything else in a certain direction? And by the way, I've seen AI agents that have Twitters, right? And like, we've seen examples in the community where people have, like, the AI agents have mistakenly basically been like hacked or goaded into giving out the money in a place, in a way that it shouldn't be. So I would say both are critical. Both are important to get right. But like, with any like form of technology, I think you start somewhere and where is possible. So like the satellite, the example of like the rainfall, the idea that like, oh, in the course of this growing season, is there enough rainfall? The cool thing is that could just be satellite data. Like you don't need a huge, like you're right on the roof example, is my roof that or not? Absolutely. You have to figure out provenance. You might want a third party actually to take the picture rather than the person who's like insured as part of that. But the weather indexing example is like a little bit satellite data that Google has or anyone else has, or there's a satellite company with an API that you use Xpert to get. These are all examples of how you could do it. So my instinct is like, those are things that like to get this perfect and do it at scale. Absolutely. You need to solve both of those problems over time. But like the hallucination could be solved a little bit, especially, sorry, the goading problem, which maybe is even a step further than that, could be solved by not putting the AI like online, right? Like only having like an expert be able to access it and parameterize it or things like that. And it doesn't accept other inputs, you know, after that. But then like when you have that satellite information or things like that, those are ones where they're very like precise and they're not super complex problems. And that's, I think, where a lot of this stuff starts. It doesn't start with the hardest problem on day one. It starts with like very like specific problem. Hey, Google has the data and there's a human with some funds in a wallet. Let's combine those 2 together. And like that works together.

Cool.

Well, I really appreciate your time and I know we've run a little long. Maybe just final opportunity to take the floor if there's anything we haven't covered at the intersection of AI and crypto that you'd like to highlight or, you know, any just, um, Elaborations on the positive vision that you'd want to try to inspire people with?

Yeah. All I'll say is we're scratching the surface on how crypto and AI can come together. I think crypto is a really powerful financial system for AI agents, and x402 is just the start of that. You and I talked a little bit today about things like reputation and other things that haven't even been really tested yet. So I would just say, if you're building an AI today, I'm super excited for you to be able to try out x402. That's x402.org. You can Also DM me directly on Twitter if anyone wants to be able to talk more or engage on anything. But like, there's so much innovation possible because we're finally bringing these 2 technologies together. And so super excited to see what people build with it. Cool.

Nimil Dalal, Developer Platform Lead at Coinbase and one of the lead developers of the 402 Payment Required, AKA X402 protocol. Thank you for being part of the Cognitive Revolution.

Thanks so much.

If you're finding value in the show, we'd appreciate it if you'd take a moment to share with friends, post online, write a review on Apple Podcasts or Spotify, or just leave us a comment on YouTube. Of course, we always welcome your feedback, guest and topic suggestions, and sponsorship inquiries, either via our website cognitiverevolution.ai or by DMing me on your favorite social network. The Cognitive Revolution is part of the Turpentine Network, a network of podcasts which is now part of A16Z. Where experts talk technology, business, economics, geopolitics, culture, and more. We're produced by AI Podcasting. If you're looking for podcast production help for everything from the moment you stop recording to the moment your audience starts listening, check them out and see my endorsement at aipodcast.ing. And thank you to everyone who listens for being part of the cognitive revolution.