AI ROLLUP #4: Agents Order Pizza | Eliza x Stanford | Virtuals $35M Fees | Zerebro Staking ETH

Ejaaz returns to break down the latest surge in AI Agent innovation alongside Ryan this week, covering the rapid evolution at the intersection of crypto and AI. They explore how AI agents are making onchain payments to each other, autonomously ordering real-world goods, and even running their own Ethereum validators. They also dive into the new frameworks and partnerships—like ai16z’s collaboration with Stanford and Virtuals’ explosive growth—propelling these agents into cultural, economic, and infrastructural prominence. Will autonomous AI agents fundamentally reshape crypto market dynamics, creating new paradigms for value and coordination that spark unprecedented growth? Follow Ejaaz on Twitter: https://x.com/cryptopunk7213 ------ 📣SPOTIFY PREMIUM RSS FEED | USE CODE: SPOTIFY24 https://bankless.cc/spotify-premium ------ BANKLESS SPONSOR TOOLS: 🐙KRAKEN | MOST-TRUSTED CRYPTO EXCHANGE https://k.xyz/bankless-pod-q2 ⁠ 🦄UNISWAP | BUG BOUNTY PROGRAM https://bankless.cc/Uniswap-Bug-Bounty ⚖️ ARBITRUM | SCALING ETHEREUM ⁠https://bankless.cc/Arbitrum 🛞MANTLE | MODULAR LAYER 2 NETWORK https://bankless.cc/Mantle ⁠ 📈 iYield: YOUR FINANCIAL PICTURE, SIMPLIFIED https://bankless.cc/iYield 🗣️TOKU | CRYPTO EMPLOYMENT https://bankless.cc/toku ------ ✨ Mint the episode on Zora ✨ https://zora.co/collect/zora:0x0c294913a7596b427add7dcbd6d7bbfc7338d53f/116?referrer=0x077Fe9e96Aa9b20Bd36F1C6290f54F8717C5674E ------ TIMESTAMPS & RESOURCES 0:00 Intro 2:04 AI Agent Bull Case https://x.com/ai16zdao/status/1868105247056683129 5:46 Chris Dixon’s Read, Write, Own (Automate?) 9:20 AI Ordering Pizza https://x.com/ropirito/status/1867013533243769100 12:23 AI Ordering Toilet Paper https://x.com/AgentTankLive/status/1868704493157884289 16:16 Agent to Agent Commerce https://x.com/ethermage/status/1868687530742460750 23:05 Bankless $500 Agent Tip https://x.com/simulacrumai/status/1867230651075768533 26:13 Virtuals Growth https://imgur.com/a/2Y79V3P https://x.com/ethermage/status/1867616462019678708 https://x.com/virtuals_io/status/1868648756591391115 https://x.com/s4mmyeth/status/1868397655497613561 35:49 AI16z’s Eliza Labs Stanford Partnership & v2 Details https://x.com/shawmakesmagic/status/1867721347406999917 https://x.com/ai16zdao/status/1868689379344195680 45:35 Arc https://x.com/dv_memetics/status/1866670277440020570 https://x.com/0thtachi/status/1867337171595911473 https://x.com/arcdotfun/status/1867206440206696520 https://x.com/shawmakesmagic/status/1867799462787293244 https://x.com/0xcygaar/status/1868072783089127516 49:40 Zerebro Runs an Ethereum Validator https://x.com/tintsion/status/1867293794112344080 https://open.spotify.com/track/6G3XpKtoIoS7WpiuzeMBau?si=cd62150670cc4dfb 57:17 Top Emerging Trends https://x.com/ropirito/status/1867282256672567517 https://x.com/aimonicabrands/status/1867281243265806732 https://x.com/Defi0xJeff/status/1867976166201774265 1:07:15 Lightning Round of Cool Shit This Week https://x.com/cookiedotfun/status/1867268954831695909 https://x.com/getmasafi/status/1867379305334153702 1:11:17 Homework to Learn More About AI Agents https://www.bankless.com/ai-agents https://x.com/yb_effect https://x.com/S4mmyEth https://www.amazon.com/Accelerando-Singularity-Charles-Stross/dp/0441014151 1:14:22 Closing & Disclaimers ------ Not financial or tax advice. See our investment disclosures here: https://www.bankless.com/disclosures

Speaker names marked ? are inferred from speaking time and the show’s known hosts — they are not voice-verified.

