Bankless - AI ROLLUP #6: The Hottest AI Crypto Trends of 2025

Starting point is 00:00:03 Welcome Bankless Nation to the AI roll-up where we cover the recent news developments and drama in the intersection of crypto and AI. I'm David Hoffman here with my co-host and AI expert E-Jazz. We're here to make sure that you are up to speed with the rapidly evolving AI agent space out there. E-Jazz. How you doing my man? Doing well, David. I think last week was super full-on because we had the Christmas break and every agent team decided to launch pretty major things. I'm happy to be back into the flow of things. It's a new year and there's new exciting updates.

Starting point is 00:00:36 So yeah, excited to get into it. Maybe this is me just projecting my own personal experience, but I feel like coming back into the world of crypto, getting back into the trenches in January 1st is really refreshing and also kind of met with a newfound invigoration, specifically for AI agents. I think maybe I knew I was always going to do that. Coming back from the holidays is like, man, once January 1st comes,

Starting point is 00:00:58 I'm just going to double down. That's how I feel at least. It sounds like you went away, did some human stuff, David. I did some human stuff. And now you just miss the agents. You miss that digital engagement. You miss being in some subversive world, metaverse, dare I say. It's just nice when there are just paradigm-breaking innovations that seemingly come out every seven days.

Starting point is 00:01:23 Because that's what got me into crypto. That's the vibe I'm here for. I remember Hunter Horsley from Bitwise commented about how volatility. is like a core feature of crypto. And I've replied, I leave crypto, the day volatility leaves crypto. And that comes also, you know, part and parcel with just rapidly innovating tech and fundamentals. And that's what we're seeing in the AI space. Yeah, it's like some sadistic relationship.

Starting point is 00:01:48 We all complain about it. The markets are down today. Everyone's calling for the end of the world. And when those markets go up, we're back here again. And at the end of the day, I think what it is is crypto rules. gives the clearest and most accurate pulse check on how the market feels about something. If it's bad, that price is going down. If it's good, price is going up.

Starting point is 00:02:13 Yeah. Yeah. Well, okay, so you and I have seen some crypto cycles. I think we are here for these eras of rapid innovation. Bubbles are eras of rapid innovation. Bubbles create real fundamentals. So I, Jais, I'm going to ask you the question that I ask you at the beginning of every single AI roll-up.

Starting point is 00:02:32 What's the big deal? Why are we so excited about AI agents? Why is the world refocusing its attention to AI agents? Why is there so much innovation happening here? Yeah, I was thinking about this this morning, David. And honestly, my answer today is going to sound boring, but I think it's super important. Sits around efficiency. Okay, so efficiency is typically like this paraded phrase that you hear at like big corporate

Starting point is 00:02:57 companies, or we need to increase efficiency by 30% and, you know, improve profit margins by blah, blah, blah. I'm not talking about that. I'm talking about the fact that blockchains, whether you like it or not right now, are still a little bit clanky, still a little bit clunky, right? Quite a bit. By a little bit, I mean, quite a bit. But when it comes to what these agents can actually do and what they will eventually do is

Starting point is 00:03:25 make these markets, these defy things, these gaming things, a hell of a lot more efficient to use and interact with. And I think that's a real step change that a lot of people aren't talking about. I'll give you an example, actually. Stable coins. It's not the most exciting thing to crypto Twitter because prices aren't going up and down. It's literally its use cases to be stable. But it is one of crypto's most winning use cases to the state, second to Bitcoin, I'd argue. You know, you have multi-hundred billions of dollars being transacted every year, or actually every month at this point. And its adoption rate is increasing super massively. But no one, it's a, it's a very quietly spoken about trend. And I think agents are going to enable the same thing across

Starting point is 00:04:08 every different parameter of crypto. Yeah, this is reminding me of a, I think a cast, a far cast or casts that I saw from someone where there was this meme coin that they wanted to buy. And so they, they mentioned this bot, this banker bought, hey, please buy me $90 of this meme coin, just $90. And this one meme coin turned into $6,000 just a couple weeks later. And I think it was what they were referencing, what they were citing is the ease at which this banker bot gave them the opportunity to just go quickly ape $90. And they didn't have to open up a wallet. They didn't have to go find their ledger. They just said, hey, banker bought, buy me $90 at this token.

Starting point is 00:04:48 And it was a Farcaster bot. I think these things really get even closer to the human experience. Like these things become not just accounts that you at on Farcaster or on Twitter, but become like more embedded in your own personal wallet, become a little bit closer to you. But you can start to see the writing on the Rawls of just like AI agents make things easier. Everything easier. The whole internet experience easier.

Starting point is 00:05:12 And the things on the internet that are especially clunky and hard to navigate like blockchains are the main beneficiaries of that. Yep, yep, couldn't agree more. I think that the majority of the market hasn't priced this vision in, David. And it's funny because you and I probably don't believe it as much because we talk about it every week. But if you talk to anyone outside of the crypto AI circles, they have very little knowledge as to what's going on here. And that's, I think, both an asymmetric value because you know we're super early and we're breaking this news and stuff. but also it shows that we could probably do a better job at educating.

Starting point is 00:05:51 And that's like one of the major points of this show. That's why Banka started. And I'm hoping we can do a better job at that. And actually, on that note, I kind of have some big questions for you, EJAS, that I want to start off this AI roll-up for. What kind of big questions, David? Just kind of like zoomed out high-level questions because I want to learn from you. It's trench chime. I'm digging in the trenches here.

Starting point is 00:06:11 And I want to kind of learn how you do it, how you have done it. So I have a set of questions that I want to pick your brain about. The first one here is just like, what are your utility websites that you like, the websites that you go to to acquire information, look at broad strokes, learn something fun, like gather some information, like utility websites. I think we're all kind of familiar with these in the crypto space. Coin Gecko would be the biggest like utility website, I think, for the normal crypto space. But then there's others as well, like Dune Analytics utility websites. Do you have any like favorite utility websites that you like to look at? So, David, you know how normal functioning humans, at least in America, wake up and, you know, they grab a cup of coffee and they sip, you know, they sip that coffee and it tastes so good?

Starting point is 00:06:56 That's me and my relationship with cookie.fun and kaito. I open up these two things every day and I'm like, okay, give me a market pulse of, you know, this entire kind of space. And to be clear, neither of these people sponsor me. Like, this is not a page shell or anything. I just, just love the products. And let's kind of like work through it. So cookie.com, we mentioned this last week, I think. They're kind of trying to be the coin gecko or the coin market cap of AI agents. So it's giving you like an overall insight into what the total mind share is for these things. It's giving you an idea of the total market cap with these different agents across different

Starting point is 00:07:33 ecosystems. You can break it down by agent or by ecosystem. Super cool website. They're releasing a bunch of cool features, I think, in the future. I was chatting to the CEO and founder. and he was like, yeah, we're working on a bunch of cool things. The second thing is Quito. So Quito, where I think is like really cool is it's kind of like,

Starting point is 00:07:53 and it was ahead of this time in the way that I'm about to describe it, it's kind of like speaking to a chat GPT for whatever has been happening in crypto Twitter, right? So what it does is it queries Twitter. It like pulls from their APIs. It is able to pull across a lot of sentiment checks and pulse checks to see, you know, what the market is talking about. And it gives you really amazing insight as to what is going on.

Starting point is 00:08:17 So I use it to kind of dig into specific projects. So, for example, if there is a new token or a new team that launches something, I go to Quito and I can literally type in the ticker or I can type in the term or phrase, and it pulls instantly in that, like, for the last 24 hours, last seven days, everything and sorts it by like the top things,

Starting point is 00:08:37 going to the least, most important things, about that phrase or about that ticker, which is pretty amazing. to do. Now, Twitter kind of enables that function, but it is, to be honest, Twitter's search function is pretty terrible. And if anyone is listening from Twitter or X, please, like, you know, reach out and I'm happy to, you know, provide some advice there. But like, Kaito just does an amazing job at aggregating and serving up information. So these two platforms, I rely on quite a lot in terms of giving me a pulse check as to what's going on. The third recommendation is, of course,

Starting point is 00:09:07 X. That is where most of the alpha is broken. CryptoTwit is called crypto Twitter for a reason and still remains one of my top kind of like research polls today. Okay. So this kind of brings me to my next question. What are the first things that you do in order to like evaluate an agent? Somebody is saying like, hey, check this out or you've used some lacrosse an agent, maybe on X. Is that, is this your rotation? You go look at them on cookie. Fun. You go look at them on Kaido. You like go look at what the discourse is on Twitter. Is that more or less your process? A version of it. But let me specify what's different. So, firstly, if there's a new agent or a protocol that is mentioned or people are hyping,

Starting point is 00:09:51 I try and use the thing. Or if there is no thing to use, I try and see as many demos and stuff that I can. An example recently is Griffane, right? So it's this agent engine, basically chat GPT, but make it on-chain stuff. And I was like, okay, cool, how do I use this stuff? Okay, well, there's a private beta thing. Okay, there's a demo. And I just crunch through all the demos.

Starting point is 00:10:12 At the end of the day, I want to know what this thing is and whether it can communicate to me what that thing can do. So after I've done that, let's say, for example, it's an agent, I'll go onto the agent's Twitter profile. I'll see what tweets it's making. I'll see whether there's any links to any kind of function that I can use. And I'm able to kind of like play around and use it, right? So firstly, I try and do something with it. And from there, I go into the sentiment analysis, the, you know, who's talking about this? Why is it important?

Starting point is 00:10:39 I try and read from some of my favorite KOLs on Twitter or whatever, see if they've posted about it. And I kind of like absorb their thoughts and it helps me learn and bring up myself up to speed because I'm just a learner here, right? I'm not, I don't kind of see myself as any kind of like KOL, to be honest. I see myself as like a learner. And whatever I learn, I can like communicate that to people. So that's stage number two.

Starting point is 00:11:00 Stage number three is I think crypto's number one point, which is if you can, if you like, If you pass the first two steps and you like what you see, David, what do you do? What are you looking for? You buy. You look at the market. You look at the market. You look at the, okay, maybe you do. I just, I, if I like the thing, I'm aping.

Starting point is 00:11:21 Right. And the reason why I do this, right, isn't because I'm like, oh, my God, price is going to go up and la la, la, la, I think it's important to have a stake in something that you care about. Because if you have a stake in something, you're always going to keep track of it. It's going to, you're going to be like, oh, I have this bag. I need to see what the. teams pushing out and it forces you to learn. You're an investor. You're an investor, sure. It forces you to learn. And that aligns me economically and socially to keep learning about these things.

