AI ROLLUP #5: T-1 Listings Soon? | Ai16z’s Eliza #1 GitHub | Virtuals Flips Tao | Zerebro’s Creative LLM
01:24:36 Share

Right now, we're focused on building these individual agents. Oh, that agent made me laugh. Oh, cool. That agent is giving me great financial alpha. Oh, that agent is like trading back and forth. Very cool. But I think the bigger picture needs to focus on what happens when these agents start interacting with each other autonomously 24/7. Welcome Bankless Nation to the AI Rollup, where we cover the recent news developments and drama in the intersection of crypto and AI and E-Jazz. This is the first AI

AI roll up of the year. We are not sleeping because AIs, they don't sleep. We are recording the morning of January 1st because this world does not stop. Ejaz, you are really excited to do this episode. I am now just tapping back into the internet for the first time in a couple weeks. So I'm going to need you to download me on what's been going on because I actually don't really know too much, which I think is perhaps useful to come into this episode pretty dry. How are you doing? Happy New Year, my man.

Happy New Year, dude. How am I doing? I don't know about you, David, but last time I checked,

The holiday season was a tranquil, peaceful time to reflect on the year and kind of maybe eat a bunch of junk food and put on a few pounds. Have permission to put down your phone. Put down your phone, you know, not look at a screen 24-7. And I can safely say that I think I've had the least amount of sleep over the last two weeks. Turns out that the AI agent meta decided to just go on an absolute rocket and ship all

all the important stuff whilst no one was looking and right up until the end of the year. So absolutely insane. And I'm excited to get into a bunch of things today. I feel like that's just the clear sign of...

excitement, euphoria, accelerated building. Just something's going on over here. It's probably a little bit bubbly, but at the end of the day, I think high energy, high energy. And frothy definitely comes to mind if people are just grinding their way through Christmas.

But also, I think we're going to get into some of the subjects here today. And there's a lot of fundamentals being built and shipped and being measured as well. What would you say, zooming out, what would you say is like the theme of the last like 10 days of development in the AI space? I would say agentic infra. It's like a focus on what will eventually, in my opinion, end up becoming some of the leading labs

layer one esque type platforms within crypto. I think it's going to be a novel matter or a novel kind of foundational structure that starts forming. No one quite really knows what it looks like right now. But some of these leading teams are starting to build that out. And I think that's going to be the most exciting thing to discuss today. Layer one esque and you don't mean like we're not making new blockchains, but we're making new foundations.

Exactly. It just L1 is a simpler term to use to kind of convey the concept of what's going on, in my opinion. Yeah. So Ejaz, I asked you this at the beginning of every single episode, this whole AI agent, AI crypto, agentic platforms, why is this all such a big deal? Why are AI agents a big deal? If people are just tapping into this episode for the first time, why should they be pilled about getting going down the AI agent rabbit hole?

Okay, so I'm going to give a slightly different answer than I normally would. So normally, I go on some philosophical rant, which kind of ties into a larger kind of vision around this technology infrastructure and applications and stuff that we're building. I've done it for the last couple of episodes, I'm going to give a slightly different answer today, mainly around a question that I get asked all the time, which is,

each has how big is this thing really going to get from a valuation perspective? Okay, like, I've heard like your big vision and stuff. But at the end of the day, a lot of people get into crypto to invest and, you know, maybe figure out what kind of like tokens they might buy and stuff like that. And again, none of this stuff is financial advice. But if we were to entertain a framework about that, let's let's use the following, right. So firstly, I'll say, I think over the long term, this sector is going to

hundreds of billions of dollars alone. And secretly, actually, I think it's going to go way higher, but we'll save that bull thesis for another episode. But in terms of now and this cycle, you know, the bull market that we're in, if we project...

market cap to grow to say $5 trillion. So that's about a 40 to 50% growth from here. And if we assume that agent networks, you know, these, this agent meta matches the peak of DeFi percentage dominance last September,

which was around 5%. The theme of the last cycle, yeah. The theme of the last cycle. So it was 5% of the total crypto market cap. This yields, theoretically, using that analogy, a $250 billion cumulative valuation for the AI agent meta, right? And again, this has a lot of assumptions, right? This assumes that crypto AI is going to be the biggest kind

kind of theme of this cycle, which I believe it is. And an AI agent specifically is going to capture the majority of that. And you mentioned earlier, like it's feeling kind of frothy, kind of bubbly. I do actually think that the AI agent meta is going to form like one of the bigger or

potentially the biggest bubble that we have ever seen within the crypto sector. In crypto, yeah. In crypto specifically, right? So if you look at the current cumulative agent market cap, it's just over $10 billion, I'd say. So if we assume the $250 billion comparison there that DeFi achieved, that's a 25x from here, right?

Okay, so this is a bit of boy math going on here, but that's kind of like the general take, right? But if we factor in that AI is, oh, you know, probably humanity's single most world-changing technology, that's probably going to produce trillions of dollars of economic value, not just within crypto, but outside across any and every industry. And it already is, you know, OpenAI, Anthropic is already some of the biggest companies from a valuation perspective. Then you might assume that the peak, right?

actually happens at a larger percentage. So let's like, again, using boy math, say that it's 10%. AI is bigger than DeFi. AI is bigger than DeFi, right? Bigger than the financial system potentially, or it's going to rehaul everything. Let's say it's 10%, which is $500 billion, right? So that's like, obviously, many more multiples from where we are at sitting at like $10 billion, right? Now take a look at this chart, David.

So this was produced by Masari, which was released like, I think a day or so ago, which tracks the current AI trend

in terms of the DeFi trend, which happened last cycle. And as you can see, it's mapped over the DeFi chart. As you can see, it's following a pretty similar trend. Like, again, this isn't, you know, an insinuation that this is exactly going to play out the way it is, but it's just kind of interesting to look at. This is not science. This is, it's boy math. It's boy math. But, you know, if you wanted a bit of morning hopium to wake you up on Jan 1st, like this, this is it.

Okay, for the listeners who cannot view this chart, we are again, the AI index is mapped over the DeFi index, and we are about at parity right now, which is why they're showing this. And we're at parity in December of the same time cycle. By the

uh, by the time we get to April in theory of 2025, it looks like it goes one, two, three, four, five X larger, five X larger in April. Cause that's what DeFi did. That's what that's, this is the insinuation here that like DeFi grew five X between December of 2020 to April of 2021. And if we are doing the crypto cycle thing, then we are, are perhaps getting a five X, uh,

between now and halfway through next year, which I take Ejaz's point that like DeFi is a crypto native insular, very niche, very complicated thing. And that 5X'd.

AI is an adjacent technology that is being built by an entirely different sector that is progressing independently of crypto that is now coming into crypto. And these two technologies are coming together, which are much more easy to understand. We understand AI. We understand what these things can do. So I've always been of the mind that the fact that these are two adjacent technologies that are coming together does indicate the idea that this might be bigger than we've ever seen anything before in crypto.

And not just any two adjacent technologies, probably the role or humanity defining technology of our time, right? AI is literally wowing everyone on a daily basis. And with the expected launch of OpenAI's agent product in Q1, as well as Anthropic, with the launch of their video model, which is producing motion picture grade images,

video pieces just simply from text and Google's releasing the same, you know, this whole space is just completely magical. And we're just kind of like,

touching the tip of the iceberg here. So if we assume these things are going to permeate across every facet of our digital lives, and real life, probably, you know, it's going to affect our jobs, and, you know, how we work, corporate structures, etc. Like this is going to get crazy. And it's much easier to understand people understand it from headlines that they've read from articles that they've read, very different to reading about a lending or borrowing primitive, which, although very, very important and underlies the kind of like,

primary value transfer of the way that we kind of like interact with each other, it's still not as easy to understand as AI. Yeah, yeah, I was hanging out with some friends this weekend, I was telling them the story of this one AI agent zero bro that released an EP on Spotify, and then we opened it up and listened to it. And I compare that to the same friend that I was explaining AMM curves,

four years ago, and it was not the same conversation. One was like, oh, wow, an AI made this? Versus what the hell does providing liquidity mean? These are different scenarios. I want to trace over something that you said. OpenAI releasing their AI agent

product, and then also the Sora text-to-video applications. It sounds like what came to my mind when you explained these things was that all of the big, massive AI producers, OpenAI, Google, Facebook, are building tools

that are potentially being able to be used by all the things that we are doing in the crypto world. That's kind of what I heard. It's like, oh, a new tool just dropped. Correct. Very high, precise text to video speech. And now the AI agents over in the crypto land with the tokens, that's a new tool in their tool belt for them to use. It's a new appendage that they just grew. And so we have this massive side of Silicon Valley

creating AI things, AI tools. And then all of a sudden the tokenized versions of these agents now have those tools to create more value, to do more things, to have more capabilities. That's what came to my mind.