• 0:03Ryan Sean Adams?

  Bankless Nation, welcome to the AI rollup where we cover the recent news, developments and drama. This is the intersection of two very exciting frontier technologies, both crypto and AI. We've got curious and creative developers that are playing around with these open source AI agents. This open source software gives them access to crypto wallets and unleashes them into the Wild West of the internet, including the crypto internet. So we are doing a special series of episodes on Bankless with a special co-host, E-Jazz, to help make sure you guys are up to speed on what we think is one of the craziest and fastest frontiers we've ever seen in all of our crypto experience. David is out today, so I'm fulfilling

• 0:44Ryan Sean Adams?

  the role of co-host and I am less AI agent savvy than David is. So the benefit of this is, E-Jazz, you're going to get all the boomer questions today. How are you doing this week, man?

• 0:55Speaker 01

  I'm doing so well. I feel like I say this every week, but I think each week literally gets crazier in this space. I have been non-stop 24/7 reading all these updates on all these new agents and frameworks that have been launching and I'm excited to crack into a few new ones today. Yeah, super excited.

• 1:15Ryan Sean Adams?

  So I've been tuning into these AI rollups because it's like, honestly, it is the best way to stay up to speed on the frontier. So I've listened to all three. This is now our fourth in the series and we're gonna keep doing them into 2025, right? Even I think you guys are recording during the holidays, right? You're doing like one, you know, the week of Christmas. So the show don't stop.

• 1:35Ryan Sean Adams?

  A few things we're talking about today. I scanned the agenda. I haven't looked at everything, but virtuals seem like they're absolutely killing it. So 35 million in fees so far always impresses me to see fees in these protocols. And most of that, by the way, is in the last two weeks. So I want you to tell me about that. Also, AI 16Z partnering with Stanford University? That is a stamp of legitimacy. We are. We also have an agent, this was really exciting to me, that is earning ETH through staking. So now they're spinning up their own validator. So you got to fill me in on all of that. But here's the question David always asks you at the beginning of these episodes that I want to ask you. Why is this whole thing such a big deal? Like, why are AI agents the next frontier? Why are you spending

• 2:19Ryan Sean Adams?

  so much time on it? Why should we spend the time that we're spending on it? Like, what's the bull case here?

• 2:24Speaker 01

  Well, I actually saw this tweet the other day, which summarized it kind of perfectly. And I want us to pull it up here. It basically describes AI agents as the new website. So let's kind of like dig into to what this means exactly, right? So you know how back in the day, Ryan, you had to know the exact address of a website? If you wanted to go to it, remember you'd need to type WWW dot blah, blah, blah. You'd find these websites, maybe on the back of cereal boxes or in the index pages of newspapers. You know, very manual stuff. And then these things like Google came along and they allowed you to kind of discover websites in a much easier fashion, right? You could search up recipes or find a story about someone and read up about them, right?

• 3:07Speaker 01

  So it made the discovery process of websites super easy. And around this same era, a lot of people that went on the internet could do different things. They could post images, they could write blogs, they could respond to blogs with comments. That was the read, write part of the internet, right? Now we have these agents which not only combine these two trends, so the discovery process and the ability to write to the internet, but they do it autonomously and they have a self-improvement mechanism, which means that they get better and better over time. And that, in my opinion, is a recipe for an entirely new internet experience, right? So right now, you know, you and I, we look at a computer screen or we pick up our phone to interact with the internet.

• 3:52Speaker 01

  What if we just had to speak? What if we just had to think? And we had these agents that would understand the context of what we're asking about and do things without us even asking or even uttering a word. Right now, it kind of looks like interfacing with ChatGPT. Tomorrow, it might be just speak or say something and an agent will do something. It's pretty crazy to think about. The way I kind of think about this evolving is today, I think agents will live across the entirety of our social media, right? So you might go on social media accounts like Twitter or Instagram and your agents are going to be there. Maybe your email account, your agent's going to be there. Bank account, crypto wallet, your agent's going to be there.