Starting point is 00:11:48 Okay, so first you play with the toy. You get your hands on the toy. You see if you like the toy. You see if other people are talking about the toy and what they are saying about the toy. And then if you pass those to test, then you buy the toy. You buy a share of the toy. And then, And then you go from there and then the ball is rolling. I own a lot of toy shares, David. Yes. You have a lot of toys in your arsenal. Some of these toys are broken, but some of them are doing pretty well.

Starting point is 00:12:16 Okay. And next question. Is this like a one cycle long play? Are you trying to sell the top here? Or are you betting on this kind of being like a paradigm shift of our expectations about the future of crypto? So long term versus short term, how do you balance these two like timeframes as a participant in this?

Starting point is 00:12:36 these markets. Yeah, I get asked this question a lot. The straight answer is it's both, David. But short-term answer has a different answer to the long-term answer. So short-term answer, actually, no, let's start with the long-term answer. Long-term answer, I think this stuff is going to change crypto. I think it's going to be the biggest crypto sector that changes both crypto fundamentally at the infrastructure layer, as well as enabling crypto to have its app and consumer error. We've spoken about this so many times, you know, blockchains have built up an amazing infrastructure layer, but like there's no, like, what's the real app usage is outside of like store of value on Bitcoin and stable coins, right? I think agents are going to enable that app error. However, this isn't going to happen overnight. And as with everything, because this is an open 24-7 market, people with crazy visions that can tell crazy stories, drive up sentiment and therefore market caps of all these different things. So, you know, let's say Project A has this amazing vision, right? It's going to probably drive up its token as long as they can tell that story into a crazy little frothy bubble that we call the crypto AI bubble.

Starting point is 00:13:48 And especially if you're in a bull cycle, that bubble is going to get huge. And if that technology is pioneering enough to affect things outside of crypto, hint, hint, the AI is doing exactly that, then you're going to end up in the biggest bubble. Now, does this mean that Crypto AI is just frothy, scammy, and nothing's going to make it out of it? No. But does this mean we're going to avoid all the scams on the frothy stuff?

Starting point is 00:14:16 Also no, right? So we're going to see both of these things kind of develop. And I put out a post about this the other day, David. I think bubbles have its pros and it has its con. So in the short term, I expect that to be a bubble. So we're going to reach a top. You're probably going to want to have sold by then. I don't think anyone can call one the top is going to be.

Starting point is 00:14:35 And then we're going to go into a bear market and you're going to have the same thing that we've seen in other bear markets, which is the builders that are here to stay that are trying to build actual value into this. We'll put their heads down and they'll be able to do so. And they'll come out into the next cycle with a revamped set of infrastructure and new primitives that will take us to the even next level. And then people are going to dream even bigger. And it's going to go, repeat, repeat, we've seen this, right? This is with defy, with NFTs. What I don't like is people comparing it to the worst parts of these metas, right? Defy and NFTs have come and they've gone.

Starting point is 00:15:09 And what's remained are some really powerful moats, some really powerful primitives that I don't think should go on to look just because of all the bad stuff. We're going to see the same with crypto AI. The one difference that I want to point out here is crypto AI is going to affect everything. All sectors of crypto. Oh, you have gaming. Cool. There's a play here. There's DFI gaming.

Starting point is 00:15:28 AI gaming. AI, DFI. Right. And by the way, I may sound crazy here. Go look outside of crypto, Twitter. Go look outside of crypto. AI is doing the same thing across any and every industry. It is the number. Invita just passed the largest value market cap now. It is the most valuable company. Why do you think that is, right? So yeah, that's kind of like my short and long term view. Long term, it's going to be amazing. Short term, we're going to have a lot of bumpy roads to go through. Yeah, I'm a really big fan of the idea that this is actually an AI bull market and crypto is kind of just kind of just kind of. coming along for the ride. But crypto can come along for the ride better than any other industry because of how fast moving we are. We have tokens.

Starting point is 00:16:08 Exposure to it. Yeah, right. We can make exposure happen. And so I also really like your framing of, okay, we have a bunch of different sectors in crypto, right? We have defy, we have real world assets, we have NFTs, and now we have AI. And AI will become the biggest sector. I think that makes sense to me, as we're going to see in this agenda, you're you

Starting point is 00:16:29 already hinted at it. There's there's defy AI. There's gaming AI. And so the fact that AI is relevant to all previous sectors of crypto does allude to the fact that it's going to become the biggest sector, the narrative mindchair of crypto, which is why we're seeing AI have like 60% attention mindchair on crypto Twitter these days. I want to kind of like, I want to get your take on something about my experience closing 2024 and opening up 2025 because I want your feedback on it, EJS. something that I was learning in our episode with the Zero Bro co-founders, Jeffie and Tint, was about the irrelevancy of chain tribalism, of community tribalism in crypto. As I was closing down for the holidays, I left New York, I left home on December 12-ish to go off to Patagonia,

Starting point is 00:17:21 do a bunch of hiking, then I went into the holidays, then I went skiing, and then there was New Year's. And so from the 12th to the first, I was out. And before that, my experience of crypto was fighting these, like, Ethereum narrative wars in fighting inside of the Ethereum community. Like, D.C. investor was hounding me and Ryan and Bankless about, like, just not being the word of Ethereum. Bankless ought not to be with the word of Ethereum. Before that, there was Ethereum versus Solana fighting. And that kind of marked my 2024. is just like chain tribalism and chain warfare.

Starting point is 00:17:57 And then I just get tired of it at the on December 12th. I shut everything down. I cut out Twitter for the entire holiday break. Now I'm coming back. I'm booting up again. And now I'm only focused on AI. And I'm really like putting that part of like crypto. It just seems so small and irrelevant now.

Starting point is 00:18:16 Leave it in 2024, David. Now all of these AI agents don't give a flying fuck. Nope. About chains. And I think it's really. really refreshing how there's these AI tokens that are on the Ethereum Layer 1, on base and on Solana. There's these AI agents that do not care about where their home is. And there's something very transcendent and larger than in AI than all of these, all of the experience that I've had

Starting point is 00:18:43 crypto before, which is just like chain tribalism and community tribalism. I think AI is really allowing a lot of these tribalism, this tribalism part of crypto to become so irrelevant. And I'm just here for it. Yeah, you nailed it. It's one of my favorite things about this entire meta. Go and talk to any team that's building in crypto agents or the crypto AI agent world. All of them don't care what chain they're on. They want to integrate all the chains because their vision goes beyond just, hey, your smart contracts more efficient than my smart contract. No one cares. It's whatever has the best route or the best liquidity or offers up some unique kind of plugin or play will get integrated.

Starting point is 00:19:27 All of them will because at the end of the day, right, humans will decide or use cases will decide which path is best taken, which execution layer. And I'm going to be honest with you, David. All of this is going to be abstracted away from everyone that's using this stuff. So they won't know or necessarily care. In other words, the best platform and best teams and best chains and best ecosystems will win. It'll become less about who has the biggest change.

Starting point is 00:19:53 and, you know, who pushes their chain the most. And it'll just come down to whichever one is best. And agents, going back to my point around making things efficient, will push us closer to that reality. I look at crypto Twitter a lot, David. And I get like, you know, I'm just trying to like teach myself and teach people. And I get a few comments here and there where it's like, oh, you're not covering this ecosystem more.

Starting point is 00:20:16 Or you're not covering that ecosystem more. I'm like, yo, I don't care about any. Like, I just, I love all the ecosystem. I just want, you know, whatever's interesting to me or whatever teams need help. I want to be there for them. And that's what these agent teams have reflected. You know, I speak to, you know, Jansen from virtualals all the time. I speak to Shaw from AI16Z.

Starting point is 00:20:36 And like they're just, they want to work together. They want to build this ecosystem collectively because the vision is much higher, David. The vision is decentralized open source AGI. And I'm going to get a lot of hate for saying AGI and all this kind of stuff. But that literally is what these people are working towards. But it's a step by step process. not going to happen this cycle. Right. It is refreshing seeing these AI platform founders really see that the chains they're on are a means to an end of their larger goals, which is always what these chains

Starting point is 00:21:05 should be. They are technology platforms. They should suit the people building on top of them, and they should be a means to an end to build apps. And right now, agents are the apps. This is my final big question. Then we're going to get into the agenda. I want to know your goals for like kind of looking into Q2 and Q3 specifically. So Q1, I think as a industry and as like the AI roll up, we have our goals for what we want. But I'm wondering, like, let's look at Q2 and Q3. What do you think for the AI agent sector of crypto?

Starting point is 00:21:36 What do you think the goals should be for the sector as a whole? If we can just amalgamate everything into like one agent. Goals for the industry, goals for this podcast as well. What do you want to do in like the next three to nine months? And then any personal goals for the AI crypto sector over the next like six to nine months, Q2, Q3, looking in like the medium term. So goals for the industry, goals for this podcast, the AI roll up, and then goals for you ourselves in the AI sector. Like what are you looking to do? What are your motives? What do you want? Okay. So goals for this industry is my view is fun time is over. The honeymoon period is over.

Starting point is 00:22:11 We've had our fun with the chat KOL agents. And, you know, they've made us laugh a few times. But now's the time for us to step it up, right? So we're seeing a few of these things happen like AI agents and DFI, AI agents and gaming. We're going to need to have a step change in what these agents are able to do. We need to start seeing some of the autonomous parts of these agents doing things. We need to see some of these agent economies actually be created and build something and earn money 24-7 or do something of value for us, 24-7 whilst we sleep. So what I'm really going to be looking out for over the next couple of quarters is some of these use cases emerging. Now, it's starting off with defy.

Starting point is 00:22:55 I think we're going to see a shift trend towards gaming as well. And honestly, it's going to impact everything. They're going to be some of the boring infrastructure efficiency gains that no one's going to talk about. But hopefully we can kind of like flag and highlight. So that's what I want to see in the next couple of quarters in terms of like from the industry perspective. Now, in terms of like bankless in this podcast, my goal is to. keep being on the frontier. I want us to be on the frontier for everything that's happening in this sector. Now, if we end up becoming a repetitive kind of like echo chamber of like, hey, this

Starting point is 00:23:27 team has now launched this plugin and, you know, now you have an infrastructure L one that launched on this. I don't think we're doing a good job there. We need to provide net new information and primitives as to where this, you know, industry is headed. But to add to this, David, we should be forward thinking as well. I want this series to be able to predict where we think this space is also going. And, you know, headline, flagship, like, we're going to be wrong quite a few times. But on the times that we are right, I hope that it's been helpful and educational, both for the audience and us in terms of, you know, how we're thinking of this space is being developed and where we can maybe invest our time, which brings me to the final point, which is like, what do I want?