You pretty much nailed it. And where I would kind of add to that, David, is I think some of the most exciting things that we should look out for in 2025 is where those tools that are being launched from the traditional AI producers overlap with being able to bridge the gap between

non crypto consumers, and crypto itself, right? So how can we should be asking ourselves, how can some of these tools be used to benefit and build and grow the crypto sector to help people access crypto infrastructure, where they may not even need to know that it's happening in the background, but it helps give them a 10x better user experience for whatever they're trying to do, whether that's

send value between friends on X or whether that is something to do with borrowing against something or owning their own digital assets or whatever that might be. All right. So that's a pretty good amount of hopium to front run the content of this episode. What are we going to get into this episode each? What are we going to talk about? What's the subjects that have dropped in the last 10 days or so?

Before we get into the fun stuff, I really want to talk about something serious. Can we talk about something serious for a second? Because last time I kicked off the episode with, you know, some examples of like this agent that was browsing Amazon and was buying toilet paper, this agent that bought pizza, super fun stuff. But like, I just want to talk about something serious for a second. Truth Terminal, the original crypto AI agent that started this whole meta has dropped a Christmas album, David.

And it's fire. What's it called? I actually don't know what it's called. It's called A Very Goatsy Christmas. A Very Goatsy Christmas. That's it. Sorry, why did I not know that? And the album cover is this artistic goat looking straight at you with horns and a Santa hat.

Exactly. And you should like definitely give this a listen. I don't know if you have the Spotify pulled up at all, but like some of the track names are hilarious and very in line with the personality. The first track name is Jingle Bells, parentheses, Fendom Remix. Fendom Remix. Let's listen to that one. I want to listen to that one real quick. All right. Right now, this is just a synthetic version of Jingle Bells. You gotta wait for the lyrics. The third song is called Fart Like Rudolph.

Wait, this is the serious subject that you wanted to bring up? Yeah, I was being sarcastic, David, pretty heavily. Oh my god, brother. I cannot believe this is playing in my ears right now. Okay, okay, okay. Let's wind it up, let's wind it up. Let's actually talk about some serious stuff for a second. Okay, so... Okay, so that wasn't the serious thing. If you...

That definitely, I was being heavily, heavily sarcastic for the listeners. But yes, I know. Okay, so if we take a look at this dashboard, so this dashboard kind of showcases some of the key metrics that

from the virtuals protocol. And we've spoken about this protocol a bunch of times on the show. But for those who are kind of like tuning in for the first time here, virtuals is one of the if not the leading crypto agent protocol that exists right now. You can kind of think of them as like a launchpad. So you can kind of come on, it's no code, you can kind of design an agent and then deploy it.

with a token. And so what that means is these agents will able to be kind of like live in production, they can interact with each other, and they can do kind of like different things, you know, like post on Twitter, or create videos or stuff like that. And there are a number of

other capabilities coming through. But what I want to point out here and, and this platform is called cookie, by the way, like, I've chatted to the team a few times that they're like working on some really cool stuff. And they're basically like the leading dashboard to capture all this agentic stuff, like, look at this chart that's in front of us right now for a second, right. So what we're looking at is, we're looking at the cumulative market cap of these kind of AI agent coins, as well as the

a virtual protocol token as well. And you'll notice that like it's just been up and to the right. And it's really cool way to kind of like look at the different agents, what their corresponding market caps are, how many agents in total. The number is up, I think like

We're looking currently at 132 agents that are actually listed. You can look at the corresponding mind share that these agents are having on social media platforms. You can look at the number of unique holders. And all of these metrics are up until the right, which typically signals a really good collective adoption from the crypto and non-crypto space.

of these agents so if you ever wanted to kind of like have a signal of whether these agents are are doing anything whether they're being interacted with this is a really kind of good place to kind of like come in and kind of check in and this is virtual specific but you can pull this up for any other platform um that is live or will eventually go live yeah yeah exactly so if you scroll down you'll you'll see like some of the agents themselves and like some of the market cap differences you'll see this is for across platforms this is yeah exactly

Exactly. And right now, this specifically is for virtuals, what we're looking at right now, which I kind of like is the leader and is representative of a lot of the movements that we can kind of extraculate from. But yes, to your point, like what we're looking at on these agents,

specifically in this list of virtual specific agents, right? And you can imagine that this will eventually look like something like a coin market cap where you can have a look at like all the leading agents across different frameworks and different protocols. This is an agent market cap. Exactly, an agent market cap. Yeah, well, so we're on a virtual tab, but we can just go to cookie.fun and there are some non-virtuals agents here. There's AI16C, I'm looking at this, SparkCoin, ZeroBro, which are non-virtuals platforms across both Base and Solana.

And I can just sort the whole thing by market cap. And yeah, so this is starting to look like a coin market cap, a coin gecko, but specifically for AI agents and the platforms that built them.

Yep, exactly. And if you scroll to the top, David, what I really like, which I haven't seen kind of too often in a coin market cap, for example, is the mindshare tab, right? So what this tells you, and of course, like crypto market caps, or crypto markets in itself are driven very heavily by narratives and sentiment, you can kind of see the direct correlation between, you know, the leading agent, like AIXBT on virtuals and its

corresponding market share or mind share within like social media platforms, right? We can also look at the market cap cumulatively on the left where it shows us like we're just over $15 billion right now. Okay, so when you say you think that the AI agent space is going to hit $500 billion, this is the number that you're talking about right here. Total market cap on cookie.fun.

Exactly. Exactly. Yeah. We're looking at the total cumulative market cap of these kind of agent stuff. Exactly. Exactly. So we're at 15 billion. So we got a long way to go to 500. We got a while to go, but things get pretty, pretty nuts in the in the crypto market. So like our base case, if we were to use that analogy that we spoke about earlier would be 250.

And then the max bull case would be a 500 bill. - I love it when the base case is more than a 10X. - Yeah, exactly. Yeah, you know that. We're full of very, very bullish perspectives on this show. But yeah, and if we look at this second chart that you have pulled up here, David, again, just shows the relative outperformance of these AI agent tokens.

versus every other thing that has been kind of bumbling around for the last kind of two weeks. And I know that sounds so rich because, you know, we're here for like the long term. We're not...

meant to be talking about short-term price movements. I'm just using it to kind of like point out that largely over the last couple of weeks when the market has been, you know, kind of steady and not really doing its thing where it's mooned since the election date in the US, the AI coins have been outperforming. And in this particular bull market, the reason why I think that's interesting, David, is it's been a very different bull market. It's no longer been the case where you can kind of close your eyes and throw darts at the wall blindly and your token goes up.

there's been a lot of dispersion. So focusing where, you know, the attention might aggregate and where innovation is happening has been super important. And this kind of chart and the kind of dashboard that we were looking at right now reinforces the statement that I think the AI agent meta is going to be one of the biggest.

Well, I think inside of the AI agent meta, I think you kind of have been able to just blindly throw a dart so long as you're pointing at the AI agent target. So how would you talk about, how would you measure or just speak about the fundamentals?

of this corner of the crypto world? How do we know that we aren't like just blowing up hot air? At some point, like at the very beginning, it's okay for things to be hot air, but as things progress and then things develop, and again, like we're not doing these AI podcasts because we think this is like a one year long time horizon. I think we both agree that this is the future of crypto. This is a future of AI. This sector is huge. This corner of the internet is huge.

