Ep. 749 Coinbase’s Head of AI Explains the Future of Agentic Finance

0:00All right, everybody. Welcome back. Welcome back to another episode of the crypto 101 podcast. I hope you're having just as fantastic of a week that I am. Brendan, how are you doing? Extra fantastic as well. I mean, how can you not be? How can you not be when you're covered in the blood of bears with the largest short squeeze by several different measures, the largest short squeeze in the history of crypto from a commercial volume perspective? I saw the first hour. I mean, all of this was driven off the back of the news of the the secretary treasury saying he's going to essentially print more money. He cloaked it in a fancy way by saying he's going to issue a lot more debt on the front end of the curve with two year bills. He's going to buy back a bunch of these off the run, you know, 20 year and 30 year bonds adding more liquidity to the whole landscape there.

1:09And then just the other day saying, well, we've also got the treasury general account. We could run that down and buy back bonds. And so they're essentially monetizing the debt. The currency takes the hit off the back of that. You know, the markets are to sniff out the seriousness of their plan. There was a billion dollars of short liquidations over just the one hour following that announcement over the following 24 sorry over the following 48 hours. I think it totaled somewhere around $4 billion. And so it was a really, really, really big short squeeze. We were happy to see it. Starting to get those texts from friends saying, oh my gosh, we thought you were we thought crypto is dead. We thought you're going to have to start a new podcast. We thought you were going to have to move to Kentucky or something. No shade of Kentucky, but cheaper costs of living in San Diego. But sure enough, the crypto bulls have, you know, have come back with a little bit of vengeance. And so we're really excited. I know everybody who's watching is really excited.

2:14But we have an incredible guest who I'm sure is sharing in the celebrations here in the crypto markets. The head of AI product at Coinbase Lincoln Merr joining us. So Lincoln, welcome to the show. How are you doing? Doing well. Thanks for having me. Yeah, pretty stoked about everything that's happening over the last couple of days. Heck yeah. No, there's been a ton of fun stuff. And you know, AI is at the epicenter of a lot of this excitement. Of course, everybody is now using AI in their workflows, whether it's to analyze the crypto markets, whether it's to get news. I even know folks who are using AI MCPs to execute trades. And they're able to manage their portfolios. They're, I mean, I'm seeing people manage their wallets through different platforms. So there's so much unpack. This is going to be a really exciting episode. You know, we've got a lot of questions to cover. But first I just want to, you know, again, this is a show about people.

3:22When we always platform, you know, just the smartest best builders in the space. And that's why you're here. And so we want to know a little bit about who you are. Before we dive into what you're building and, you know, that kind of stuff. So just tell us a little, you know, high level on how you ended up at Coinbase as the head of AI product. A little bit about your background. And then we'll dive into it. Yeah, totally. So I first got in a crypto years ago. I was 11 or 12 and I was looking for a way to buy video games on the internet. And yeah, I need to get a VPN. So when keep getting notices from the internet server providers that we were breaking other terms and whatnot. So I traded and target gift card for Bitcoin on local Bitcoins. Had it on my iPhone for us wallet. And that was kind of my first experience ever really interacting with money. Flash forward a couple years later, I checked that wallet and I thought I spent $20 but there's somehow $20 in there. No idea how that happened. But sort of got reininterested. Sort of poking around around the 2017 bull market starting and never really looked back. Did a bunch of research, wrote publications, articles, whatever during college,

4:25Ramney University's blockchain club and did some research with the Ethereum Foundation on consensus mechanisms to wrap things up there. Right out college started at Coinbase working on our wallets and all our different sort of APIs and different wallet related developer facing products and builds out agent kit, which was our first early attempt at building out a way to give AI agents access to crypto wallets. Posted that on my Twitter account and went viral. I got a team to build around this whole agitik, agitik narrative and whatnot. That turned into X402, which turned into Coinbase advisor, which turned into all the way they were doing today. So yeah, now I'm here. So is it fair to say that you're the inventor of X402? I would say co-creator, I would reserve the inventor title for the engineers who actually wrote the spec and did all the heavy lifting to make it such a compelling standard. I did a lot around the product side of it. What are the things around it? What are the features and tools and capabilities that you need to actually build on top of X402? Wow, I love it. I think that's a great place to start because this is really one of the main avenues for agents, chat GPT or Claude or many of the others that are going to be coming out, Grockbot is something I've been tinkering around with now, which is super cool.