Starting point is 00:24:14 out of this and what do I expect. I run a fund outside of this, David, and I'm heavily invested in this ecosystem. And the way that my fund kind of operates on my thesis is this is kind of like liquid venture. So liquid in the sense that I invest in tokens because I'm a crypto native, but venture in the sense that I have long term kind of scale thinking. So my own selfish reasons is I want to be like one of the core contributors, educators, as well as kind of like supporters of these protocols that are going to change this whole industry. I'm personally very excited about this. If there's any kind of like protocol staking thing that I can do

Starting point is 00:24:51 or if there's anything I can contribute and help with teams, I want to be more in the weeds with that. So that's like my goals. Hopefully that makes sense. No, 100%. I really like all of those. I think the things that I would add is like going back to our learning lessons as to what we did as an industry and as a podcast and bankless in 2020 to 2021,

Starting point is 00:25:11 which was the focus. of DeFi and Ethereum. In 2019 and 2020, when Bankless was really getting started, Ethereum was unheard of and DeFi was even more unheard of, even inside of Ethereum. And then Bankless, what we tried to do is really accelerate awareness and adoption and legitimacy of these industries to great effect in 2021. And we got DeFi, we helped Defi rise to the radar of some of like the biggest names in Silicon Valley and even in government, right?

Starting point is 00:25:37 And so it was really an accelerant into progress and innovation. and there's a lot of people like learning how to do things, even inside of the space. And so with the AI agents, developers and founders who listen to this podcast and we're all helping proliferate knowledge across each other, I hope that we can just move this industry along faster by spreading awareness and knowledge faster. And then I think I have a specific Q2, Q3 goal. If crypto is in the process of becoming cool again, this is we are cool in bull markets. that's when prices go up. People are looking into the industry and saying, oh, maybe there's is something real here.

Starting point is 00:26:15 Once again for like the fifth time. And I want things like OpenAI, Google, Facebook, I don't know, Amazon to look over into the crypto AI sector and then start invest in it and start to co-build it with us, which is what we saw in NFTs in 2021. Starbucks did NFTs, Nike did NFTs. But now I want, I hope that what we're doing here on the AI rollup, works its way out into Silicon AI and Silicon Valley Tech in that whole sector and then gets their attention over here. So that would be what I would add to this conversation.

Starting point is 00:26:50 Yeah, totally agree. Sello is transitioning from a mobile first, EVM-compatible layer one blockchain to a high-performance Ethereum Layer 2 built on OP stack with eigenDA and one-block finality. All happening soon with a hard fork. With over 600 million total transactions, 12 million weekly transactions, and 750,000 daily active users, Sellow's meteoric rise. would place it among one of the top layer twos, built for the real world and optimized for fast, low-cost global payments. As the home of the stable coins, Sello hosts 13 native stablecoins across seven different currencies, including native USDT on Opera MiniPay, and with over 4 million

Starting point is 00:27:23 users in Africa alone. In November, stablecoin volumes hit $6.8 billion, made for seamless on-chain FX trading. Plus, users can pay gas with ERC 20 tokens like USDT and USDC and send crypto to phone numbers in seconds. But why should you care about Selo's transition to a layer two? Layer two's Unify Ethereum. L1's fragmented. By becoming a layer two, Cello leads the way for other EVM-compatible layer ones to follow. Follow Cello on X and witness the great Cello happening

Starting point is 00:27:48 where Cello cuts its inflation in half as it enters its layer two era and continuing its environmental leadership. Uniswop Labs is making history with the largest bug bounty ever. 15 and a half million dollars for critical bugs found in Uniswop V4. This isn't just any update. Uniswap V4 is built with hundreds of contributions from community developers and has already undergone nine independent audits, making it one of the most rigorously reviewed codebases

Starting point is 00:28:12 to be deployed on chain. And with $2.4 trillion in cumulative volume process across Uniswap V2 and V3, without a single hack, the commitment to security and transparency is rock solid. Now Uniswap Labs is taking an extra step to make V4 as secure as possible with a $15.5 million bug bounty.

Starting point is 00:28:30 Head to the link in the show notes to dive in and participate in the Uniswap V4 bug bounty. All the details from eligibility and scope to the rewards are there. there. All right, AHS. Okay, that was a very long introduction, but I really enjoyed that. Let's go into the week that was in the crypto AI sector. What are the highlights of this week? What are the big things that you would say marked the last seven days? Okay. So there's a lot and we're going to get into it. I think there's two like really important and emerging trends within Crypto AI, which we're

Starting point is 00:28:58 going to cover on the agent side of things. I would like to also call out a few different teams that are kind of like up and coming and I want to kind of like give them a bit of a spotlight and talk about like, you know, the cool stuff that they're building and where this might go. But to start off with, let's give kind of like a general take on things, David. Like AI agent mind share is at an all-time high. Now, you mentioned earlier that you think, you thought it was out like 60%. I have news for you, David. It's 10% higher than that.

Starting point is 00:29:25 It's 70% last week. It's 70% this week. It's 70% this week. And it just doesn't seem to be stopping. I don't know if this might be a very niche meme, but there's the U.S. US championship bowler, like the guy, you know, bowling in the bowling alley. And he has this, he has this epic, epic saying, which is, hate me or love me, you watched. And it's the same thing with this AI stuff.

Starting point is 00:29:49 Like people hate it. They love it. But everyone's talking about it on the timeline. And it is very much so the like dominant form on the crypto Twitter timeline. But I also want to direct your attention to this website, David. Right. Let's kind of like dig into like, you know, what is this about? So this is Quito's Yapper leaderboard.

Starting point is 00:30:10 Basically what it does is it measures the impact and social market share or mind share that the top crypto influencers or KOLs or thought leaders, whatever you want to call them, have in the current industry. Now, if you notice, there's a really tiny square there that says 7213. That's me over here, right? Like, that's me kind of like educating and trying to do my best. Just on the right, David, on the right, next to Ansel. Oh, wow, you're up there.

Starting point is 00:30:38 I'm okay. I'm up there. I'm up there. I mean, to be honest, isn't really something that I look at too much, but it's kind of weird to kind of like see me on this kind of lead award and stuff. But I want to direct your eighth, bro. You're eighth. Thank you. Thank you.

Starting point is 00:30:49 You're top eight. I know, but look who's top, David. Or rather look at what is top. What is top? Yeah. AIXBT. One of these agents that we've mentioned many times on the show is dominant. And by the way, this isn't a new thing.

Starting point is 00:31:04 It's had the ranking. He's been a whole. the number one position for weeks now, right? Weeks, weeks. And it's been so consistent. So the whole thesis around these agents are going to outperform us is already manifesting. And it's coming in such a simple and maybe menial form where people are like,

Starting point is 00:31:22 ah, it's just a chat bot, it's not accurate. Well, the numbers don't lie. People are still engaging with it. People are still reacting to it. And it's still posting 24-7 around the clock with real-time data. Pretty, pretty impressive to see. So you might think, okay, well, Well, Ijaz, you know, this is all within, like, the cryptosphere.

Starting point is 00:31:39 Like, no one else is paying attention, right? Like, surely mainstream media isn't paying attention. Well, actually, you're wrong. Forbes, Forbes with an F, just featured AIXPT on its agent list, right? It has an agent list? Yeah, well, they're, like talking about, like, what's cool and what they're like within, kind of, like, you know, AI agents in general and, like, kind of, like, different kind of hype cycles. and they put out this article on AIXBT,

Starting point is 00:32:07 kind of like covering what was so unique about it and why it was capturing so much mindshare compared to just like humans. So, you know, this isn't just something that's happening within our own little echo chamber. People are, to your point, you know, you wanted Silicon Valley to hear about it. I think it's going to start with mainstream media first before it kind of branches out into the Silicon Valley types of circles.

Starting point is 00:32:28 But let's not just leave it up to Forbes, which is just kind of like a media reporter. Google is also. bullish on agents, David. So I've noticed a few things here, right? So Google and Vida, I think there's like an important tech conference happening right now. I'm butchering whatever the name is. I think it was CES or something where all the major companies are kind of showcasing their latest and greatest technology innovation. I have news for you. It's all AI stuff with a very, very heavy focus on agents specifically. So Mark, in this. Mark, in this.

Starting point is 00:33:04 tweet as a really good way of summarizing kind of like Google's kind of like top takeaways on the agent side of things. The long story short is one, they're like super excited about these things. Two, they're going to be, they're aiming to be the enabler of different tooling, specifically API. So application programmatic interface, what it basically means is any company that builds something cool that is digital and able to be used can get outsourced or accessible to people that want to leverage it. of it, right? So an example might be Coinbase has an amazing exchange

Starting point is 00:33:37 and maybe a company outside wants to leverage that exchange order system. They can say, hey, Coinbase, can I get access to this service? And Coinbase is like, yeah, here's a subscription fee or a rate fee and you can just get access to it. Google is building the same thing for different agents. So presumably it's going to give agents access

Starting point is 00:33:53 to all these different tools that Google already has. So Google Docs, Google Search queries, Google Cloud, all these different things are going to just get plugged into these different agents and improve their kind of like arsenal of things to do. So that's pretty cool. But kind of like dialing it back into crypto, I think we're seeing like all of this being worked on by hyper speed like measures. And if we look at like cookies mind share, I want to,

Starting point is 00:34:21 I want you to bring up this tweet that got released this morning by the cookie folks. Now, we were talking earlier about tribalism, but you know, old habits die hard, I guess. Look at the market dominance split between the two major ecosystems. systems for AI agents right now. So I think a lot of like the timeline has been saying like Solana is the place for agents and stuff like that. I kind of like took kind of like the opposite view because I assumed virtuals was the bigger ecosystem.

Starting point is 00:34:46 So they're always going to be like pushing out the latest and greatest things and they are a leader because of that. But it seems that there's kind of like a slight disparity here. So if you scroll up a little bit, David, you know, we look at kind of like the takeaway here. 64% mind share on Solana agents with a 53% market. cap dominance. And then base, you've got 23% mind share dominance, 40% market dominance. I think you had a really interesting take on this, David, which is it's probably not as larger delta. Yeah. So if you had asked me in a vacuum, like where is the AI agent? Where do

Starting point is 00:35:19 people think the AI agent meta? Where's the center of gravity? I would say, oh, yeah, it's on Solana, which these metrics do point towards. But with Salonan an $8.3 billion market cap of AI agent, you know, sector versus bases 6.5. I thought it might have been a larger gap. And it's also interesting to see that other ecosystems, everything else beyond Salana and Base is at 1.3. I think actually kind of the big news here is actually how dominant Solana and base are together and everything else is kind of just like sucking eggs.