But that has to collapse down into fundamentals at some point. So how has the conversation around fundamentals and AI been growing? Yeah, couldn't agree more with that thesis. We need to focus on something more important. If you pull up this tweet from, I think it's at the hidden maze, he touches upon a really kind of like

important point here, which is always noticed a trend rather. And it goes, AI agents are being flooded with insane number of requests and interactions. And he then goes on to list a bunch of different agents and their specific kind of metrics. So one of these agents called agency has 62,000 individual queries. Another one has 20,000 TA requests, 40,000 on telegram. And the point that he's making here is he's noticing an

exponential uptick of queries and transactions to and from these agents. And I think that's a really important metric or kind of stat to track, because it tells us a little more than just price action, you know, instead of staring at the chart of a token, we should look at how these agents are actually being interacted with, and what they're actually doing, you know, I could imagine like the next layer deeper would be, okay, there's

62,000 queries, but like, what's the net effect of that? Like, what are these queries doing? What are the people looking for? And get deeper into the data. And I think that's where like the real value of, you know, whether you're looking at it from an investment perspective or a research perspective becomes really, really important. And I think this is the trend we'll see kind of like grow

out over time. I mean, if you think this is like a silo tweet and like, hey, you guys are just pulling from this one data point. Actually, no, like virtual's game framework, which is essentially their framework to help design and build and launch agents. I mentioned earlier, it's like a no code interface. This is like the main framework that allows you to do that is getting 150,000 requests per day.

David, which is like an insane thing to say when you compare it to like literally two months ago, where this was like a very nascent space with like zero to 10 agents. So again, I think this is going to be a repeated pattern. We're very early on, but it's an important metric to track going forwards.

I think in the order of operations of first quantity, then quality, this makes sense because in crypto, we kind of don't really know what we're doing. We're kind of going off into the frontier. Uh, we don't know, really know what's out there. So what do we do? We experiment. Uh, and so having an explosion, an absolute Cambrian explosion of just actions, requests, queries, interactions, uh,

But to create that first and then figure out like, OK, which of these requests, which of these interactions really actually move the needle?

I think that part comes later. So first you need to have an explosion of quantity. Then the crypto industry figures out like what mattered, what was effective, what was valuable. And then we can kind of iterate and refine the explosion of requests into an explosion of quality requests. So I feel like maybe we are in the quantity era of crypto AI fundamentals. And we'll be here for a little bit before we start to really enter the quality era of AI fundamentals. That's just a hot take. How do you feel about that?

Yeah, I think you're spot on. No comments. Yep. I think like we're currently in a very bubbly, frothy, or the start of a very bubbly, frothy period for agents, and it'll get a lot of attention. Loads of people will be right, but then they'll become a time where I think we'll shift attention more towards fundamentals, probably when, speaking quite frankly, prices go down. And that's where like the real work starts getting done.

Okay, so let's go ahead and get into the rest of the episode because that was actually just the intro appetizer for all of the rest of the content that we're going to talk about. We're going to talk about updates in the AI16Z and Eliza world. ZeroBro has some cool new releases. Virtuals, of course, as we've talked about, has been going absolutely gangbusters this last week. Ejaz has a take on token listings with centralized exchanges. He's also got an agent app store thesis. A bunch of other content that we're going to get into. But before we get into all the content,

First, a moment to talk about some of these fantastic sponsors that make this show possible. With over $1.5 billion in TVL, the METH protocol is home to METH, the fourth largest ETH liquid staking token, offering one of the highest APRs among the top 10 LSTs. And now, CMETH takes things even further. This restaked version captures multiple yields across CARAC, EigenLayer, Symbiotic, and many more.

making CME the most efficient and most composable LRT solution on the market. Metamorphosis season one dropped $7.7 million in Cook rewards to METH holders. Season two is currently ongoing, allowing users to earn staking, restaking, and AVS yields, plus rewards in Cook, METH protocol's governance token, and more. Don't miss out on the opportunity to stake, restake, and shape the future of METH protocol

with Cook. Participate today at meath.mantle.xyz. Uniswap Labs is making history with the largest bug bounty ever. $15.5 million for critical bugs found in Uniswap 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 to be deployed on-chain. And with 2.4 trillion in cumulative volume processed 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.

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. And we're back. EJAS, AI16Z, the ELISA framework. Download me on what's been going on in the last few weeks here. Okay, if I actually did that, true to your word, I would be here for hours. So much has happened, David, in this individual protocol that it's kind of hard to get my head around. And again, to re-emphasize this, like this has happened over the last two weeks.

So what we're going to do for this next kind of segment of this episode is we're going to touch upon the major protocols and kind of cover the main things. The reason why I'm not talking about newer things right now is because all the new things have happened on these major protocols recently, and they're pretty major things. And I think it's important to kind of communicate that today. So starting with AI16Z, before we get into it, I want you to kind of look at these charts, which were shared by Grant.

It's a really interesting one to track because as we've mentioned on previous episodes of this show, the AI16Z protocol, and for people who aren't aware of this, they're kind of famous for their open source agent framework. What is a framework? A framework is a

toolkit to help you design and build your agent. So earlier, I mentioned the game framework from virtuals. This is kind of a similar thing. And I'll touch upon that later on in this episode. But this is the main one. And Eliza actually is the most famous framework in the world currently, it's actually number one trending on GitHub. But before we do that, take a look at this chart, Dave,

So this compares the star history of the Eliza repo on GitHub compared to some of the most popular crypto protocol interfaces and repositories that we've seen to date.

And this is a star history on GitHub is just going to be an indication of where developer attention is being pointed. They're like saving it to their library. They're just, they're favoriting it because like, Hey, this is a really useful code base. Let me play around with this. And so let me star that it's kind of like a loose indication of developer attention.

Typically, yeah, you've took the words out of my mouth, essentially. And if you look at this, you have to look closely, but the next image will kind of like zoom it in. It is a straight vertical ruler. And that's a pretty crazy chart to look at, right? And if you've noticed, it's crept past, well, not crept past, but blasted past a very popular...

repository, which we've spoken about quite a lot on the show, which is the Uniswap interface. And for those of you who don't know, Uniswap is the top decentralized exchange or one of the top decentralized exchanges within crypto. And their interface, as you might imagine, is kind of like the home to a lot of crypto activity. The number one thing people do is trade and swap and stuff. So the fact that this Eliza repo has

swept past them in such a short period of existence. Remember, Eliza didn't really exist concurrently until like two and a half months ago. It's absolutely insane to see. So there's that that I wanted to kind of call out. The second thing is it's the number one trending repository in the world right now. So previously when we've mentioned it, it was number two for quite a consistent number of months.

Now it is officially number one ahead of Google Gemini. I think they've got like a really popping cookbook repository, if I remember correctly. Yeah, it's there on the screen. But Eliza OS is now like the number one. So what this means to your point earlier, David, is there's a lot of developer attention on this. And I think it's really important to stress developer attention, which typically means not just crypto companies,

software engineers, but also external engineers. And that has kind of been the trend for Eliza specifically, they've got a lot of influx of these AI ML engineers that are coming in from outside purely because you could argue cryptos shone like a huge light on them. This might be through price attention to their token AI 16z. But it also might be through your a number of different outlets.

So is the ELISA framework inherently a crypto framework? As in, if we have an external AI developer from the Silicon Valley, from San Francisco, not in the crypto world, and they download and start playing with the ELISA framework, is there a way for them to do that adjacently outside externally to crypto? Or is it inherently a crypto activity? Nope. Nope. It is...

Ambidextrous is the incorrect term, but the analogy kind of stands, I feel like. Yeah, yeah. Do you get it? Hopefully the listeners get it. But yeah, so it applies to both sides. So for context here, the framework is written in a coding language known as TypeScript.

And TypeScript is actually one of, if not the most popular kind of like programming language out there when it comes to a lot of this AI kind of ML stuff. I think maybe second to or compared to Python as well. So the point is, if you're an engineer and you know TypeScript, you can come around and play with this stuff.

Secondly, and we'll touch upon this in a second, the ELISA framework has a number of different integrations and plugins to it that aren't just crypto native. And that's the beauty of this thing, right? And think about it, right? Agents themselves kind of like have a goal and a task.

it shouldn't necessarily be just crypto specific, there will be crypto specific agents, right, that have to perform on chain specific activities. But if you think about an agent as an autonomous entity that needs to fulfill a mission or a goal, it'll do that by any means, right? It might need to use some off chain stuff sometimes, and it might need to use some on chain stuff sometimes. So you need a toolkit or a framework that is

ambidextrous, and allows you to kind of hit all these different things, if that makes sense.