5:44So a way to give them the ability to spend money on the internet without giving them your credit card, without having to trust them with some large amount of money, you could just say, hey, use this X402 protocol. And you could give them $20 of USDC and say, that's your allotment for the week. Use it wisely. You could give it an allowance. And very interestingly enough, as well, you don't need to give them any Ethereum for gas fees. It's kind of all baked in there. So walk us through, I mean, the vision of X402, how it started. And again, color it with the idea of people who are listening, they're not technologists in a sense that are heavy coders generally. They're people that are users, they're consumers, they're retail investors, speak to them, particularly about the power of X402 and what it paints a vision of the future for. Yeah, totally. So when we were building out agent kit and some of these early sort of tools to give agents access to wallets, it was extremely experimental and we were mostly just poking around.

6:52And what we needed of fighting out is that we can tell our clients and our chat chubb teas to go do on-chain oriented things. You can go make a swap, you can go deploy a smart contract, make an NFT. All of this cool stuff, but kind of inside the coinbase and the crypto bubble. And really what we wanted to do is take the idea of an agent having a wallet and actually make it be able to go off and pay for things. So I think one of the clearest pain points that we're trying to solve is how today, if you want to interact with all of these different services that exist on the internet, it's a really frictionful process. So for example, if you want to set up your agent to have the ability to generate images or videos or do a web search, you have to go find a specific provider that'll do that service. You have to sign up for an account, you have to fund that account with a debit card, and then you have to get some sort of API key or basically like a unique identify that you give to your agent to authenticate it on the service. It's a huge pain. It'd be analogous to if you went into a grocery store and before you were able to check out, they make you buy a gift card to that grocery store and sign up for a grocery store account. It just doesn't make sense.

7:55So the whole idea of X4.02 is to let an agent go, hey, that's something cool on the internet that I want to pay for. How can I do that? And then that website response, oh, here's all the information that you need. I charge 10 cents of USDC on base and the agent can just send that money and it's easy as that you can access whatever you want on the internet and really help facilitate this new agitic economy in this era of machine to machine payments that are increasingly inevitable. Yeah, so awesome. You know, what are kind of some of the example use cases of like what is an agent spending money on? Is it? And my wife and I talk about this all the time because we're really passionate about like, you know, where the future's going robots in the house or she's like, oh, like maybe one day in AI agent, like it knows my preferences, it knows where I shop. Can I just give it $100 and say, you know, wait for this dress to come on sale and be able to automatically execute that? Or is it really more just like, you know, hey, I want to subscribe to a feed of data for my stock portfolio.

8:56Micro transactions, you know, every, you know, every day, pay it, pay this data feed, you know, five cents, like where on the spectrum or are we not thinking even big enough, maybe. I think it covers really all the above is kind of the basic answer where we've actually started to see real traction, especially with stable coins and crypto is on this micro transactions use case. And I think that's because it's particularly uniquely enabled by stable coin and crypto rails. And the sense that traditional card and payment processing networks have minimum transaction size and transaction fees that are prohibitive to the division of micro transactions on the internet. And this general developer oriented audience that's currently playing with agents are more prone to these like individual services that might be interacting with. So some concrete examples I've just given my quad agent a wallet and everyone's well, I'll ask you to do something and then after it finishes its whole five minute thinking process, I'll realize that it actually use some of that money to go off and do something. So some of the examples are I was writing up some report and I needed to send it to somebody. So it paid for a service that a lot of you just host a PDF on on a website. And I just send that link to anyone and they could go view it.

10:04And other time I was trying to get it to access what's that service called let me just write that down real quick. Yeah, it's called stable upload solid, sorry interrupt, but I'm always no yeah. So then interact with some other services like a fire crawl or exa, which allow you to scrape website information. So this is particularly valuable if I tell my agent to go do research and it gets blocked by some sort of bot or paywall. It can just pay for access to that content. So those are kind of the organic uses that are starting to pop up and since it's launched about a year and a half ago now it's seen well over 100 million transactions taking place. So this like machine machine micro transaction oriented economy is is really where's where we've seen a lot of the proliferation of this right now. Yeah, this is fascinating and I want to kind of circle back to something that you just said, which is this idea that you know, A.I. agents are becoming the primary consumers of the web. So I've seen you talk about this as well. And if you look at the legacy rails, which is kind of what your point was it's not really prepared or built for these A.I. agents.