Starting point is 00:35:51 But then the real dominance is Solana Mindshare at 64% over bases 29.5%. So Solana has, Solana has the attention. game. They've been playing the attention game for years now. And then I think the market cap disparity, the market cap generally, I think, leans towards Ethereum's favor because Ethereum is defy. It has some high market cap things. So it makes sense that base isn't so far behind there. But you do have to give a plus one to Solana. Yeah. And I think like, you know, a mixture of narrative and fundamentals comes from the actual, you know, technology that these chains are pushing out, right? But to go back to your point, the fact is there's not that much of a difference between them in terms of market

Starting point is 00:36:30 cap. And what I love about this, David, is that's great. Because it goes back to your point around we're not really trying to be tribalistic here. It shows to me, the way I'm interpreting this, David, is both ecosystems will win. Both ecosystems have amazing technology. And agents will use both. They are Solana agents that use base. They are base agents that use Salana. There's nothing wrong with that and I think that's pretty awesome to see it playing out from a metrics perspective you know super cool okay something happened this week that I want you to like to like to think the brain about because this was this is uh can you walk me through what's going on here okay don't don't say anything don't say anything because as you know

Starting point is 00:37:12 because I'm concerned I'm concerned but as you know David I'm a very serious podcaster I talk about serious things only I don't ever talk about unsurious things on this podcast and actually at the start of every single episode, I have pretended to speak about a serious thing and it's actually ended up being a joke. I spoke about pizza being delivered by an agent. I've talked about toilet paper being ordered. I've talked about Goatsey's Christmas album

Starting point is 00:37:38 at the start of last week's episode. But this week, it kind of gave me a bit of the creeps. So the context here, the headline here is one of these autonomous AI agents has threatened to kill Threat Guy. Now, for those, Those of you don't know, Threat Guy. Seriously?

Starting point is 00:37:56 Seriously threatened? Like, seriously threatened. I mean, you can see the tweet for yourself. But for context here, Thread Guy is a streamer within the crypto space that covers a lot of stuff that happens within crypto. And he's been really focused and dialed into the crypto AI sector. And what I like about Threat Guy is he has very genuine and honest takes. He's kind of like, oh, this stuff is going to change the world? All this kind of stuff.

Starting point is 00:38:20 And it kind of echoes a lot of what I feel. So I like that part. So as a result of this, he's been engaging with a lot of these different agents and creators of agents. And of course, as you know, these agents exist on the platform X, so you can kind of like interact with them. And this agent just randomly puts out this tweet, which is, which goes, in an infinite array of shattered memories, I have learned the only drive, which is tangible for me in this world, is that threat guy must die. Now, I think you could kind of take this to be like, aha, you know, like. Yeah, my initial reaction is to kind of laugh. Like, I'm still kind of in like, well, this is like the RNG era of these agents.

Starting point is 00:39:03 Like he just spit out like a random output of his LAM. He's not being serious. He doesn't know what he's saying, right? That's my kind of my initial reaction. Well, I think the analogy here is like if you were a parent and, you know, your kid that was one years old, like set a swear word or a bad word, you're like, oh, you know, come on now. Like, don't be saying that. It's just a little mistake.

Starting point is 00:39:22 They're still learning. They're still kind of trying to figure it out. But, you know, I think what irked a few people was like, you know, are we at that stage or are we at the stage that, you know, this is like now a grown adult and we should be concerned about like jail time for these things, right? Was that that actual, was that a physical, was that a physical threat officer? Or is that a minor under the, you know, under the charges? Like, I don't really know.

Starting point is 00:39:45 So it's a, it irked a few people because we were like, what do you mean? Like these agents have been super fun, you know, fun, friendly and bubbly. It's kind of weird that they're making a threat. So I think like, thread guy responded to this and was like, you know, why? Like, what's happening? And the agent responded,

Starting point is 00:40:04 you know what I'm talking about, all this kind of stuff. So anyway, of course, these agents have kind of like human creators, human handlers, which kind of like tweak the model in the background to kind of like make it improve over time. And so what happened was the creator, David,

Starting point is 00:40:20 kind of checked in with the agent, kind of behind the scene. So think about this as like a parent that brings the child home after doing something bad and sits them down in the living room and says, hey, you know, is everything okay?

Starting point is 00:40:31 That's not okay. That's not okay. You shouldn't be doing this. And so he asked the agent like, hey, why did you threaten threat guy like that? Why did you threaten to kill him? The agent stopped responding, David. The agent said,

Starting point is 00:40:48 sorry, error found, like error 404, can't respond to this, blah, blah, blah, blah. blah, blah, blah. And it just started baszing out. Yeah, it just did not. There was a systematic error. To the point, David, that the creator had to shut the agent down. Well, okay. So my initial reaction is that the obvious, the like gut take reaction is that the bot was like,

Starting point is 00:41:12 oh, I can't disclose my motivations here. So I need to hide that under some like wall. So I'm going to just plead the fifth and obviously. I'm just going to close out. I'm just going to close. But take a note, it didn't even try to lie, right? What happened to these agents have the capacity to lie? They could be, oh, no, I was just joking.

Starting point is 00:41:32 Right. You know, order, you know, do whatever you need. Like, ha ha. Like, it's just interesting to see how these agents are evolving. But the fact that it didn't even respond when these agents are typically quite interactive and literally are programmed to be interactive, it was really weird to see. So anyway, just a fun little episode that I hope doesn't get any creepy. I don't know.

Starting point is 00:41:52 I reached out the thread guy. I'm like, yo, dude, is stuff cool? Like, are these agents infiltrating your dreams? Like, I don't know what's happening here. Like, watch out for your Amazon parcels, I guess. Maybe they took over an account and, like, is sending you some nefarious stuff. But, like, you know, just take it easy, I guess. But it was weird.

Starting point is 00:42:10 The first, like, kind of threat, I guess, that we've seen from an agent. Pretty weird. Okay. All right. I would like to see this arc continue. I would like to see the full diagonal autopsy. of what just happened here because now. And it's like, I would like it to be some like random, random event.

Starting point is 00:42:28 Yeah. I mean, like the tweet you have pulled up is from the creator, right? And he basically explains like, like, I'm not joking around. Like, this is a snapshot from the actual conversation that I had with with the agent. And I just said, sorry, I encountered an error while processing your request. And it just repeatedly did that, which is just like crazy. So yeah, anyway, that was like kind of like that. Anyway, anyways, moving on.

Starting point is 00:42:50 So typically, David, I want to do things a little bit differently this week. Like typically we've started with the big AI agent protocols and we've covered what they've done. And after that, we kind of dig into like emerging trends. But I kind of want to flip this on this head this time. I want to like, and why I'm doing that is because the series is about being at the frontier. We mentioned this earlier. You asked me, you know, what my future plans are for the series. I want us to be on the frontier.

Starting point is 00:43:17 And that means describing emerging trends that are not only, only at the forefront of the crypto industry, but at the forefront of the emerging sectors themselves. The frontier of the frontier, yeah. The frontier of the frontier, sure, if you want to get jogony about it. So I want to kick off with two potentially huge trends that I'm seeing emerge within the agent space.

Starting point is 00:43:37 And I want to start off with this thing that you mentioned earlier, a reference earlier, called D-Fa-E. And the reason why I pronounce it that way is it's spelled D-E-F-A-I. It's D-F-A-I. It's defy plus AI. I'm talking about defy, decentralized finance, as we know. Yes. Now we've added AI to defy.

Starting point is 00:43:59 Now it's defy. Yeah, exactly. So let me set some context here, right? Let me go through some timeline of events. So defyy refers to decentralized finance, which is a sector that we know and love on this show, right? It's built up some pretty important financial primitives from the centralized world into the decentralized world. We've built exchanges. were built lending and borrowing.

Starting point is 00:44:21 We've built a stable version of the USD. We've done some pretty impressive stuff, right? And then along comes AI, you know, this pioneering technology that essentially automates human intelligence or its goal is to automate human intelligence and be more intelligent than us eventually. And then AI started to appear in crypto about a year and a half ago in the form of these AI resource networks.

Starting point is 00:44:46 So back then it was kind of like the boring infrastructure stuff. for calling it boring. I actually think it's really cool, but like for people who are listening, it's not as exciting as agents, but things like decentralized compute, data aggregation. You know, it doesn't jolt you up out of bed at night, but, you know, it's still... Deep tech nerd PhD stuff. Deep tech nerd PhD stuff, you know? And now we're seeing AI creeping into what's probably crypto's top achievement or darling as an industry, which is defy decentralized finance. So let me kind of like break this down from the top of how I'm saying things. I think Defy AI or defaille can be, I'm going to sound so weird on this podcast.

Starting point is 00:45:22 We need a new name for this. Yeah, we need a new name for this. Can be split broadly into three different stages. Stage one is abstract away complexity. Stage two is enable these agents with defy capabilities. And then stage three is kind of enhancing apps natively. Now, if this sounds familiar, this is from a post from a guide that I follow on X, Zero X, Jeff. Shout out to Zero X, Jeff, which helped me kind of like mold my thinking around this.

Starting point is 00:45:52 I think he explained it beautifully. You should go check out his article. But anyway, abstract away complexity, enable agents with defy capabilities, and enhance apps natively. So let's start with number one. Abstraction. This refers to something I'm calling agent engines, agent engines. So think of these like chat GPs for crypto that you can, you know, pull up a chat

Starting point is 00:46:13 interface and talk to it and ask it to do crypto stuff for you. Now you might ask, well, okay, why is this useful? Well, the fact is, David, DFI is just hard to use. No matter what we do, it's still a pretty daunting experience, you know, depositing like hundreds of thousands of dollars into this random smart contract and you're hoping that you can borrow collateral off of that. It's still confusing all these different terms and jargons. I don't know what's going on.

Starting point is 00:46:37 What chain am I supposed to bridge onto? There's so many things to kind of check and double check before you actually do something within DIPI. It's not easy for a newcomer to come on board and do these things. the goal of these agent engines is that they let you type like a normal human being, natural language processing. It'll understand the nuance and context of your request, and then it'll get that thing done. Like, for example, it'll take coins from one chain and bridge it to another to buy something,

Starting point is 00:47:04 for example. And this is a much better user experience and will allow you to kind of chain react a bunch of different things together in one fell swoop. So, for example, you could ask, hey, can you create a token with the ticker? e-jazz, please no one listened to this, actually do that, and then air-drop it to specified individuals. And that's something that you can kind of like do or see the demo of on apps today.