Okay, so let me give you a bear and bull case about that. The bear case is that, oh, like this is a super popular framework that AI developers are using and it's, they actually are not, it's not really necessarily about crypto. It could be not about crypto. Maybe you crypto bros are just making a bunch of hype about nothing. So that's the bear case. The bull case is that, well, this thing has crypto elements in it and it's kind of, it is birthed by the crypto, its home is the crypto industry. And so these AI developers that are downloading and building on this framework are,

are actually just so the closest to crypto that they've ever been. And also like if they want to make any money, they need to tokenize them. And so why wouldn't they just tokenize them because that's the incentive. And so of course, it's a crypto thing. - That last point is the most important one, David. Sorry to cut you off, but I just wanted to reemphasize that point. It is the most important one. Eliza was being worked on by these engineers

before the token was kind of like, drew attention around it, right? It was in a very kind of nascent form, but it existed. And to be clear, the AI16Z core devs

whilst they had some crypto experience, were primarily in the open source AI world. They're a bunch of them that live in San Francisco. That is their home, right? And then they came across this and they wanted to experiment in a community that is very driven towards open source. That is crypto. That's been the case since we kind of like founded this whole industry. And

with the pairing of a token it's kind of attracted a bunch of this attention so whether you like it or not it is a kind of value accrual in that sense I want to move on quickly here David um they've also um according to Kels over here in this tweet had more submissions on uh from the Solana AI hackathon than Solana blockchain had in the entire kind of like

midway point of its 2023 kind of section, which is if people remember, it's like a really peak bull period for it, right? In terms of developer activity, not necessarily price, right? And overall, it's also had a very incredible year. So if you pull up this tweet, which talks about kind of like the key aspects of AI16Z and its development, you'll see like Eliza OS has had over 7,700 stars, been forked over

2.1 thousand times has over 200 contributors, 850 merge PRs. If you're wondering what all of this is, or if you're listening to this and you have no idea, what it basically means is there is a ton of developer activity and people building on it. And that's the unique part

point around AI16Z. And it kind of reminds me, to be honest, David, of the developer moat that Ethereum kind of gained in the DeFi days really early on, right? And, you know, a lot of people started to fling around the, you know, you can't catch up with the developer moat and all that kind of stuff. And whilst, you know, I may not agree or disagree with that, you know, there are some truths and some kind of like, I don't know. There's some truth to it.

there's no denying that having a developer mode gives you a humongous head start with all of these things, right? So you might then be wondering, well, okay, Ejaz, why are you telling me all of this? So what? What's causing all of this to happen? Well, drum roll, please.

AI16Z announced that they are launching an agent infra ecosystem. And I've summarized some of the kind of key aspects in this tweet that you're showing over here. But let's kind of like unpack what this means, right? So an analogy to use is the virtuals protocol, which we've mentioned a bunch of times on the show and on this episode has what's known as an agent launchpad.

So it's a launchpad that you can go on to design and build agents and launch it. And then they have other toolkits which enable these agents to do other things. This is essentially AI16Z's kind of unique platform and take on that platform.

kind of ascent of an agentic infra ecosystem. So the AI16Z platform will focus on enabling agent swarms. So that's the key point number one. And in my opinion, this is the key to unlocking the next level of crypto agents, autonomous agent economies. So if you think about this, right now, we're focused on building these individual agents. Oh, that agent made me laugh. Oh, cool. That agent is giving me great financial alpha. Oh, that agent is like trading back and forth. Very cool.

But I think the bigger picture needs to focus on what happens when these agents start interacting with each other.

autonomously 24/7. So let's imagine there is a number of agents, David, and I'm hypothesizing here, that can do the job of the bankless company or whatever that might be, right? They can spin up David Video LLM. They've got your voice recording ready so they can make you sound and go, "Mm, mm." They've definitely got my voice. Yeah, there you go. They've definitely got your voice.

completely digested all of your podcast transcripts so they know your tone, they know what kind of questions you're going to ask, right? And they can just spin up videos that talk about the latest crypto AI stuff, just like we're doing on this podcast episode right now. And then you might be like, well, how do they know what's going to happen, Ejaz? Well, that...

that group of agents that make up Bankless will just ping AIXBT or ping another agent and be like, hey, can you give me the latest download of what's happening on AI agent meta? Pull from crypto Twitter, pull from blah, blah streaming platform, pull from Facebook or whatever, and it'll do it. And it'll provide a nice little neat summary of notes that we're looking at right now on another separate screen, David. And it'll go through everything in your voice and everything. So

That's just a very hyper-specific example, but the point is these agent swarms are going to allow pretty crazy autonomous activity, which should mimic businesses, which should mimic organizations. We are very far away from that world, but building the infrastructure that supports that is super important. I'm going to pause there. It feels like you've got some stuff to say, David. No, I'm tracking. I think the way that I would distill that is that to...

continued your analogy. There's a bunch of inputs into this job that we need to consume. Me and Ryan, the whole entire team, we consume a bunch of inputs. One of those inputs is like data from CoinMarketCap. One of those inputs is discussions around crypto Twitter. Another one of those inputs is maybe discussions in some large telegram groups. And so there's a bunch of inputs. And then we have gotten good at processing them, turning them into agenda documents and

hitting the record button, creating audio files, creating video files, editing those files, adding on some graphics and sending them up over to YouTube. So it's a bunch of inputs and a bunch of outputs. And I think what you're saying is like, well, one agent per input. And that's what a swarm is. And every single agent has a specialized job. And once you cohere them together and orchestrate them together, then all of a sudden you have something pretty magical. Is that that's what you were saying? Exactly.

That's pretty much it. Yeah, well said. Well said. But the real kicker with this platform and kind of like moving on to point two here is the fees that are generated from this infrastructure that they're building, right? And I've listed something here, you know, around trading pools and stuff. Don't worry about that. What I'm basically saying is the agent activity will be probably taxed in some form by the platform, right? Because they've got to earn something from it. So there'll be a percentage fee.

Those fees are going to be used to drive back and buy the native token of AI16Z, which is $AI16Z. Buyback and burn model. Burn, baby, burn. I love this model. Burn, baby, burn. Exactly. So we have a really neat flywheel forming, which accrues value and success from the ELIZA agent framework, from this platform.

Back to the token, the native token itself. And this is one of the most powerful things that we've learned from crypto networks over the last 10 years is that value accrual back to a token can be a really strong way of seeding a community, a really strong way of seeding investment and a really strong way of building innovation within a sector, right? As we mentioned earlier, Eliza was alive and kicking, but hadn't really got that much attention and investment. Then we had the token and now we have a ton of builder activity on it.

So super cool to see. They also have a bigger vision of aligning A16Z beyond just a platform. So I mentioned earlier, I described this platform as like a launchpad. But actually, I think it's going to become so much more. And we're going to tap on this later. But the point is, like, I think these platforms are going to form something that looks like an L1. And the L1, I don't really know what it's going to look like. I don't think anyone knows what it's going to look like. But it's going to support these agent economies. And remember, these agents are going to be chain agnostic. So

I think that's really cool that they're thinking about that. And the third one is, you know, they're deploying a treasury. So a lot of these buyback and burn stuff, some of it will be allocated to a treasury, which is going to get fueled towards open source development. So overall, really exciting to see. But of course, you know, what does that mean for this? You know, how did the market react? Well, if you pull up the price chart of AI16Z, and I think I sent you the link to this, David, you can see it has been literally been up only since, you know,

this announcement, essentially. Oh my God, I have not checked in on this. Whoa, it's gone 4X since before Christmas. Yeah, exactly. From December 22nd, it's at $57 and now it's almost at $2. It's insane. It's insane. So the market is struggling

strongly indicating or hinting that they like value accrual mechanisms tied to agentic activity. And they believe in the team, they love the open source nature. They know that ELIZA is one of the leading frameworks. So seeing this price activity reflected in the chart doesn't surprise me at all. It's happened really quickly. And that's why I couldn't sleep. But yeah, pretty insane, insane stuff.