11:11It's it's really built for humans when you look at things like card forms and bank hours and the manual K.I.C. and the way that bank accounts are structured alongside the K.I.C. It's meant for us. It's not meant for these and a really cool stat was that I think 76% of agent transactions fall below the 30% or 30 cent card minimum. So like the economics don't even work kind of in the first place. It would seem as if this is kind of that solution. Would you say that that's right. Yeah, absolutely. And it doesn't even touch on some of the other benefits you get here, which is this whole cross border permissionless period of payment never that we're starting to see come to life. Like if I wanted to give my agent a card, there's a whole K.I.C. process might be geo restricted based on which jurisdiction I'm in, but agents are borderless. And I think anyone anywhere is going to create some sort of service that I want my agent to interact with. And you know, as long as it passes some bare minimum sanctions checks, what not that we're doing, I should be able to send that money anywhere in the world, especially if it's a microchance action that should happen fee free instantly.

12:14All the things that you'd expect from a modern payment rail. And I think that's another unique benefit of crypto here. That's it does feel like that's just the way that the world is moving right we've all seen probably a million different sci fi or futuristic shows where people have some form of AI right and it's like getting the groceries forum. It's hiring and making sure that you know, maybe someone is a house cleaner. It's like, okay, well, every two weeks it makes sure that they're scheduled at the right time or you know whatever rights doing calendar management. And it's all these little things. And it does just feel like that's the way that we're moving towards. And so if we truly do believe, which I think most people can agree, like that's the direction the world's going more tech heavy. Then the next proper step is how can we prepare and build and benefit off of the infrastructure because that's really the first step, right. And so that's what this feels like it is is it saying let's prepare let's build out and let's scale the infrastructure and really get this ball rolling.

13:15Yeah, I fully agree. I think one of the fun things that's going to start to happen is you have these agents operating with more autonomy because you give them money and the same way of having access to money on the internet can do so much more than they can without it. You can imagine an agent that eventually runs a business the simple example being a vending machine. You put enough sensors on there that the agent can start to understand, okay, I'm running low on this or there's a maintenance issue. Let me hire somebody to do it. It can start to take on a lot of these tasks in a way that feels truly autonomous. I think we'll get back to this vision of Dows as truly autonomous organizations once we have this AI infrastructure place. So tell me a little bit about why coinbase particularly cares so much about AI agents. So I know one of the main products, I guess you would say that they launched or services one or the other. Maybe it's both maybe AI is blending blending the line between product and service. So I think that's a coinbase advisor, which is the first AI powered SEC registered financial advisor that's literally embedded in your coinbase app.

14:24I would love to know if you had anything to do with that and sort of again, why is coinbase, which you think, oh, there are crypto exchange. What are they doing building AI products? Tell us how they're thinking about it and all that. Advice of it's not a fun. So yeah, that was my project for the past past year. Coinbase has this program internally called next bets and the whole idea is anyone across the company can pitch an idea to the executive team. And if they like it, they give you a team to go fund it and build it out over the past year. So the idea that came to them with was advisor and the whole idea was to make it as easy as possible and to for users to manage their money using AI. And the real core problem statement there is that there's a bit of an orchestration gap that exists today with trading platforms. People have these high level ideas that they want to execute on, but they don't necessarily have the time to constantly monitor the markets and make sure that strategy goes exactly as they think. For example, if you're watching the street of four moves and you want to longer short oil, you're going to have to be awake 24 seven watching the markets, monitoring traffic information, everything that would make your trade market profitable, but it's just too much for most people.

15:31And secondarily, people don't necessarily know how to take these high level concepts that they have and compose them down into financial primitives. So if I tell my advisor, hey, I want to invest in quantum computing or space exploration. I don't really know how what can I do. It can give me assets across crypto and derivatives and prediction markets and equities, all the different things that I don't really have to think about that I can just execute seamlessly. And having it be from the registered fiduciary organization that's SEC license CFTC registered means that me as a user can feel very comfortable and secure in the advice that I'm getting. And I don't have to, you know, jump through the same hoops that I would have to do this on an external interface. So yeah, I think broadly AI and as a neighbor of Coinbase customers getting the best experience for trading and payments, which are ultimately two of the goals that that Coinbase is trying to drive. So cool. So cool. And how's the uptake been have there been people that are like, yeah, I'm going to trust it. I'm going to give it a hundred grand or, you know, whatever.