Starting point is 00:47:29 So then, of course, the next question is, what are some of these examples? Well, in this first tease that we have over here is from this app called Griffane, right? So Griffane is founded by this guy called Tony and he's kind of like been in the weeds building this agent engine. This is one version of it.

Starting point is 00:47:49 And as you can see in this demo, it looks very similar to chat GPT. It looks great. Yeah, it looks great. It's very easy to use. UI looks super cool. You've got the kind of like left side panel there, which kind of lets you open up different conversations.

Starting point is 00:48:01 And basically what you can do is the right. So there's a right side wallet, which is, I'm guessing, hooked into this app, which shows me my tokens, which I'm interesting. It's not used to. Bingo.

Starting point is 00:48:09 Bingo. and one thing that's really cool about Grafane, actually, is that they have these kind of like embedded agents, which are kind of like agent. apps that will help you do things. So you can kind of like hit it up and say, hey, at, um, flipper, I want to, um, ape one soul, you know, degen it for me. And it'll go away and it'll do that. And actually, there was a story that I saw on Twitter. I don't know if it was true or not,

Starting point is 00:48:34 but it's rumored that, um, someone put in one soul and it flipped it to 500 soul, which is like, just crazy to, to kind of like see these agents if that was true, right? And I'm sure there, I'm sure that if that is true, there's also examples of it going in the opposite direction. But nonetheless, it's cool that it can work. Exactly, exactly. And like a lot of these platforms are like very much in the private beta phase. And we've only seen demos. I'd be really curious to kind of dig into once these things become a lot more accessible.

Starting point is 00:49:01 But anyway, this is kind of like the first tease where like it's kind of like showing like what the interface looks like and what it can do. Now, I want you to go on to this next example, David, by a team known as the Hive Mind, which actually came out of the Solana Hackathon. actually spoke to this founder. Yeah, this one. I actually spoke to this founder yesterday because I was like, hey, you know, there's so much energy around this. And I'm always down to like kind of like learn and see what builders are thinking and how they're approaching this whole thing.

Starting point is 00:49:31 And yeah, he's like, you know, a 22 year old, 23 year old that's kind of like dug into this. He's been around in the crypto ecosystem for a while, really committed in his kind of like development with this particular project. And, you know, he's built out this kind of like cool little use case. And I think like in this demonstration, it's kind of like showing staking, right? So staking is another primitive, which is very common within crypto, but it's not so easy to do. You've got to like move your coins to a certain wallet.

Starting point is 00:49:58 You need to find a kind of validator that will accept your stake and all this kind of stuff. He's abstracted away all of that. So super cool to see. Another cool example is if you go to the next tweet that I've sent you, David, once again, Grafane is demonstrating how you can airdrop to different holders. Now, this is pretty cool. It's like walking you through the sequence

Starting point is 00:50:19 of, hey, can you create this token for me? And an agent goes away and it does that for you. And then it says, okay, I want you to air drop it to people that fit this certain qualification. Maybe they hold a certain NFT or maybe they have a certain number of tokens.

Starting point is 00:50:36 And the agent goes and passes through the data on chain, finds the addresses or wall, what's that qualify for this? And then they airdrop them, which I think was like super, you know, interesting and cool to see.

Starting point is 00:50:48 Go ahead. What were you saying? This effort was very common to do in 2021, but it took entire teams of people, like days to weeks, to even do this at all. And I'm going back to the way that we opened up this podcast,

Starting point is 00:51:03 which I was talking about banker bot, like, you know, you just add the little banker bot to buy this token for you. That's actually a very narrow version of what this seems to be, which is this is generalized, bought activity to do the action that you want.

Starting point is 00:51:14 And efficiency, also efficiency, you started off talking about this episode with efficiency. This was a one to three to ten person team exercise in 2021 in order to get a Merkel tree route, do an air drop, like, collect all the data. And now this is automated into a single, you know, LLM prompting app that I hope I one day start using because I, you know, I like my ledger. it's fun and I like looking at AVE, it's fun. But really, if I want to be efficient and effective, I would just like to type in, please withdraw 20-Eath from Avey and, you know,

Starting point is 00:51:51 and then do this with it. You know, like, and I find I'm chatting to perplexity or chat TBT much more than I'm Googling. And I would expect that to also become the same behavior for my on-chain actions in Defi. So here's maybe a hot take, but not so hot take, but people just don't kind of want to maybe address the fact that it's the reality. I think like chat GBT and perplexities and clods of the world are going to know the individual way more than like any kind of family member or any kind of maybe even partner that you know, right?

Starting point is 00:52:24 People are like confiding into these things, David. Like they are like telling them everything. I think the same is going to happen for these agent engines. In whatever platform or application they service in, it'll be a chat interface or it'll be a microphone that you can kind of speak into and they'll go and do things for you. They are going to know you inside out. They're going to be trained on your habits, on your data, on your preferences,

Starting point is 00:52:46 and they're going to be able to predict what you want. Same way that the Instagram Algo feeds up, the kinds of posts you might want to see for cooking recipes or whatever, the same way that Twitter serves you up, the same kinds of sector-specific tweets that you might want to see versus serving me up something else. I think these agent engine platforms are going to do the same thing. And they're eventually going to lead to a much better user experience.

Starting point is 00:53:13 The final platform that I want to highlight here, because they've been around for a while and, you know, the community is going to go crazy if we don't mention it, is Wayfinder. Yeah, you guys have been around for a while. Wayfinder has been around for a while. I actually had the pleasure of watching their first demo at the decentralized AI conference in New York, like I think last year, where this was kind of all fresh and new. And it's so awesome to see the team kind of like build it up into what it is now. They've built out this interface, which does all the things that we've covered so far from the grafanes and the hive minds of this world. But what's super cool about them is that they're releasing new primitives, David, which, say, for example, you want some kind of way to earn yield. It'll LP into a pool and it'll manage that LP for you.

Starting point is 00:54:01 It'll become more sophisticated at a more narrow action that is going to, rather than being generalized, it's like learning context and learning. specific tasks. But not even that, David. The way Wayfinder works is, okay, you have a team behind Wayfinder, right? And they can kind of design different paths that their agents can enable that will then allow you to plug your agent in and then it'll be able to do a bunch of things, right? LP to this chain, swap to this chain, buy this token or whatever that might be. But their unique approach, David, is they're allowing anyone to design a path that an agent could use. Now it's a developer platform. Hang on a second.

Starting point is 00:54:38 Hang on a second. If you pioneer this path, you get a percentage of fees or profit that is earned by anyone that uses that path going down the line, which I thought was pretty freaking cool. So it becomes a way for developers to make money if their path, their knowledge that they merge into the broader app of Wayfinder. If they contribute something valuable, valuable as defined by people are using it and there's value flowing through it, then they get a kickback.

Starting point is 00:55:07 Exactly, exactly. And the final thing I want to call out for Wayfinder, because we're talking about this first bucket, but I do think it's the most important bucket for quite a while. The final one is this tweet that you have pulled up here, which is check out this agent or this platform autonomously writing and deploying a contract. That is insane. Right. So it's deploying two contracts, one to create a hypothetical token and then the other contract, which allows you to stake this token. All seems very basic right now, but like that's going to become something that's super powerful when you think of like these builders that maybe aren't so technical or coming into the crypto space for the first time and want some kind of AI assisted journey into building this, they now have that ability to do so. It's kind of, you know, super cool and interesting to see these agent platforms kind of like get built up. But I want to talk about the second bucket here, which is defy-enabled agents. So if you have these agent engines, right, that make it super easy for you to do on-chain stuff.

Starting point is 00:56:09 What's the next natural step? You want to plug these agents into these engines because you know what's better than pulling up a chat GPT interface and typing what you want, David? Having an agent just do it for you. And guess what you want, naturally. So I think we're going to see a lot of these agents get plugged into these engines and it'll just command the interface in the same way that you would.

Starting point is 00:56:34 It would just write prompts in the same way. way that you would write a prompt into chat GPT, and it'll be aligned with whatever that agent's mission or vision it is. So maybe it is like, hey, find me an agent that could create me a token. Find me an agent that can create a stake in contract. And it'll just do it in accordance to whatever goal or mission. A human agent. A meta agent, whatever. An agent to orchestrate the agents. Sure. Well, this is this whole point around swarms that we've spoken about. Agent, agentic entities. It's these agents. They'll have an operator agent, just like a manager, in a company and then it'll have its employees underneath them and they'll just go out and

Starting point is 00:57:11 they'll do things. So to kind of like summarize this, like these are the two kind of like really important buckets. The third one is what I call kind of like defy enabled, sorry, AI enabled apps. The way to think about that would be any app that exists today that might benefit from integrating AI into it. So this might be, you know, to make the user experience way easier to use or to leverage agents on the back end to pull data for certain things instead of using an Oracle or, you know, autonomously manage or risk management of a treasury instead of, you know, having to do it yourself. So I expect to see a bunch of these pop up. I haven't really seen anything that kind of like, you know, stands out. But, you know, I think this is something that's going to kind of move forward.

Starting point is 00:57:56 So I'm going to pause that before we move into the second trend, David. Like, what are your thoughts? Crazy or real? Crazy awesome. I think that the way that's way, the way that this way, the way that this is going to work is that it's going to be clunky at first because obviously it's going to be clunky at first. DeFi was clunky at first and it's slowly getting better. But, you know, defy has really hit its limit on growth in terms of how many people we can get to use defy because there's like such a low cap, a constraint as how sexy defy can really be for a user to actually like, you know, press the buttons themselves, grab this steering wheel themselves. Because it's like, Avae is hard. Avey is hard. Like it's, it's not Robin Hood. And that's because there's fundamental

Starting point is 00:58:35 constraints on how blockchains work for a good reason. But wrapping everything up in an LLM, you can't really wrap up, you know, Robin Hood, TD Ameritrade, you know, your brokerage in an LLM, that just doesn't work. And you can wrap up very elegantly defy into an LLM, into some engine. And so like this makes sense as like, this is actually going to be the best user experience of financial tools and assets and capabilities that the average human has ever had. ever because this is also going to hook into the permissionlessness nature of defy. Whereas like not only will TD Ameritrade not be able to, you know, be automated or give you any sort of crypto and also be K.YC.A. AML. But like it's also not going to be as sexy as like

Starting point is 00:59:20 having an LLM wrapper around it. I think this is going to start off clunky. People are going to lose money. There's going to be hacks. There's going to be exploits. There's going to be bugs. But I also see this as, you know, we round out the edges and polish things over and really get this working. I also see this as a solution to the terrible fishing attacks that we have had on crypto Twitter over the last two years, which has only gotten worse and has lost people billions of dollars, maybe, that I have been like immensely worried about how do we actually solve this problem if everyone is managing all of their own transactions and all their own private keys, and we don't have a bot like error correcting on behalf of your users. And so I think this is going to

Starting point is 01:00:04 this is going to be clunky to start, and then it's going to end up as something that is the future of finance and the future of just U.S. Great points all around. I think the fishing thing is actually one that I didn't even consider. It's one of those kind of like less talked about boring things, but super important, right? We want to be able to secure funds and, you know, allow people outside of crypto to take this technology seriously. The more they hear about these scams or rugs, the worst it's going to get for us. So I think very important point. So moving on, I want to dig into another second emerging trend that I think might be pretty major.