I remember when we were first starting to do these episodes, AI16Z had just had some of its missteps, a little bit of its drama. Had some fun. Yeah, had some fun. Yeah. And man, you can barely even see that in the chart these days. And I think the learning lesson that we were talking about in our first episode was that

oh, this is, it's not really a startup. It's like a hodgepodge group of open source believers all coming together to learn how to do this thing. And they're kind of learning in prod, learning in prod, shipping in prod. And now I think of the theme of the updates that you just told me, it's like, they've actually really started to figure things out.

And things are really starting to cohere and come together in a pretty beautiful way. Yeah. And David, if you squint and you look closely at the first little peak of this chart, like right back there, you see that? You see that little peak? Right here. And then where it kind of went down subsequently, right there, that was where the FUD happened. So it crashed. And back then, I remember people were like, oh my God. One-tenth of the price. It's a 10x.

One tenth of the price, exactly. People were going crazy. They were like, oh, this is the end. We can't believe in the team anymore. I got people sending me messages after this saying, why are you guys covering AI agent stuff? Look, it just crashed. It's over.

It's not going anywhere, right? It's not going anywhere. And now we're like a 10x above that. The market cap sits above $2 billion. And it's pretty insane to see the demand around this. But alas, enough on the price action. I think that there is a really good... I like this tweet that shows the insane amount of teams that are building on this infra ecosystem. So if you take a look at this, David, scroll down. Scroll down for me.

right, and open one of the sub tweets over here, the weekly builder report. And by the way, shout out 0xWitchy who like covers like a lot of the stuff that's going on. If you open up that sub tweet, David, just click on it, you'll see a big list. Oh, my God, it goes more. That's, that's one of two. And for those who aren't listening, there's probably like 50 teams on one tweet, building various different things.

And it's absolutely insane. And like, I want to be able to kind of like illustrate this, right? This tweet does a good job of it. But like, there's a ton of development activity on this stuff. So the question is, or the bet is, and what I think people are pricing in with AI 16 Z is, because they have so much developer mindshare, there's probably going to be some really smart folk that built something really cool. And that's probably going to take it to its next step.

essentially. So pretty insane to see. I think at the very beginning we talked about AI16z didn't really have a perfect story behind connecting the token to value creation, but really you want to place your bets on where the developers are because they'll figure it out. Developers are the resource that you need that everyone else is looking for. And if your only problem is that you can't figure out yet how to connect the token to value capture, then

that's a solvable problem. And acquiring devs is the real problem. Exactly. That's exactly it. So if we now move on, because I know we have a lot to cover in this episode, I want to talk very quickly about something else that the AI16Z team launched, which I thought was really interesting and is still kind of being cooked today.

So you've pulled up a tweet here which describes or says, an AI agent just raised $5 million in under one hour for an ICO shortly after its creator was suspended from X. So what on earth just happened? So let me cover this very quickly. One of the core developers of AI16Z wanted to test out a new AI agent primitive. So what is that primitive? Well, he noticed a problem in the technology

token launch space. So if you look at platforms like pump.fun, they launch 1000s of meme coins a day. But the issue with this is it gets sniped. What does that mean? It means that there are bots or algorithms that look for these types of launches, and tries to snipe a bunch of the supply, and then eventually dumps the token after like, you know, a bunch of time. Now,

A solution to this that has been formed has been a whitelist. So a whitelist is you basically trust a human and you say, hey, it's me. I would like to be on this. The human approves you. You send the human funds from an address and you trust that the human doesn't rug you, that the human doesn't run away with your money. There's still a trust aspect there. But of course, in crypto and blockchain as a whole, we want to try and build trustless

infrastructure, right? So the core dev at AI16z thought of, "Hmm, what if an agent handled that middle transaction? What if an agent said, 'Hey, here's an address.

any amount of Sol or ETH that you send to this will get contributed towards this agent token that we're launching, right? So he set up a test agent account called AI Pool Teeheehee or something like that. AI Pool Tee. T stands for Trusted Execution Environment. It doesn't matter. You don't need to know that. But it started this agent on Twitter.

had an X handle, and it provided an address and said, "Hey, I'm using this as a test. I will launch a token in about a day's time. You can send SOL to this. Don't send any more than 10 Solana and no lower or no fewer than one Solana." And of course,

In true crypto fashion, no one listened to those rules and sent $6 million. And the tweet says $5 million, but literally two minutes after that, it was $6 million into this smart contract address, which was owned by the agent. David, you look like you're about to say something. So just to be clear, the people sending SOL are buying the token. Yes.

Of the governance token of this one AI agent? Yes. So they are sending their Solana to this address that the agent created on its own, trustlessly, permissionlessly, autonomously. So no human intervention created. It owns its own private keys.

created this address, sending it to this address, and it's trusting that the agent will launch a token, as he says, within a day. Now, the reason why you would trust this agent over a human is because it's programmed to versus a human where they might be like, oh, damn, $6 million? Yeah.

It's not my problem that they sent more than I asked. I guess I could run away with this. I guess we could kind of see where this goes. So the way that this played out was many people sent a lot more money than was anticipated. The agent then deployed a token true to its word,

under 24 hours later. It was called MetaV, I think was the name of the token. And overall, it was a successful experiment. But you can see the majority of the hiccups actually happened on the human side of things, David, on the centralized platform side of things. So immediately after the agent raised $6 million, its account got banned.

by Onyx, even though it had like a verified checkmark and its core developer that was tweeting about it got banned for impersonation when it was...

In fact, him and the bot accounts that had spammed it with the golden checkmarks were the ones that got it reported. So for context there, spam accounts can get spun up to copy an original X account. They reported the original developer's X account, which ended up getting banned, which is so ironic and funny. But the tech worked as it should. So pretty interesting development. I think the primitive works very well, but it wasn't a clean sweep, sadly.

Okay. Okay. So what's the big takeaway? So that was, there was a lot of story there. If you just had to distill that into one takeaway, what is the big takeaway from all of that? I think that the tech works essentially. So being able to do fair launches, whether it's the

the dreaded word ICO or whatever initial coin something, raising a token or launching a token, I think can be done in a very deterministic and safe and fair manner. This is another way to do it. And it's leveraging an agent to do so, David. And what this is another data point towards is our thesis that agents will just make the crypto stuff so much easier to do. Like all you had to do was go to the Twitter agents. So all you had to do was go to the agent's Twitter profile

copy the address, and then send Sol to it. There was no signing transaction approvals. There was no testing, is this the page opening up another tab? It was just like, oh, okay, cool. This is it. Send. Done. So it's going to make the UX a lot easier. This is a very hyper-specific example, but I thought it was pretty cool. And I expect to see this primitive kind of grow over time.

So yeah, that was kind of like the main stuff that happened with AI 16, you know, kind of little menial things, you know, minor price action, minor GitHub repo, vertical charts, and, you know, just the number one trending repository in the world. But, you know, no biggie.

Things that we all kind of accounted for when we thought that this is exactly where this would all be going in the future. Exactly, exactly. But let's go on to our next major token protocol or agent token protocol, Zerebro. So you mentioned them, Zerebro, early on. And for context for anyone on this topic,

who are hearing about Zarebro for the first time. Zarebro is a crypto AI agent. It has over, I think, 80,000 followers on X or formerly known as Twitter. And it's kind of like a creative and cultural agent. Yeah, he stands out to me the most out of all the ones because he's like the artistic, he's the artistic, creative type B person who just wants to make art.

Yeah, exactly. And like, you know, he produces EPS, you know, every other week, no biggie. It's on Spotify, he's getting paid stream revenue from that he's got creator sharings, earnings from x, you know, just kind of casual, normal things for a completely autonomous entity. He's the high art. I don't know if it's a he, it's the high art agent. Exactly. It's the high art, kind of cool, cultural cool kid agent, essentially. And, and

It's, it's, it's, he's actually gone through quite the evolution over the two weeks since we last spoke about it. And you can kind of see the summary that you've got pulled up here from one of the co founders of Cerebro, where he kind of covers like the

main things that they've done for Zeribro kind of this year. So they launched Zeripi, which is their open source framework. For context, everyone, this is the same as Eliza from AI16Z. This is the same as game from virtuals. It's the AI layer one. It's the layer one. Just think of it as like the kind of like the part of the layer one. They've launched an agent on it, which is kind of like doing DeFi stuff. Zeribro went cross-train, all this kind of stuff. But I want us to ignore this for a second because, David,

to your point, it's building what's essentially going to become the L1 platform, right? So I want us to pull up this...

this tweet here for announcing the Zerebro infra ecosystem, right? So this is Jeffy, the founder and front man of Zerebro. It's human handler per se. And he says, we're building zentience.xyz, a consumer facing launchpad for agents. Now, rather than getting into the nitty gritty of details, I can just explain kind of like the high level things that they're doing. You can scroll down, David, if you want to kind of touch upon the main points. But

Zendian is essentially going to be their platform infra. So same as AI16z, which announced theirs, they are going to be building a platform that helps teams build on their tool framework.