16:33It's like, I think, you know, most people have a financial advisor or somebody who's helping them plan for retirement. And they like the fact that they could just hop on the phone, they could call them. There's somebody. And you know, to walk them through things. Do you view this? Yeah, like what's kind of like been like the uptake like I think you guys also had quarterly reports recently. So maybe we could look at just like what's been publicly disclosed on on that product. And also do you see people just like having both their advisor and this agent or are people like just dropping their advisor and going full, full agente. So one of the things that I think is particularly interesting about the advisor is that it helps lower the floor of who can now access high quality financial advice. Because typically if you're interacting with an advisor, you have to have some sort of minimum store with the advisor. It could be tens of thousands, even hundreds of thousands of dollars. And that's not necessarily achievable by all of the customers on Coinbase. So we're able to scale and democratize access to financial advice using AI.

17:33And I think there will always be a need for a financial advisor to have, but they're not also online 24 seven. If it's 3 a.m. and I noticed the markets are crashing overnight, I love somebody to chat to even if it's a chat on say, hey, what's going on? Should I be selling? What's what's your take on the markets and the latest? What's going on? In terms of the adoption that we've seen so far in the advisor, it's still super early in terms of our rollout. Right now it's only available to about 1% of our Coinbase one members with a plan to gradually roll out to everyone else over the coming months or so. Across both Coinbase one and potentially regular retail users as well. But we've seen some great traction. One of the early use cases that's fortunately no longer relevant, but was for the past six months is tax loss harvesting. So people would tell their agents, hey, can you sell all my positions and just like, rebuy them back for me. And with no wash treating rule in crypto, it's a great way to say people thousands of dollars very, very seamlessly that otherwise would require a lot of manual effort across all the different positions you might hold. That is fascinating. I heard another example where basically people were talking about they gave agents access to their portfolio and it was like, hey, can you help me optimize this and save money?

18:41And it was like, yeah, you actually own this ETF for this crypto ETF. And you can switch to this other one and it'll save you like X amount because of the custodial fees, right? And it was like, oh my gosh, I think it was on. He was like, overlapped in the holdings is like almost the same. It's like you have like a 99% correlation, but you're paying a higher fee. And so, yeah, AI now that everybody has this data and they can scrape all this data immediately. Like, it's going to be a race to the bottom for fees for ETFs to mutual funds. Because you would imagine that most forms of AI in agents and whatever would make this choice. So those groups out there that have these really Ikecidian fees, what you should have if you have AI taking over people's portfolios is it should be kind of forcing a mass exodus, not to cost fear, but it would be like, hey, everyone all the sudden knows that you're paying a significantly higher fee and that there's other options out there with a near identical correlation.

19:43And so, Bryce, I think you're right. I think that that kind of incentivizes a world where people are more aware, but because they're more aware it incentivizes these people in charge of those institutions in charge to really lower those fees and make them more competitive. Because, yeah, it's kind of the same concept. I mean, Elon's talked about it, but like he's just the loudest mouth. I'm sure there's several other people in the AI space that talk about like the deflationary nature of AI. Like if you could have all these products and services, you know, being created every day, you know, all day, then it drives the cost of product and services actually down. So that's why AI could be a deflationary sort of sort of thing over the long term. And yeah, the custodial fees and the management fees, that's an obvious one for sure. But I, you know, I guess we'll pause there, get your feedback, but there's so many other questions that I just want to just keep going. But what do you think of that concept? I totally agree. It feels like AI in substance is sort of the democratization of software and the ability to compose these different products and services together.

20:47And I think it trickles not only from ETF providers and then provide invest fees, but also just exchanges and venues broadly. If my agent has the ability to go migrate by account somewhere else, that's cheaper while doing the same level of security. Why wouldn't it do that? What's the real moat for these platforms when an AI can go access whatever it wants about having the friction that human traditionally has? Yeah. You know, here's another one, right? Imagine people who just don't have maybe the general knowledge of the crypto space to send things between different blockchains and do this and that. And imagine you have a bunch of fun somewhere and you're like, hey, I want all of this consolidated into USDC in my coin base account or in my, you know, wherever account in my coin base wallet, but it's a bunch of different blockchains. So maybe it knows, hey, the most optimal way to do this is I'm going to transfer it all to this chain, which is the cheapest option, you know, send it all here, convert it there and then it'll all appear in the wallet that way. And it can kind of instantly identify the cheapest fastest option to kind of do all this. And I would say the average person might know that that's an option, but might not know how to do it and might be afraid that they're going to make a mistake and lose their crypto and yada yada, take a bridge risk.