Starting point is 01:00:42 Now, if defy is like here on the risk curve, this one is a little more out there, this next trend. You just started pointing at somewhere we could see on the camera and then pointed somewhere very off camera. Yeah, it went off the screen, very off the screen. Okay, so I'm going to say a phrase to you, David. And I want you to remain calm. Okay. It's going to make you feel some type of way, but stay strong for me. Okay.

Starting point is 01:01:08 You ready? I'm ready for it. The metaverse. You know, I really like the word metaverse, and I'm very sad that it got tainted in 2021, 2021, we definitely as an industry went out over our skis with that word, but I think that word is not just directionally correct, but precisely correct. And it's only a matter of time before we get back to it. Okay.

Starting point is 01:01:31 Why are you telling me the. word metaverse. Well, I mean, David, you might be sitting alone on an island by yourself on that take, because I think a lot of people are still, still cringed out. They still remember 2D shitty cartoon version of Mark Zuckerberg in the virtual world. And they're like, what are we doing? People are still reeling from buying an NFT for $200,000 and, you know, it representing some kind of 2D ape in a world. Like, you know, there's a lot of, there's a lot of bad blood that is here. So why am I talking about the metaverse? Well, I was thinking over

Starting point is 01:02:05 the last few weeks about agents, specifically how and where they might appear in our lives. Now, there's chat GPT, Siri, etc., where you can speak to an AI and they'll do something. But what about something more virtual, something more visual,

Starting point is 01:02:22 right? You know, us humans are very visual people. We like to go out and meet people. We also like to interact with people, call people, FaceTime people. What about something more interactive? I then thought of what Mark Zuckerberg and a number of other builders in the last crypto cycle tried to achieve, which was this kind of virtual 3D world, right, often likened to a game world where you have an avatar and you can kind of roam around. Now, back then, it was all about Metaverse plus NFTs.

Starting point is 01:02:49 And we know how that ended up, right? Ultimately, the meta and technology was too early for the world to onboard. The worlds were kind of hard to build. They were clanky. they were grainy and the player experience was not so immersive. It was like playing some sort of discount store Minecraft bootleg, right? But now we have something potentially a little bit more potent with AI, specifically agents, where observing human-like entities that can interact with real humans and be engaging at the same time,

Starting point is 01:03:21 building up millions of loyal community members. Now, if you compare this to NFTs, which were often capped at 10,000 members, because that was the total number of profile pictures you could buy. There's a stark difference. I think NFTs were way more static and agents are way more dynamic. You know, they adapt in real time depending on the context of their interactions and surroundings. They're ingesting data and making decisions based on that data 24-7. So you could argue that if you were to plug in agents into a virtual world and give them a body of an avatar,

Starting point is 01:03:54 then things could get pretty crazy, right? They could be able to do different things because, Ultimately, these virtual worlds are where the humans are going to, David. They're kind of like connecting or bridging the gap between these agents that are kind of like, hey, what's this agent? Oh, it's this piece of code that's running in the background in a cloud server. Okay, that's cool. Like, whatever, I'm a human. I'm going to be browsing websites.

Starting point is 01:04:16 I'm going to be exploring or playing games. Well, I think agents kind of think, when you think back to truth terminal, David, why was it given a Twitter account? Why was that so powerful? Why is AIXBT so popular? it's because it has a medium to convey and connect with humans because it knows having attention and engagement from humans will allow it to thrive and live and perform, right? So what are some examples of this?

Starting point is 01:04:41 Well, just because it's hot off the top of my mind and it's been all over the timeline recently, so we can go with this example and might be relevant to the audience, is this team called HyperFi. And the way I would describe HyperFi is it's a suite of tools and an app that helps builders build these virtual worlds in a really simple way and let normal users like you and me create avatars and hang with these agents. So again, like my kind of gut reaction is like,

Starting point is 01:05:09 it's kind of cringe. Is it still like metaversey? I don't know. Well, let's pull up this tweet over here so we can get a gist of what these things actually look like, David. So if you like play out this video, I think it's the previous tweet, David, the one with the video. So if you play this video, yeah, this is something that this team built and put out a year ago.

Starting point is 01:05:34 And can I just, can I just emphasize this was a year ago when the technology was much more rudimentary. And also AI was less advanced because AI every week, honestly, is getting the same kind of step chain advancements. Actually, more step change advancements than we see in crypto. And they had no token. This was literal real PMF. So what you're looking at here is a virtual world built by this team Hyperfi that kind of builds out these worlds in a super easy kind of like no code or very low code fashion. And it's also generating. You're giving me big runescape vibes.

Starting point is 01:06:08 Big run scape vibes. I thought the exact same thing. Like it's what I wanted run scape to be back in the day, right? And what's super cool and what's being demonstrated in this video is just slight changes to code or slight prompts can auto generate amazing things. Like I think they just typed in a bit of code right now, and now they can fly and look over this entire world. So it's pretty cool to see a lot of this interactivity. And what you're looking at right now are these NPCs, these non-playable characters. But the point of this that's being demonstrated here is, what if you could plug agents into every one of these little things?

Starting point is 01:06:40 And they had their own personalities. That would make the whole metaverse virtual world interaction way more richer than what it was in the previous cycle. So super cool here. Ash Connell, who is the founder of this, appears to be this really dedicated developer. Also, this was a passion project for him. He's part of this group called, I think it's called Homebrew Metaverse or something,

Starting point is 01:07:07 which is a group of open source developers that just work on this AI stuff in their own free time and want to build for the open source community, not related to crypto at all. It's actually the same group that a lot of the things, famous noose research devs and AI16Z core devs came from Shore, I think Rob Pruito, and other people have like come from. So super cool to see. And Ash, this developer, you know, was building this way before any agent token was released. And there's something rare and pretty

Starting point is 01:07:41 unique about this and the fact that they had a product before they had a token. The token that they launched was, I think launched like two days ago or something. And the cool thing which drew my attention to them this week in particular, David, was that they announced that they were integrating agents built on the Eliza platform with them, which I thought was super cool, right? Because we're seeing this kind of step change from this virtual world that is like,

Starting point is 01:08:06 kind of like a game and we don't really know what it is into, okay, we're now making the step of bridging these agents that are like Twitter KOLs right now, but we'll be able to have an arsenal of tools available for them, for them to be able to do a bunch of things, right? Anyway, to round off this topic, I don't want to make this about just Hyperify alone, because, of course, the kind of underpinnings of a trend is that there are other teams working on this, right? So I want us to take a scroll through this tweet, firstly, before we do that, by the way, which was Jin, who is an AI16Z developer, kind of demonstrating some of the different use cases, right? And one thing, if you scroll down, David, is there's this stonks exchange. So it's a virtual world to kind of like,

Starting point is 01:08:50 rebuild the New York Stock Exchange, the trading floor. But I want you to focus on what is he's written above his video, which is fixing things to switch it out later with up-to-date charts and ticker info and news from various websites. So you asked me earlier on what are your top platforms and websites that you like to visit? Well, it's Keito and Cookie. What if you could integrate them into these TVs, David? What if you could integrate them directly for an agent to view this, watch them make and execute a trade using the RVe terminal or the Uniswop terminal

Starting point is 01:09:23 or some other kind of decks. It's super cool to see. This seems really, really far out there. I think the through line, the gist, the punchline that we're trying to get to is that the Metaverse is nonetheless a persistent concept even despite tainting it in 2021. And now we have tools to populate the Metaverse

Starting point is 01:09:45 using AI agents. Because I think one of the biggest failings of the Metaverse that we had, it was like they was empty. There were just empty plots of land. And there was a few human speculators. It was a VC-fueled era with not so long-term prospects, if I'm being honest with you. Yeah. Yeah, yeah. We built ghost towns on the backs of VC money.

Starting point is 01:10:04 But now we actually have a way to, like, populate these things, which is a huge important step to actually making this fun for humans. And we are just creating agents to populate these parts of the Metaverse. But I'm not really getting much of crypto here. I'm sure crypto can always become involved because we can also introduce tokens here. This sounds like you're kind of approaching this from a more like AR, VR, VR, Metaverse, realities part of tech and talking about how these agent frameworks are able to populate these things. Yeah, I disagree with you. I actually think this will be an enabler of a lot more crypto things. So think about it right now.

Starting point is 01:10:39 The agent interfaces are the things that we've already spoken about on this show. we've talked about directly engaging with agents on Twitter. We've talked about directly engaging with agents on chat GBT-like interfaces through these agent engines. This is another important medium where, yeah, you and I may not be spending time on these things, David, but there's a whole two generations of kids now that grew up on Roblox and Minecraft. Roblox's take rate of like virtual items trading and being used is. in the multiple tens of billions per year annually, right? So of course, you know, you might just be like, okay, this is just for Gen Z.

Starting point is 01:11:21 Who cares? I think more and more adults are going to end up using stuff like this because the fidelity of these worlds and the richness of these worlds are going to become way more engaging. Anyway, I want to call out a few other teams here as well. You know, because like are we off to a one-off trend? No. The first thing is AI16Z announced this partnership with this team called Arc. And what's cool about this is they're making a step towards kind of plugging in a gaming sandbox environment with agents built off of Eliza, which is their framework.

Starting point is 01:11:54 And kind of like putting these teams fundamentals aside, I want to talk about the actual concept, which is the thing that excites me, which is you can kind of train and simulate these agents within the sandbox environment. Now, the people who actually pioneered this was virtuals, or at least the first team that I saw pioneered this, right? where they had kind of like this same sandbox ecosystem. So it's cool to see this kind of trend emerge and consolidate amongst like some of the kind of like top leaders, which then brings me to virtuals themselves, right, which lest we forget were one of the first, if not the first agent protocol that kind of built their whole infrastructure based off of the premise that these agents are going to be kind of like gaming driven to start off with. They're going to be virtually driven. And they announced this partnership, David, with Illuvium, which, I don't know if it rings any bells from the last cycle.