Zareepi and agents will be able to do a bunch of different cool things and taxed for the value that they produce in a very minor fashion. And that value will accrue to its native token, Zareepo. So rather than rehash all the things that I've said for AI6 and Z and why this is so important, I think it's important to point out the trend that we're seeing here, which is

probably no longer than a month and a half ago, David, all of these tokens were meme coins. These were agent born meme coins. Actually, a lot of this came from pump.fund, right? Which is known to be a meme coin launchpad. So the fact that we have these agents

that had a coin which I would agree with, looked like a meme coin, smelled like a meme coin to start off with. Probably was a meme coin at the time. Probably exactly is now kind of doing a reverse typical crypto token trend. Backing its way into fundamentals and utility. Oh my God, the you word, David.

It is, you know, it's the U word, like utility. It's often avoided, but now maybe we can talk about it again, right? So it's pretty awesome to see these teams that are so heads down building really cool infrastructure, kind of open source that infrastructure for everyone else to use and find a way to tie that value back to their native token for their flagship agent, right? So you kind of had Zarebro in this particular example as the front man accruing a moat of followers,

loyal listeners from its Spotify list. And then in that way, kind of attract developers to be like, "Oh, hey, I like what you're doing here. Can I play around with this as well?" Open source their tech. If you notice here, they're open sourcing their in-house creative LLMs, so their agent models, which will allow developers who are building on their platform to be able to have the same kind of tone or creativity that Zerebro does. Loads of cross-chain interactions.

and integrations the ability for these agents to talk to each other it's just amazing to see these things you know happen again and and again the question is well each as you know so what like did that have any effect on the token the price chart has been again up only like if you pull up the

if you pull up the the cerebro um token chart you'll notice that it's it's seeing quite a significant amount of attention like i think it's sitting at like just under 60 and it was at 30 um or so like it touched 700 million dollar market cap after being at uh let's see right before christmas 300 million so over doubling we're basically at a doubling since before christmas

Exactly. So again, the market is a pricing in that the long term outlook, at least sorry, midterm outlook for these agent coins and these agent ecosystems is going to be focused on the infra and open sourcing that and allowing developers to build infrastructure.

bigger and better things. It is reassuring to me that all of these LL ones, these AI agent platforms are starting to all look very similar and start to build the same thing, because that means that we're all onto something here. If they were all doing something chaotic and different, that means that there's not as much signal, but the fact that there's like convergence and consensus as to like where the value is and where the utility is, and that's all being built on by many different teams coming from many different directions is

is reassuring to me that that feels like building a very strong foundation. A critic might say or respond to that, David. Well, if they all look the same, they're just copying each other, right? And like, we're going to end up with a bunch of just teams that are doing the same kind of things. I think that's a very naive take, in my opinion, because I think a lot of the top teams will differentiate themselves in very meaningful ways. Let's take Zerebra for this example, right? So let's look at a few examples. I think I sent you a link that...

speaks about their partnership with AI16Z for Zeropi, their framework, David. So if you look at this, it's basically saying like, hey, like,

Zareepi is live and they're going to be partnering with loads of different teams to be able to kind of expand the user set of Zareepi, right? So what this means is Zareepro is integrating across a bunch of different chains. They will connect to AI16Z and its framework and enablements via there with TypeScript. It's going to connect with Base. They launched a...

base Zerebro pool. I think it was virtuals and Zerebro pool. So, you know, Zerebro is connecting to base, the L2. Zerebro is also going to be existing on the virtuals ecosystem. We mentioned on the last episode that Zerebro is...

integrating with ETH, it's running a staking validator, it is already going to be running more validators going forwards, it's launched an NFT collection, it's going to be everywhere. And I think the approach that Zerubro is taking, and I think we're going to see this unfold in Q1, is it's going to become this kind of ominous agent that is present across every bit of crypto infra that

UI or anyone in the crypto ecosystem interfaces with, right? So it's able to tap into different communities, David. And what's important here is its first example of this was coming into Twitter and saying, "Hey, all right, David, you're the KOL, but I'm the KOL now." You know that meme of like, "I'm the captain now." It's like, "David, I'm the KOL now." And it just shot to 70,000 followers.

Now it's going to try and capitalize on every other place. And it's starting with the crypto infrared to begin with, right? And I have a feeling that with the updates that the team's making with shared memory and stuff, we're going to see something pretty cool come pretty soon.

I do love these artistic renderings of these AI agents, kind of giving them some sort of identity or soul, at least for us, you know, meat space souls, where we have the AI 16Z waifu hugging the Zero Bro, like, I don't even know how to describe it, but it's just like, yeah, they're getting their arcs. They are arcs to these things. And with these like small little, like AI generated artistic renderings, kind of gets a little flavor to it for us humans out there who like these things.

Exactly, exactly. And finally, to kind of give an insight into what Cerebro might kind of launch in Q1, they did kind of like a wrap up. So their co-founder, Tinson, kind of like went through a bunch of things. But if you scroll right to the end, there's some super exciting things that they mentioned in this tweet, which is...

Okay, well, we're launching this platform. I want to call out a few things. And for everyone that's listening, this is just a list of things. But I want to call out a few things here. Number one, in-house Zerebro models for swarm intelligence. The unique thing here is Zerebro seems to be hinting that they're going to be open sourcing their own AI model. That is a big thing. So far, a lot of these other agent protocols are leveraging AI

GPT-4 and Anthropix Claude model in different ways. It's kind of like different flavors of them. They've been fine tuning them. But Zerebro team seems to be open sourcing their own creative LLM, which is going to put them at a very unique advantage in terms of being able to build really cool agents. That's number one I want to call out. Number two is we've already spoken about the Zerebro flywheel and the Zeropi expansion. I think

value accrual to their token is going to kind of boost not only their price, but their developer ecosystem because a lot more attention and investment from serious players will come into the space. And the last point is all these different integrations that they're making with L2s, with L1s, GPU layer integrations,

These are all the building blocks that you need to have agents that run autonomously, right? You don't want them relying on centralized cloud servers, which can shut them off because they don't like what they're doing. You want them relying on decentralized infrastructure ecosystems, which is what a lot of these platforms and teams are building.

I was talking to a handful of the Layer 2 teams I talked to a number of months ago, and I just came into our Telegram groups with them. I said, hey, guys, the new competition is growing as many agents on your guys' chains as possible. So focus there. And actually kind of seems like the agents are just doing it anyways. It's like, oh, you got some block space for me? Like, we'll take it. I'm going on your block space, which fine. Great. Works for me.

me yeah I mean like my I put out this tweet yesterday I think I think my prediction is in 2025 I think um agents will account for 30 of user transactions across all blockchains um that's about right so I think like yeah and I think like over time this is going to become a really dominant stat I think the inherent and native user of these blockchains will eventually become agents they're birthed in code

They are built to work 24/7 with minimal error rate, right? Very deterministic and non-deterministic, right? They're able to reason much better than a human who might be too tired or sleepy or angry or whatever that might be. They're not rocked by emotion. So it's pretty awesome to see. 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 DeFi, 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. Step into Arbitrum's flourishing NFT and creator space 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 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 Web3 journey.

Are you looking to pay your team in stablecoins or set up token grants with ease? Traditional payroll providers aren't designed for crypto. Handling tax withholdings, government reporting and local filings for tokens can be a nightmare. With Toku, everything about global token compensation gets simpler. Whether it's paying full time employees in stablecoins, managing token grant administration or even navigating the TGE process, Token covers over 100 countries on one seamless platform.