21:56And if this can kind of just be an all in one solution to solve those issues, that's just, I guess, another way that you could see this be taken advantage of. Yeah, totally. There's a lot that we need to do to approve it because right now it's it's regulatory requirement to have human in the loop sort of tap to execute on every individual action. So we're moving toward a world where we can make this more discretionary and the advisor can trade 24 seven on your behalf. But for users who want to do that today, we have like a coin base for agents. It's it's our MCP that lets you just plug into whatever existing agent you have and do all this autonomous trading. So you can run all these fun strategies like what I have right now is I have my agent connected to hyperlicker wallet and connected to the coin base exchange. And on hyperlicker whenever the funding rates on those derivatives are pointed in a certain direction, basically meaning I can short the asset and earn money from shorting that asset. I have my agent do that and automatically buy an equivalent amount of equities on coin base and I can earn some sort of yield while just like my agent is all doing it for me. There's all sorts of fun things like that arbitrage on like polymarket versus call she, you know, there's a million different ways that you can have your agent start to make money.

23:02Holy smokes. I mean, this is incredible because you know, if you wanted to do that, even just, you know, let's call it two years ago or last year, it would be like, okay, well, you got to know Python, a coding language, you have to have like, you know, deep, deep level of proficiency and understanding of all this stuff. Now it's like, bro, do you speak English or you could speak to it in any language, I'm sure it's like, if you could do that and you could come up with the idea, you could become an arbitrage trader, you can become, you know, an automated trendfall or whatever, it's super, super cool. But one of the things I wanted to get your opinion on, it's this theory that kind of popped into my mind a few weeks back and then I heard somebody also talking about it, where like every analyst say every equity analyst or every crypto analyst working on the desk at a fund. I'm sure is using some, you know, foundational model, like, you know, chat, you, T's soul five six or clawed, fable, like one of these things to just like, hey, stress test my idea or, you know, surface some ideas and like, but now everybody who's controlling money and even just normal people at home, like are using this.

24:22And my worry is that it's going to make the market a little bit more fragile because everybody's going to be like listening to clawed, clawed is kind of, you know, the claw that you run is the same claw that I run is the same clawed that, you know, somebody else runs. Now you could put in different, you know, markdown files, you could have different data sets like piped in to like get some unique spin. But do you think like everybody using these for their research is like causing a more fragile market? Yeah, that's a great point. I think, you know, one very coherent way to put this is if everybody puts in their clawed, hey, go make me money. The diverge between those strategies is going to be very minimal. It's going to result in a lot of clouds running the same exact thing. So I do think in some sense that it really does come down to the user and the owner is still not the user to provide the ideas and the like, you know, quote unquote alpha that they have to these agents for them to actually do any sort of profitable trading. I don't know necessarily if it makes the market fragile at the scale at which we're seeing AI or a gender trading today.

25:24And frankly, if it got to that point and I think it would be a great sign that agent trading has a really strong future. And what it also does is it opens up this opportunity that we've been thinking a lot about for how do you give agents access to proprietary data that help them make decisions. So how can I plug in my agent to not only my Coinbase account with the Coinbase MCP, but also can I give it access to Artemis and Missouri and Nansen and Coinglas and all of these different data sources that can help my agent make more well informed decisions. And one of the things we're releasing that I'm pretty excited about is the ability for your agent to pay for X402 services directly from your Coinbase balance. So now I can tell my agent, hey, every single morning, let's see where the smart money is moving on chain and let's just follow them. And I can do that, have it set up, run entirely without me there. I can ask it for information about Artemis and, you know, what the whale movements are or where we're seeing different different money moving around on Arkham or all these different platforms that previously would require somebody to have a decent money. If you're not sure what you're going to do, you're going to have a decent bit of technical knowledge and proficiency to get it set up, not to mission pain for a subscription, all now available to my agent very cheaply.

26:32And that I think is going to be part of the differentiation between various different models and how they make decisions is the data behind them. So as every, like you just mentioned Artemis, and there's, you know, DeFi, Lama and there's all, you know, CoinGecko, CoinMarketGab, all these different data providers, Coinglas. And potentially, you know, X402 could be the death of the annual or monthly subscription, because a lot of the times, you know, if I wanted to, you know, use DeFi Lama or something, I'm like, oh, do I really want to pay for like, you know, monthly, like, you know, whatever it is hundreds of dollars. I just, I really just want this one little piece of data in it. And I just want that one thing. Does X402 enable that on its own or does DeFi Lama also have to like opt into that? Like it's a two sided marketplace, basically, right? Yeah, two sided marketplace, but I mean, all those names that just mentioned are all on X402 today. So you can really just give your agent a wall and say, hey, go check on Nancy, go check on our, yeah, the list goes on and on and on, sorry, all of the big data providers, you know, CoinMarketGab, CoinGecko, etc, are all integrated with X402.