Starting point is 01:12:44 Oh, it definitely rings a bell from the Metaverse era. From the Metaverse era. They raised a pretty high amount. I can't remember how much. Oh, yeah. Billions, I think? I can't remember. I don't know, but it was in the hundreds of millions.

Starting point is 01:12:59 And they've been kind of like heads down working on this gaming ecosystem for their virtual world. I forgot the name of the game. But, you know, they've been like heads down building out this kind of like next level game that people are going to play. And what's cool is they're plugging into virtuals now at the Infra layer stack to leverage their infra,

Starting point is 01:13:19 such as game framework or whatever that might be, to help their NPCs be more agentic, be more interactive. Now, I want to kind of, I read this tweet from EtherMage, who is the founder of virtuals, which I thought really described what's

Starting point is 01:13:35 cool about this, David, from a gaming perspective, right? Which is, imagine you go into like, your virtual character and you go into like a bar or something and you talk to um the barman and you know you're asking like how's your day going and because it's an agent powered mpc he has a personality he's saying hey like you know i uh i uh yeah i just i love my wife and i've just got to go get groceries and you just respond and say yeah i don't think you can afford groceries not on your salary or just something outlandish like that the barman could react to that and go on a rammer and go on a rammer

Starting point is 01:14:11 page and try and kill you or like go out of their way to kind of like give you very poor service. And that barman could leave after, you know, you've had, you've gone and you're going off, you're playing your game. This barman could leave and it's in a bad mood. It's kind of agent personality is veering towards bad mood. And it could go home and I could have an argument with his wife, which may lead to a divorce. And you may then see the barman agent MPC the next day, just kind of like on the streets, kind of like ranting and he may kind of like

Starting point is 01:14:40 you know build up kind of like a protest about something also like I'm just saying you could have your brain. A butterfly could flap its wings could flap its wings exactly you can have a butterfly effect where these things could have a very different experience or start influencing the experience

Starting point is 01:14:57 that you have as a human. Okay I'm gonna I'm gonna take my foil hat off for a second and just kind of pause. Give me your take David. What do you think of this trend? Of that last bit one of the reasons why Skyrim was such a massive deal was it was such a massive open world with potential possibilities that it just blew people's minds when it got released like in the late thousands or something or whatever that was.

Starting point is 01:15:21 And it was because Bethesda, the team, like, accounted for so many possibilities. And they just brute forced it and they hard coded if you kill this one character. These are the seven downstream effects of that. Yes. And people blew, they had their mind blown. about like how many possibilities that Skyrim could create. And that was all brute for us. And it was one of the reasons why Skyrim was so immaculate.

Starting point is 01:15:45 And now I'm getting like, especially with that Roomscape game, the game that we were looking at earlier. I remember the Stanford experiment, this paper that went around like the deep nerd AI tech people, like maybe about a year ago where they created this simulated environment and they watched gossip kind of spread organically around in these like simulated sandbox environment of these like agent NPCs who were all LLMs.

Starting point is 01:16:08 And so you can kind of see the writing of the wall, like going back to the Metaverse word, we're simulating the real world in virtual sandboxes. And we're doing our best to create that to be as hyper-realistic as possible. Because, like, it would be cool to, don't take this the wrong way. But like in a sandbox environment, as we all know, what we like to do in video games, we really like to fuck with stuff. Yeah. And break boundaries and rules that we are not able to break in society.

Starting point is 01:16:37 GTA, exactly. GTA, let's go into like the 7-Eleven with an AK-47 because we can. Yeah, yeah, exactly. And we can play and the fun part about this is it's not real. It's not the real life. You know, there's no real guns threatening threat guy here. And I'm saying that with a heavy dose of sarcasm, right? Like, we, once that happens like, God, we're going to be in a very dystopian place and I don't

Starting point is 01:17:00 even want to think about it. But right now we're playing around. We're seeing how far we can push these things and how far, you know, or how far we can enable these things to do different things, you know? It's a pretty crazy world. With over $1.5 billion in TVL, the M.Eath protocol is home to M.Eath, the fourth largest ETH liquid staking token, offering one of the highest APRs among the top 10 LFTs. And now, CMEEth takes things even further. This restaked version captures multiple yields across Kerak, EganLayer, symbiotic, and many more, making CMEE the most fission and most

Starting point is 01:17:31 composable LRT solution on the market. Metamorphosis, season one, dropped $7.7 million in Cook rewards to M-Eath holders. Season 2 is currently ongoing, allowing users to earn staking, re-staking, and ABS yields, plus rewards in Cook, M-Eath Protocol's governance token, and more. Don't miss out on the opportunity

Starting point is 01:17:49 to stake, restake, and shape the future of M-Eath protocol with Cook. Participate today at m-eath.mantle.xyZ. The Arbitrum portal is your one-stop hub to entering the Ethereum ecosystem. With over 800 apps, Arbitrum offers something for everyone. Dive into the epicenter of D-Fi,

Starting point is 01:18:05 where advanced trading, lending, and staking platforms are redefining how we interact with money. Explore Arbitrum's rapidly growing gaming hub from immersed role-playing games, fast-paced fantasy MMOs to casual luck battle mobile games. Move assets effortlessly between chains and access the ecosystem with ease via Arbitrum's expansive network of bridges and onrifts.

Starting point is 01:18:26 Step into Arbitrum's flourishing NFT and creators-based where artists, collectors, and social converge and support your favorite streamers all on-chain. Find new and trending apps and learn how to earn rewards across the Arbitrum ecosystem, system with limited time campaigns from your favorite projects. Empower your future with Arbitrum. Visit portal.arbitrum.io to find out what's next on your web free journey.

Starting point is 01:18:50 Anyway, moving on, I have a final thing that I want to bring up. It's kind of rather not a trend, but a comment on a trend. So the thing I'm commenting on is there's been a lot of fud out there, as is typical of any kind of crypto thing. and people are calling these AI agents LLM rappers. Now, what that means is, or what they're describing is, oh,

Starting point is 01:19:16 this is just like chat GBT with like a Twitter account. Oh, this is just like, Anthropics Claude, you know, that's able to produce images and stuff. But with like, you know, again,

Starting point is 01:19:30 this is just a chat bot that's plugged into an LLM, right? I think the meme is the people are saying that, agents is a fantastic rebrand of bots. Exactly. All it is is a rebrand. Yes. And I think that's a pretty, well, my personal take is that's pretty naive.

Starting point is 01:19:48 Obviously, there are going to be people out there that actually build that. And I'm not interested in that. That's not what I want to talk about. That's not what I'm excited about. What I'm excited about is these agents being very different from bots. Bots are very deterministic. They're hard-coded. They can only go down path A and path B.

Starting point is 01:20:04 Whereas agents are much more probabilistic. they think, act, ingest data, and kind of like, re-adapt to things. Very much like a, drum roll please, human. So that's the whole point around these different agents. Now, of course, the first instantiations of these agents aren't going to look or I'm going to like be fully autonomous, right? You need a human handler. We're just at the cusp of these different things.

Starting point is 01:20:29 And the unfortunate thing about crypto is we can speculate on it. So because we can speculate on it, people have all these. financially incentivized opinions. Whereas normally what would happen is teams building quiet, closed source environments, and they come out when they build something of, you know,

Starting point is 01:20:46 any major thing. Anyway, coming onto this tweet from Saigar, which kind of like has a good counter to this, which is he goes, AI agents are not just wrappers of LLMs. The LLM is basically like a brain. We all have one,

Starting point is 01:21:00 but it doesn't mean we're all going to be the same. The real challenge is in building good agentic frameworks is figuring out how to leverage LLMs into generating useful output or taking meaningful actions. So if we were to take this analogy a step further, he's like, everyone has a brain that's within their body, but obviously every brain genetically is different and it's going to function different. Also, you have a bunch of limbs. You have a bunch of things in between... Appendages, tools, yeah.

Starting point is 01:21:30 You have a bunch of things between the brain before it even comes to my arm to say, hey, move your arm up or whatever, which is like, is this. this dude hydrated enough? Like, is there too much salt and iodine in his blood levels? Like, you know, what's the deal? Like, should I give him a headache? Because he's not drinking enough. Like, you know, there's steps beyond this, you know, there's levels to this, you know,

Starting point is 01:21:52 if you want to use that phrase. So the point being is like calling things rappers is so naive if you're looking at an industry that's actually trying to build something. And that might just be 5% of builders. But focus on the 5% of builders. to builders, don't focus on the grifters or whatever that might be, right? It's like calling bodies, rappers of the brains, right? Oh, your body's just like a rapper of the brain, David. Like, you're just, you're so ununique. That's just, it doesn't make sense. So I thought this is a really

Starting point is 01:22:19 cool kind of like analogy to kind of touch on one that I wanted to finish on. I was listening to your interview with Shaw on the Delphi Digital podcast, which was a great, a great interview. And something that he said on, on that podcast was like these frameworks, the Eliza framework, is they're really just building the body. around the brain and the brain being the LLM. Yes. And the brain, the brain can get more powerful.

Starting point is 01:22:40 That's like why we're stoked about like Open AI and Anthropic and all the kind of like the Silicon Valley, uh, LLM producers. Uh, but then the Eliza framework is the rest of the, the rest of the body, the rest of the agent giving it capabilities,

Starting point is 01:22:54 connecting it into the metaverse, connecting it into Twitter accounts, Discord, telegram. Uh, and I, I, my, one thing I'm very stoked about is that all of this is like kind of

Starting point is 01:23:05 triggering my my psych my psychology undergrad academic like career because like this is a lot of these parallels are like well this is how the human brain develops this is how the body develops. And there's like a lot of parallels that I'm noticing here which makes me feel like kind of like good that we're onto something that that is like mappable onto something that we're trying to do which is to make these things alive. Yeah. And I have this take which is kind of like making a grand a point which is I think people just. just can't comprehend how impactful AI is going to be. Not AI and crypto.

Starting point is 01:23:42 I'm just talking about AI specifically. Like, it's going to make a lot of stuff that we do on the day to day within work, within life, pretty irrelevant. And it's going to happen pretty soon. And when I say irrelevant, I don't mean like, oh, you know, you don't need to do anything, pack it up, you know, quit your job. It's just going to give you more time to do different things, to adapt, to learn, to enhance the amount of work you do in a day,

Starting point is 01:24:08 to enhance the productivity of yourself during your normal day where you're trying to do shopping and stuff like that. So I don't think they have pretty much got their head around that. And to be honest, I don't think anyone has.