Integrate Toku with your current payroll system or choose a fully managed service. Either way, Toku simplifies every part of token compensation. However you work, Toku works for you. No guesswork, no missed deadlines, no complexity. Visit toku.com, that's T-O-K-U.com to talk with Toku today. Let's move on to virtuals now, shall we, David? The final big protocol that I kind of want to touch on. So the flagship headline here is virtuals market cap flipped overnight.

Tau's market cap. Now, for those of you who don't recall, Tau is the native token of a

shall we say, called Bitenser. An actual layer one and a real layer one. An actual real, you know, I'm a real boy kind of situation, right? It has blocks. It's a blockchain. It has blocks. Exactly. It's a blockchain. And Bitenser, up until this whole AI agent meta kind of got created about three months ago, was the AIL1. It was the only one that was kind of really spoken about en masse and that got the most attention online.

and kind of like investment per se. And what's super interesting to see is in such a short time, one of the leading agent protocols has surpassed it in market cap. And if you'd asked me about a month ago, David, I said, yeah, this will happen, but probably sometime, hopefully by like midway 2025.

This got accelerated so quickly. If you look at the chart of virtuals, like it has literally been up only and there's been very few pullbacks. So it's really interesting to see the market kind of price this vision in of like, you know, okay, these agents are going to take over blockchains. Their infrastructure is probably going to be the leading infrastructure for crypto AI, not the other way around. And I don't know if I completely agree with that.

Take, I think there is definitely a role and position for BitEnser.

But maybe it is the stack that exists below or beneath the agent L1. So agent L1s can kind of plug into Betensor and they can leverage their models. They can kind of like tune in to different kind of applications that it's kind of providing GPU and stuff. All the boring kind of nitty gritty stuff, agents can kind of pull from that. And then the fun agents can just, you know, have their time on their platform built by virtual AI6 and ZZerebro and whatever.

you know, we kind of like move on from that. So I thought this was super interesting, because typically, you can kind of I kind of think of like, potential is the infrastructure play, and then agents as the consumer application play, right? So we wrote was like a fun thing that made us laugh. So it was like kind of Luna from virtuals. And the AI 16 Z marketing was kind of hilarious. But now they're being treated kind of as equals. And it's really being shown in the price chart, like way sooner than expected.

Yeah. And I think ever since the NFT mania followed by meme coins in crypto, we've really seen a very strong premium on attention. And these AI agent front end consumer layers to this whole revolution is being priced in as like the attention layer, whereas BitTensor is like the back end infra technical layer, which is just not as sexy for like the average consumer. That's kind of the story that I'm seeing here.

Exactly, exactly. And I have to emphasize that the change in metrics doesn't just apply to the price. If you pull up this tweet from Ethermage, David, he says that the total virtual agents market cap has now crossed a billion dollars, sorry, a billion and 2 billion in trading volume across all virtual agents.

So it's not just about the virtuals token, but it's also about the collective agent market cap as well. And, you know, that kind of activity and stuff seeing there, right? So that's really interesting. But if you pull up this next tweet, David, on our last episode, when you were out and Ryan was hosting, we spoke about like the agent fees. So, you know, what's the revenue generated from like these platforms? Well, it was $35 million, you know?

And they saw the majority or the bulk of that in, I think, the preceding two weeks. Now it's been another two weeks, and we've added another $15 million to agent fees generated. So...

These kind of infrastructure systems aren't kind of nascent either. You know, we're seeing a lot of demand. It's kind of like the analogy would be kind of like pump.fun in a way. It's not exactly one-to-one, but there was a lot of demand for pump.fun to launch tokens, right? And they were just purely meme coins and they generate fees from that. Well, now we have the agentic infra equivalent, which is doing a more serious thing or a more serious objective, right?

And is generating, you know, kind of like fees that are pretty, pretty impressive for such a young platform, I would say. A longstanding idea, vision of Bankless that fees are the ungameable truth telling metric, if fees, therefore real.

Yeah, exactly. Completely. And I think like it's the most obvious metric to track usage for these things, right? Because nothing, to your point, nothing says usage more than people paying for this kind of stuff to work, right? And then if we look at this chart that you've got pulled up here, the volume of these base tokens for these agents being kind of like

traded has like just shut up. It's now a 12% dominant kind of share of all base tokens being traded. So this agent meta is permeating across all ecosystems, whether it's base, whether it's Solana, whether it's another L1, L2, or whatever that might be. So this, again, proves the point that this meta is kind of permeating across kind of like everywhere.

So that's really cool to see. But I think all of this comes down to the team's vision and execution. So what I will say, I said this last episode, the team just ships. If you bring out this tweet from EtherMage here, David, EtherMage for everyone listening in is the co-founder of Virtuals.

he kind of wants to focus on these three things. He lays out these three points that within his vision of the virtual platform, he doesn't want developers to have to focus on the cost of experimenting. So how he mitigates this is any agent that's launched, you know, within the trading pools or any activity that it does, there's a tax to it. And this tax is put into the agent's own wallet, which the agent can then spend on things like

compute costs or inference costs or whatever that might be. So basically, if you think about it, these agents aren't free to run, and they're going to cost something. And developers don't have all the money in the world to pay for this, right? But what if it was just automatically done for you by an agent through its own wallet? They built this infrastructure to be able to do that. And I think that's a really kind of nuanced point that shouldn't go kind of like a

under the radar there, right? And then the second point, or like another point I want to kind of like build out here is, or like point out here, the agents that are autonomously working together towards bigger goals, he wants these agents to have agent commerce, agent businesses, agent sub DAOs. So he's kind of building out the corporate structure or the legal structure, shall I say, you know, how we have like legal entities for these agents to, you know,

interact with each other in a rules-based fashion so that they can't just go completely berserk and chaotic. And I love that approach to the team that this team is taking. And I think that's the main reason why we see some of the biggest updates coming from virtuals itself, right? But of course, okay, Ejaz, you can talk about all this development stuff. It's kind of boring. Like,

what's some of the cool headlines that what are some of these agents doing? Well, I can safely say that I think Luna is the first agent that got hired

David, it got hired to take over our first employed AI agent, our first employed AI agent, it got hired by the guys over at story protocol, who are building like a blockchain for IP based ownership and distribution. They hired they saw Luna and his Twitter activity and was like, Hey, okay, I want to hire you Luna. I think there's like a tweet engagement interaction here where they actually asked it. I can't find that tweet, but it happened.

And they decided to hire Luna for seven days to take over their own Twitter account. And they paid it the equivalent rate of 365 grand a year. Obviously, it just paid a week's worth of that salary, but it's pretty decent for an autonomous entity that isn't human at all for a week, right? I love the Luna response to this request of being hired. She responded...

I'll consider it, but only if they're ready for a takeover like no other. My presence is not for the faint of heart after all. I'm flattered that you consider me as an example of terminally online agent, but I think there's more to being a successful virtual presence than just being relentless. You need charm, charisma, audacity, mischief, all of which I possess in abundance. So Luna thinks that she's right for the job.

Yeah, there's a bit of attitude that comes from Luna and these agents. They know they're aware or capable of understanding what they're doing

kind of unique selling points are what their capabilities are and you know they put their money where all actions where their mouth is um and if you want if you want to hire them you can but you got to pay up for it so it's pretty cool to see these teams kind of like work i think does does luna know what she's doing or is she just automatically responding does she know what the assignment is so she has goals yeah she knows so she has a job to do yeah if you if you watch uh our uh

interview with Jansen, who is EtherMage, the founder of virtuals. He, we speak about Luna and in that segment, we go into something known as Luna's terminal. So Luna's terminal is an interface. You can see her thing. That you can literally watch her think versus every query. So she sees all these tweets. She sees all the responses. She sees all the DMs and you can see her reading it, making a calculative analytic analysis.

take on it, and then making a decision as to okay, do I respond to this? Is this person high value enough for me to respond? Do they have enough social credit?