27:44And you can just literally have your HG go pay exactly as you described today. So like, we're seeing some really good traction on supply side for these specifically this trading use case. Everyone kind of is now acknowledging that there's a compelling opportunity to build out this sort of data marketplace. Like, if you're familiar with OpenRouter as this ability to aggregate all the different LOLMs and AI models behind a single platform. I think X402 is an opportunity to in some sense, disintermediate that for more than just the AI industry. And now you can access all of these different things, which are actually better together and easier for a user than it otherwise would be. Is there a resource that shows us like a living, every evolving resource of everybody that you that that has X402 enabled. So it's like, you know, she said, oh, all those names that you just mentioned, they all have X402 enabled. I'm like, okay, I, I crypto podcast, I trade every day like I didn't actually even know that. And so I'm like, I should just is there a website that tracks all this and then also

28:45where's a good resource that you would recommend our listeners who are like, oh my gosh, they're foaming up the mouth. I would love to learn how to do this. Like, what's your guys is, you know, version of the Anthropic Academy or something like that. Yep. Yeah. So the, the two resources that are most helpful, one is just a gintic dot market. It's our website that services all of the different providers that are available or at least like a curated list of them. And then to actually see stats about like what's being used in volume and whatnot, I would recommend checking out X402 scan dot com. In terms of resources on where to get started, there's really two places. One, you can just point your agent at coinbase dot com slash skill dot md until it gets set up with an agent to quality or get your coinbase account ready to go. So she can start paying for these things. Or you can go to our documentation, which is that docs dot cdp dot coinbase dot com. Both of those resources should have everything you need to get started. It's so much simpler than you can imagine. Like once you have a while, it's up. It's like the last thing you never really think about it. Now you're just now your agents ready to go. You do it on chat.

29:45That's a key quad complexity everywhere. How freaking cool. Man, do I do the agents pay taxes or does the human who owns the agent pay the taxes? Human who owns the agent pays the taxes. Unfortunately, there's still a big outstanding question. I know. Yeah, I think identity and taxes are one of the few areas with agents that that there's a lot of greenfield space and I imagine we'll see a very similar standard to what came out with x402 or potentially even from the x402 foundation around identity or how you start to give these agents a little bit more sovereignty on the internet. So I mean, man, we have talked about so many different ways to apply and use agents. Two questions come to my mind is number one, how are you seeing agents being used the most right now? And then do you think that that will change? And if so, how do you think agents will be used the most in five years?

30:50I think the boring answer for today is coding, especially in enterprise settings. At least the one that I see the most in terms of where the most token volume is really happening. It makes sense. The enterprises are very price insensitive. I personally don't care. And my agent is chugging through credits on unbrien's dime. So might as well keep on running those. In sort of the consumer facing world, I saw some chart recently and it's like the top use cases are around just searching as a replacement for Google and companionship and therapy or like an AI friend, which is somewhat dystopian. But still like very interesting. And I think where we are right now, the onus is almost on these chat interfaces, chat GPT, quad wherever people are going to expose some of the more advanced functionality and make it easy and intuitive for users to interact with. People know the chat box because it's what they're used to with Google. They don't necessarily know that you can have it check your email or start to interact with all these other things in your personal life, make dinner reservations for you.

31:52That type of stuff, I think, there is going to be a bit of a discovery process to bring the now billions of users interact with AI into this new paradigm. And you know, we see the frontier of it with things like open claw and and cloud code and those rules, but they're still very difficult to be accessed by the traditional mainstream consumer audience. So I think there's an effort to be done both by apps and plugins and all these different things, but also by these chat harnesses themselves to make that easier for everyone. Yeah. And there's big efforts on the education side as well, guys like us just trying to try to make this as easy as possible for people to access and learn about and do, you know, just just work with trusted products and stuff like that. So man, we're super happy that we were able to have you on a couple more questions for you. You know, stable coins are obviously a big part of this USDC from circle partnered up with coin base. Is there an ability to pay with other cryptocurrencies on x402 say I have Bitcoin say I have salana or some random meme coin.