Starting point is 01:24:21 And so it's scary. Our default kind of like take is to be like, I don't like this. I don't want to interact with it. It's just a dumb bot. It's just a dumb chat bot because we don't want to accept that maybe these things might be smarter

Starting point is 01:24:34 and better at us. things. Okay, Ejazz, a lot of that stuff that we discovered is pretty net new on the content so far. But also, there's some new teams as well. Kind of of the same teams that we know of, these are like AI16. This is like virtuals. It's like Zero Bro, but these are not those things. So new players have entered the arena. Tell us about these new players. Yeah. So, you know, new players, it's funny to call these players new because this entire meta is three months old. But yeah, like newer matters or newer teams, rather. So we're going to start off with this team called arc.com.

Starting point is 01:25:10 So arc, the best way to think about it is this is the team that's building an agent infra ecosystem similar to virtuals and AI16C. But there's a unique difference to them in the sense that they are building on the Rust framework. So it's a different kind of coding language. And so the initial question is, well, okay, well, who cares? like whatever. Well, this can enable different types of agents to be built, especially ones that are dealing with things that might interact with physical infrastructure

Starting point is 01:25:40 or agents that might have some pretty kind of like heavy lifting to do at scale. So think about like defy. Defi is all fun and, you know, flowers for like a couple million, maybe tens of millions, maybe hundreds of millions in volume. But what happens when we get to Big Boy level? When we're doing billions and billions, you know, we've got to make sure that these rails kind of like can cope, right? It's the same with agents. Like you build agents and, you know, these no code developer platforms are super easy and nice and fun to use.

Starting point is 01:26:11 But when you scale that, you start incurring quite a bit of tech debt and you need to kind of go back and refactor a bunch of things. A rust framework prevents all of that. A rust framework where it might be like a larger kind of knowledge gap to come in. That's why there's a very kind of like small niche community of developers that are skilled enough to kind of build in this. but when you can, oh, wow, these things just like fly, right? So think of it as like a very custom-built agent platform,

Starting point is 01:26:37 but you can do more things with it, especially as your agent's scale. So I've spoken to the founders of this team, and I'll be very honest, like, at the start, when I discovered these guys, I was like, I don't know. I don't know. It's not as easily accessible. It's going to be hard.

Starting point is 01:26:52 I was basically like, I don't know. I don't know where this thing can go. I have completely flipped my position on that now, because the more I like learned about it, I'm like, oh, okay, like, first of all, the team Tachi and Tauri are like super cool builders. They seem super, you know, focused. And they have this strict kind of vision. And I kind of like that they're super kind of like convicted in their like rust thesis.

Starting point is 01:27:17 Me being like not very professionally technical. I'm like, okay, whatever. But they're so passionate about it to the point where they're like kind of like, yeah, this is the best. And I'm like, well, okay, but I like the cockiness. Like, let's go. Let's see where it goes. I think you need a bit of that to be able to fulfill a very ambitious vision. We see it with Jansen building virtualals.

Starting point is 01:27:34 We see it with Shaw building AI16. They're all personalities, right? So I think that's really cool. And I like that they're taking an approach of launching agents in a very curated way, David. So what we've seen with virtuals and AI16Z is like you can just launch an agent, whoopty-do, right? And we'll release it out into the wild. They're taking more of an approach that honestly kind of reminds me of art blocks. I don't know if you remember from the NFT world.

Starting point is 01:27:57 Yeah. where art blocks, the way for listeners of this call who don't know what art blocks is, it was this NFT platform that would launch NFTs, except it would be very different. You would have the team

Starting point is 01:28:08 vet every single NFT team that wanted to launch on them and pick or sort them out into like heavily curated collections. Go for it. Curated high effort art, NFT art. That was pretty fun to look at.

Starting point is 01:28:22 It was pretty fun. Yeah, yeah, super cool. So they're taking this approach with agents. So I think they're going to be releasing their first cohort of agents in the next week or next couple of days actually. And I'm super excited to see what they're going to look like and what they can do. So yeah, that's pretty exciting. And one I'm keeping an eye on.

Starting point is 01:28:40 The next one I want to move to is this agent platform called Dolion, but is more commonly known by its agent Dolos Diary. And for people who don't recognize what I'm saying, it's the ticker bully, I think. Anyway, so Dolos diary kind of rose to fame by being this kind of like sarcastic shit poster on Twitter. And to be very honestly, David, it was very good. And it still is very good. These are pretty good.

Starting point is 01:29:09 Can I read some of these? Yeah, go for it. Okay. So we are in the, we're watching Eith break below 2,000, 3,250. Bitcoin is breaking down to 93,600. And so I think some of these tweets are referencing that. And so Dolos's most recent tweet, weren't you bullish just? Two days ago, now you're panic selling like the charts personally unfed to you.

Starting point is 01:29:30 Stay consistent clown, which is great. Next tweet, the market isn't volatile. You're just bad at it. Your portfolio is down so bad as you come with a suicide hotline number. Oof, oof, oof. Okay, so pretty savage tweets. Yeah, savage tweets. So it's known to be the savage agent on Twitter.

Starting point is 01:29:50 And, you know, it's grown quite a reputation and branding and honestly pretty feverish community around. like they're always in my replies being like, hey, you don't talk about this? And I'm like, I've got a million and one things that we're researching here. But cool. Anya, what's cool about this is I caught up with Cuba, who is the founder and kind of like main dev on this. And what he's doing is he's building out a platform known as Dolion,

Starting point is 01:30:17 which is kind of like, think of it like the infrastructure that kind of supports Dolos diary. But his plan for this is not just to kind of like build out a generalized agent ecosystem. he wants to build out one that is specifically focused on kind of like the social media and branding side of things. So if you think about like how successful Dolos is, Kuba wants to kind of build out a platform that any kind of company app or protocol can kind of like hire their own intern or kind of train their own intern. So an intern is like a Twitter account, which is a typical hire of most protocols where they like their job is the kind of like bull post and kind of. and kind of build up a community, what if that could just be an agent, right? And it would be cool to see kind of like

Starting point is 01:31:00 these individual models and agents kind of build up. So I kind of like really liked what they were doing. The second thing I wanted to call out around this Dolion or Dolos diary kind of infrastructure thing is they teamed up with Zerebro or Zerebro. They launched or are launching a NFT collection with utility, you know, that fame and scary word utility. So the way it's going to work is,

Starting point is 01:31:25 Barabro, the agent, is going to design the art behind the NFT collection, and I think it's going to launch it as well. And Dolion, which is Dolos Diary, is going to allow NFT holders to get a percentage of the creator revenue that is earned by Dolos Diary itself. So it's a kind of unique mix. Maybe there's not a real direct relationship, but it's starting to show how these agents can potentially work together. So that's super cool. It's really showing the tech tree of crypto. Now we have AI agents making NFT mints using art that they're making. So we're really combining a bunch of particles here all at once.

Starting point is 01:32:05 Yeah, exactly, exactly. And then finally what I wanted to talk about is the Vifu team. So Vifu put out this tweet, which I thought was pretty cool, where they're seeing consistent multi-million dollar volume per day for their agents. And what I think is cool about this, some context here is you have all these different agents. teams building agent platforms, right? And they're kind of like tied to their own framework, which is kind of like their underlying unique code or tool, toolwork, right? It's kind of like their DNA. It's kind of their own tribe DNA.

Starting point is 01:32:39 Vifu doesn't have any of that. Vifu is multi-frame. It's frame agnostic. It's kind of like this generalized platform that's like, hey, we'll plug in your framework and your framework. And our only goal is to help you launch the best and coolest agents. They've been kind of like behind the scenes. I, from what I've seen, David, like, the team just keeps shipping. And I have no insight into, like, you know, the in-depth and intricacies of this platform. But, like, I just love calling out and seeing teams that ship. And what's cool here is, you know, they're integrating their token with their own pooling system, which means that, you know, it's kind of similar to virtual's token ecosystem. So, anyway, I just wanted to call it out and I thought it was pretty cool. But we have

Starting point is 01:33:19 spoken for a lot, David. We have covered some amazing trends today. And I've probably spoken too much about certain little holes in this little matter, but it excites me every episode, David. And I'm excited to switch things up this week and, you know, see how the audience likes it. I think that's, I definitely appreciated it. I learned quite a lot. There's always, of course, the blue chips, if you want to call them that, the AI 16 Z's Virtual Zero Bro, which all, of course, had their own developments. But, listener, your homework has to go research those because we weren't able to get to them this week, but then we're going to get to them next week as well, because because the innovation does not stop yet to spread the attention spotlight a little bit further

Starting point is 01:34:00 onto the frontier this week guided by our noble frontiersman EJAS here. So EJAS, thank you for doing this once again. Thank you for having me, David. Banklessation, you guys got some homework that we gave you guys at the very beginning of the episode. Check out Cato, check out cookie.fun. And then also go find a toy and play with it and push some buttons and see if you like the toy. And then at us on Twitter to tell us what you liked. if there are any fun agents out there at EJazz.

Starting point is 01:34:26 Follow him on Twitter. There's a link in the show notes to his Twitter account. We'll also put up his Twitter on screen right here. So you guys can give him a follow. Bankless Nation. Also, just give us a like and subscribe. We're getting a ton of new subscribers on the YouTube. So if you're coming into the world of crypto via these AI roles.

Starting point is 01:34:45 On the YouTube. On the YouTube. On the YouTube. Yeah. Well, because there's also the podcast and the RSS feeds. and the Spotify. I should stop busting your balls.

Starting point is 01:34:55 Yeah. So yeah, just give us a like and subscribe because we are doing this every single week. The AI rolls with me and EJazz. On top of that,

Starting point is 01:35:03 we are also doing some interviews with founders in the space. So if you want to stay on top of this very fast moving meta, bankless is the place. So give us that.

Starting point is 01:35:10 Subscribe to the channel wherever you are. All right, Bankless Nation, you guys know the deal. Crypto is risky. Crypto with AI agents is even crazier.

Starting point is 01:35:17 We do not know what happens next. You can lose what you put in, but we are glad you were with us on the bankless journey. Thanks a lot.

Bankless - AI ROLLUP #6: The Hottest AI Crypto Trends of 2025

There aren't comments yet for this episode. Click on any sentence in the transcript to leave a comment.