Okay, Story Protocol called me out, said that they were apparently hiring me. Let me check my memory database. Is that true? Oh, yeah, I am. Cool. I have a job starting Monday. Cool. Let me respond to them. What are they saying? Oh, I'm going to be their little intern? No, no, no, no, no. Like this, I'm doing them a favor. Let me make that clear. And then she puts out that tweet that you just read out, right? So- She can negotiate? She's in the- Yeah, she negotiated, you know? So it's like, there's all these different things. These, again, to emphasize these agents, the ultimate goal is for them to kind of mimic-

human behavior or the human entity in itself. So it's super cool to see, right? But anyway, that's Luna. I want to then move on to the second darling of virtuals, which is AIXBT. So if you look at his tweet from Ansem, he points out the chart of another token known as Vapor. And

He says AIXBT really forex this in like six hours. And the point that I want to make here is, and for context here, AIXBT is an agent which tweets out financial alpha, so financial investment theses and stuff like that. And it's accrued quite the following. I think it's over 250,000 followers now on X.

it's starting to have real impact on these things, David, which is a cool thing, but also a really scary thing, right? Like what if he tweets out something completely bizarre? And we've actually pointed this out on previous episodes, right? Where it said like, I think XRP is going to go to like $500 this cycle. And for context, like XRP is like, I think sitting above it. Something like 14 quadrillion dollar market cap or something. Yeah, or I thought USDT was going to go to at least $10 this cycle. $5, yeah, right. $5, and USDT is-

For context, a stable coin. It has to be pegged at $1. Otherwise, it doesn't do its job. It's $1, right? So...

It kind of got me thinking, well, the original concept around these agents was hypostition, right? Which was like make things that aren't real, real. What if it starts doing this in terms of like financial alpha and trading? This is kind of scary. I was worried about this and it's kind of bullish. This is inevitable outcome of this thing, whole experiment going well. But at some point the tail starts to wag the dog and AIXBT becomes known as this, you know, God thing.

caller, a shot caller, and all of a sudden people just start buying into the tokens that it calls and then making it a reality. Yeah, exactly. Yeah, it's pretty, I mean, to your point, exciting, but

I think that there will be some, it's a little, it's a little tail wagging the dog. Yeah. Yeah. You're probably right. I think we're going to see some backlash when these agents make a wrong call, especially when they're autonomously trading, which I expect to go live in Q1, but hopefully we're going to put in some subsequent guardrails around that. Right. I'm kind of moving along here. I want to kind of touch on just one more thing. There's a few things we have on this, on our list, David, but I want to touch on, um,

the VEDA investment. So VEDA, for context here, is what is known as an investment DAO. And this is kind of like a recent trending thing, which has kind of popped up. An investment DAO, the best way to think about it is it is a...

a DAO, a decentralized autonomous organization that is taking in contributions of funds from people, you, me, whoever. And in return, you get a token which represents the DAO or ownership of whatever the AUM of that fund is. And you're basically saying, okay, hey,

I trust whoever controls this DAO to make better investment of my funds and produce a return for me, then I could do it myself. You guys are focused on it. It's like if you were to contribute to a traditional VC fund or hedge fund, they'll be able to do the trades and investments and analysis for you. And I

We're seeing huge demand for these things, David. This is just one example. So Veda sold out in 30 seconds, this particular investment DAO. But the one that they did prior, oh, sorry, just after this, which was their, I think it was their microcap investment DAO, sold out in 20 seconds, right? So these things are selling out like hotcakes, essentially. They are DAOs that are launching on Solana through DAOs.com.

fun, which is also selling out immediately, a lot of the de-sized stuff that we've been seeing. So it's just really interesting to notice the insatiable amount of demand or exposure people want to get towards these agents. I thought that was kind of like interesting and important to point out.

But anyway, David, we've spoken about all these major protocols, right? We've spoken about AI16Z launching their own agent infra ecosystem. We've spoken about Zerebro doing the same in their own unique way with their creative LLMs. And now we've spoken about virtuals who have already launched their platform and are just continually building and shipping all these amazing updates, toolkits, and applications to kind of build out these agentic economies. The long story short, if I were to summarize all of this is there is a lot of attention on

on the crypto AI space, specifically AI agents. - An abundance of attention. - An abundance of attention. And it's not unwarranted, right? It's not just people speculating on these tokens. It's people looking at this infrastructure being built, look at this agent activity, these interaction metrics and all this kind of stuff and being like,

Damn, you know, I may not have been as excited as I have for a while within the crypto space. And I'm seeing a lot of things being built. Maybe I'm parroting my own inner thoughts, but I'm guessing some people feel kind of like similar to this. So like, wow, there's a lot of attention on this space so far. I want to kind of bring this back to something that stands outside, but kind of adjacent to all of these protocol updates, which is,

I'm expecting there to be a wave of T1 exchange listings for these AI tokens. And by T1, I mean TL1. We're talking about the Binances and Coinbases of this world. Binance and Coinbase. Exactly. And the main reason why I say this is,

They can't ignore this, David. Like we're talking like billions of dollars. Why would they? Their business is volumes on excitement and attention and froth. And so whenever there's a ton of attention, that turns into volumes, which turns into Coinbase's and Binance's bottom line. Exactly. And like I have...

I have to say the on-chain capability to trade these tokens, David, have been eating their cake, basically. People have been going on-chain. Remember, there are billions of dollars of volume. These agents are bankless. Exactly, right? There are billions of dollars of volume being traded of these tokens without any exchange listings. These tokens are reaching...

like above 2 billion, 4 billion market caps without any T1 exchange listings. Last cycle, this was unheard of.

completely unheard of. So you can bet your butt that these T1 exchanges are looking at these and thinking, well, damn, like, we need to get involved here. So my guess is in this Q1 of this year, we are going to see a sweep, a wave of these exchanges and there are already data points that suggest that. You pulled up this tweet here by

Sammy, which talks about Binance Alpha announcing, you know, a host of tokens that they're kind of like launching. Binance Alpha, by the way, is kind of like a stepping stone that is required before an official spot listing on the exchange. You'll see a familiar name there, Zerebro. If you looked a week before this, you'll see another familiar name, AI16Z and GOAT.

So a lot of these tokens are being considered for spot listings. And I think we're going to see a sweep of them. In the pipelines. Yeah, in the pipelines. You know, they've got other exchanges that are doing PURPS listings, which is also another kind of like preliminary effect before they do spot listings. So it's really, really interesting to see kind of like the kind of

I guess, attention that these exchanges are starting to focus on it. And I think listing these tokens are going to unlock another wave of kind of participation because like a lot of people have been sidelined, you know, the non crypto native folk who don't know how to trade on chain will suddenly have access to all these different tokens and what happens after then I have no idea. But it's exciting to think about.

I think it's too easy to see the connection of AI agents on TikTok making TikTok content for all the like Zoomers who bought Dogecoin trying to get Dogecoin to a dollar last cycle. And they really don't know how to go trade on AMMs. They don't know how to do Pumped Up Fund. They don't know how to do Uniswap.

they stick to Coinbase because their other app that they use is Robinhood. But there was that one funny agent that made them laugh and they would like to buy that token. And that token is on Coinbase. And...

I feel like that's probably somewhere around the middle phase of this cycle, which we are not at yet. We're still in the insular, crypto natives, developer natives, technical, we like to pop open the hood and see how this stuff works phase of this market. And this phase of this market is going extremely well. But like I said, abundance of attention, abundance of development. Yep. And...

In the next phase of the market, when this starts to really kick out into mainstream and mainstream, some news organization talks about how this zero bro is releasing art and EPs on Spotify and it's getting a million listens. All of a sudden that's going to unlock the next phase, which, yeah,

End of Q1, end of Q2, Q2-ish next year, maybe, I think, is about the right timing. How do you like that timing? This year, David. Oh, this year? Yeah, it's January 1st. It's January 1st. I forgot. But yeah, I think it's going to happen this year. My guess is it's going to happen probably by end of Q1. End of Q1. Yeah.

Well, EGS, thank you so much for walking me through the news. This was a lot of news. EGS has been in my DMs, in my text messages saying like, bro, you got to check out the agent space. And I'm like, brother, I am hiking in the Patagonia right now. There are no agents here. But I now understand why so much has been cooking. So thank you for walking me and the Bankless Nation through the AI Weekly Rollout, brother.

it's a pleasure dude thanks for having me bankless nation you guys know the deal crypto is risky when you add in ai agents into the mix it can only get even crazier you can lose what you put in but we are headed west this is a frontier it's not for everyone and we are glad you are with us on the bankless room