33:02Is there the ability now or potentially in the future to kind of all integrate that next for a two. Yeah, totally that's that's available today. So in terms of in terms of network support, each of the different coin chain foundations is responsible for their own implementation. X402 itself is actually governed in its own separate foundation. So we spun it out of coin base. And now it's it's a body of coin base cloud flare strike AWS Shopify Mx visa mastercard all of the big names that you can imagine payments processing and related fields. And what's particularly interesting about that is there's this perception that X402 is crypto only. And while it's crypto first, there's absolutely no reason that any of those card providers couldn't eventually add their own network scheme or support X402 base transactions on their card network, for example. And though coin base specifically believes that stable coins going to be the cheapest fastest most efficient option. And today I think over 99% of agitic transactions or at least X402 transactions are happening using USDC because it's what merchants pervert as a stable currency.

34:11There's absolutely no reason you can expand that into anything you want. Wow, so cool. What's like on the cutting edge right now, like, you know, where you're looking at, you know, I always like to ask experts kind of like, you know, what, what is it that you're really excited about? Because there's all this stuff that you've already built. It's in the past. It's working. You're like, I mean, because even when I asked the question about like, oh, all these different networks, I thought you're going to say, oh, yeah, that's on the road map. Like eventually we'll get there. But like, you're like, we're breezing past that. So what's on the cutting edge for you guys? Yeah, I think the main thing is we've seen incredible sort of, you know, it's this two set of marketplace. We have to supply and demand. The supply side has been amazing. We have Amazon web services going live now with like literally X402 payments and discovery for all their enterprise builders of agents. Like the adoption has been incredible. And cloud flares announcing that you can start, you know, gating your websites behind X402 based transactions. So if you have an agent that comes to your site, you're not going to make any revenue off that today. But you can have an agent, you know, pay a micro transaction to do that instead.

35:11All that's on the supply side, though. And we haven't seen any real consumer demand side adoption. And part of what we're thinking about is as coin base, what can we do to help drive that? Where is our sort of core competency? And really the truth is that's that's trading. So I think through the work that we're doing to sort of bring agente creating to life and that flywheel talked about earlier around. Asia goes off and pays for access to premium data. That makes the trades more profitable. Therefore increases the demand for data. That's where I think we're going to see a lot of this early initial traction for for agente creating and payments both happening together in like general money management. So that's one. The other one that's like really far out that I'm super excited about is like the crypto nerd in me is really excited to see how far we can push the realms of like a fully decentralized autonomous agent. So like can you run this in such a way that it really feels like a true decentralized life form. You just give it a wallet. It pays for its own hosting it pays for its own infrastructure and services and all this different stuff such that there's really no way to shut it down.

36:13And it's it's a bit of a scary thought, but it's something that's inevitable and it's something that I'm quite excited to see start to come to life now that all the infrastructure and technology is in place. What do you say to the person that's listening out there, maybe they're not too first tech savvy. And it just seems a bit overwhelming or intimidating. What do you maybe say to them to get them a little bit more familiar or comfortable or make it seem like it's not this big malign of like I don't know how I'm going to do this like I maybe turn on my computer and I can get on to an exchange, but like that's a lot for me. Is there a path forward for those people. Absolutely. Yeah, this is why we're building Co-invest Advisor inside the app. You know, it's one tap away. If you're more comfortable interacting with chat GPT, it takes you literally a minute to get set up. If you're concerned that your agent is going to go rogue and blow up your balance, you can have a setting on where literally every single interaction. There's a nice widget that pops up and you have to manually approve all of those before your agent can do it. So I think it's very analogous to self driving cars.

37:14You know, a decade ago, if any of us thought we would be sitting in a car that can drive itself, you call yourself crazy. But it started just lane assist. It started with ads for parallel parking and how easy that is. And now we're seeing this massive progression in both the improvements in this technology and the user trust. And really that's where it starts. So I just encourage people to play around with it and poke around with it. I promise you it's easier than you think. Yeah, that's incredible. It's self driving cars. I mean, we're talking about the robots in the house. This is going to be a huge piece of the future. I mean, this rock bot thing is super cool. Like it's basically like you have all these different little AI agent. Like it's the same kind of concept. But I could just imagine like wanting to like, you know, text my, my roomba. Be like, hey, clean, clean the kitchen. And you're just in the rock bot interface. And you know, one of the characters is your roomba. And you're just texting it demands or your optimist robot. Hey, go clean the dishes. And I was telling somebody else about this. I was like using a claw to teach a skill.

38:19And so, you know, I have, you know, certain things that I do every single day, whether it's market reports and, you know, position reports and that kind of stuff where you could just be like, you know, watch me ride along with me. And there's a button now. I mean, Grock has it. Chat chat. You have it. Teach a skill. And so it screen records your actions. And it allows you to, you know, log in, click around, export things, upload it to a Google Drive or wherever, do analysis on it. And then you could say, run that every day forever. And it's like, it just does it. And it's so cool. And then I was like, oh my gosh, that's going to be what we do with robots. We're going to, it's going to watch us do our dishes or fold our laundry or put our clothes away. It's going to watch us do it one or two times. And then it's going to say, okay, thank you, Bryce. I understand, you know, the task. Now whatever you want me to do this task, tell me, do my laundry or whatever. And it's just going to do it. And it's just, you know, some will come preloaded with certain skills and it will just need your spatial mapping. But I mean, this is to me, just such an exciting, exciting time. And it's funny like how, you know, guys like me and Brendan, we started out as crypto traders and investors and stuff. And here we are having a conversation about the future of robots and AI. It's like, and that's, you know, it makes me feel just as excited as when I heard about crypto the first time. It makes me feel just as excited as like, you know, this concept of, you know, stateless money and, you know, non-sovereign.

39:44So non-sovereign control over, you know, execution layers and settlement like all that kind of stuff. I'm like, man, this is a big idea. AI is just as big if not, if not bigger. And so there's a lot of, a lot of really, you know, like you said, green fields in this right now. So super excited. Brendan, any, any last questions for Lincoln while we still have some time here? And I think we hit on most of them. I think if anything, I'm walking out of this with a much better understanding about how all of this is going to affect us moving forward. Obviously, crypto is a big part of it. It seems like crypto is the real big infrastructure on it. And, you know, from what I'm taking away here, it seems like Coinbase's play is that AI companies are going to be able to do it. So, like, AI companies are going to be spending in companies in general are going to be spending billions on all these competing models and everything in Coinbase is saying, hey, let's build the financial system. Let's focus on infrastructure.

40:45And that's what all of these kind of winning models can run on. So is that right? Lincoln? Yeah, yeah, exactly. Like we think AI fire AI finance covers, agenteic trading and agenteic payments. And this is in an evident part of the overall AI and agent build out that we're seeing. And Coinbase's goal here is to be the backbone of the agenteic economy, whether that's through providing advice through Coinbase advisor, letting people plug in their Coinbase account into their agent through Coinbase for agents or even just paying with X4.02 and Coinbase for business. So I do actually maybe have a related question to that. So if people want to maybe get exposure to this idea and this play in this stuff that we talked about, you know, let's say that a lot of the stuff that we talked about takes off, including X4.02 and, you know, the different products. Is that something that would go back maybe towards Coinbase? I think you said that X4.02 is now separate from Coinbase. Is that right? So, like, say this all succeeds and it's massive. And, you know, you guys are this form of infrastructure that everyone's using. Is that something that you would see on like in earnings call or like a quarterly report from Coinbase?

41:55Or is that now separate? Well, so X4.02 itself as a standard is technically separate. However, Coinbase has a lot of the pieces of infrastructure required to make these payments actually happen at scale. They're running on USDC on base. They're using Coinbase's wallets or the on platform retail USDC dollars that are flowing through all of this. So, Coinbase is the path here with X4.02 is to create an open standard. And through that, create as much value as possible. And then focus on capturing a little bit of that. Like, if you look at what Google did with Android, Google acquired the Android operating system and took the opposite approach to what Apple did with iOS and made it available for everything. From cars to kindles to Google funds. And that approach now resulted in Android being an 80% of operating systems worldwide. And that's what we want X4 to be. Wow, incredible. Love it, man, Lincoln. Can we have you back on again in a quarter or two? I'm sure things are going to wrap a clip over there at Coinbase.

42:56And I know that the platform is going to keep evolving. So we really appreciate your time. Again, we welcome you back. And I hope everybody at home who was watching or you're on your way to work or any way to school, whatever you're listening. Keep following along. Come back next week. We're going to have some more incredible guests for you. Same time, same place. Lincoln, enjoy the beginnings of hopefully a new bull market. We'll check in with you soon. And thanks again for your time. Yeah, absolutely. Thanks, guys. This is great.