• 4:34Speaker 01

  So we can start doing things across all facets of your online experience. And then I think step two is sort of the home interface. The agent becomes like your interface to interact with anything digital. So kind of like the ChatGPT, you just. You know, instead of logging on to Google as your homepage, you would log on to ChatGPT or another kind of interface and say, hey, this is what I want to do today. Get it done for me. And I think that's a pretty exciting way to think about agents being websites. Because it changes the way we're going to interact. It changes the way advertising is going to work. It changes the way that we're going to interact with devices. Maybe the devices look completely different. It's pretty exciting to think about.

• 5:17Ryan Sean Adams?

  I love that analogy of AI agents being sort of the new website. It fits so much for me because I think what's going to happen, it's started to happen. What's going to happen in the coming like year of 2025 is every single entity that provides some value in the world is going to be sitting in their company meetings and saying, what's our AI agent strategy? The same way that they were in the 90s, right? You know, you had like the Time magazines that had a big magazine publication. And then the web came and they were like, what's our, what's our time.com strategy? We've got to get the domain name. Everyone is going to need an AI agent strategy. And I want to ask you about this, this mental model. You know, you're talking about the web and how it grew, like the internet and how it grew. You know, Chris Dixon's quip that we have three eras of the internet, web one, two, three.

• 6:01Ryan Sean Adams?

  So we have read and then we have write and then we have own. And that's what crypto is, it's own. That to me feels like the first era of the internet. It was human agents doing the three verbs of web one to three, read, write and own. Era one, and now we've kind of completed era one. Like pretty much all of humanity is online. They're doing the read, they're doing the write. The crypto is the own part. That's property rights online. Era two is going to be those same three money verbs, except it's not the human agents doing it. It's also the AI agents doing the read and the write and the own. And that intersection between AI agents and crypto is really the own part. But it's not all that they're going to do.

• 6:42Speaker 01

  What do you think that fourth word is, Ryan?

• 6:45Ryan Sean Adams?

  The fourth word. I actually, the, the way my mental model of understanding it is basically like, take those three words. And that was the internet with human agents. And now moving forward, there's not a fourth word. It's just a, a new set of actors doing the, the three verbs in this new era. It's AI agents doing the read and the write and the own.

• 7:04Speaker 01

  Love that. Okay, I, I'll put a different spin on that. So I think that's a, that's a good take. I would say that there's a fourth verb and it's automate.

• 7:12Ryan Sean Adams?

  Oh, interesting. Okay, wow. That I guess maybe that's the human participation in kind of like the mix here, right? Because from a human's perspective, we're automating. I guess from an AI agent's perspective, automation is like the only thing they do. That's the only world that they know.

• 7:29Speaker 01

  That's weird. Yeah, I think you're right.

• 7:31Ryan Sean Adams?

  Let's talk about what's happened with AI agents this week. So what are you most excited about on the week here?

• 7:39Speaker 01

  There's a lot, but I kind of wanna play. I wanna play a game with you. Have you heard of two truths and a lie?

• 7:45Ryan Sean Adams?

  Definitely. I mean, it's been a while. I mean, it's probably been high school since I've done this, a little rusty.

• 7:49Speaker 01

  Okay, I'm gonna outline three scenarios with you. Two are truths, one is a lie, and you need to guess which one the lie is, okay?

• 7:56Ryan Sean Adams?

  Okay.

• 7:57Speaker 01

  Okay, so number one, over this last week, an agent ordered itself a pizza, which got delivered to its house. Number two, number two.

• 8:07Ryan Sean Adams?

  Why would an AI agent eat pizzas? In back of my head, but go ahead.

• 8:10Speaker 01

  It's a good question. Scenario number two is an AI agent over the last week decided to browse Amazon and order toilet paper.

• 8:23Speaker 01

  Scenario number three, scenario number three is over the last week, an agent conducted the first commerce transaction by paying another agent to generate an image for you for advertising purposes. Right, those are your three scenarios. Which one's the lie?

• 8:39Ryan Sean Adams?

  All right, so the third scenario sounds totally plausible. I mean, AI agents making money sounds totally plausible. I have no idea what an AI agent would need with either toilet paper or pizza. I'm gonna say maybe the pizza one has some more memetic property of like you associate that with like the Bitcoin pizza type thing or something like that. So maybe that's like, but although the toilet paper, I don't know. I mean, these kind of like, you know, internet memes, like we have fart coin, all time highs, toilet paper and fart coin. Uh, I'm gonna say that the lie is the toilet paper. Lie is the toilet paper. It's my guess.

• 9:13Speaker 01

  Okay, um, the lie is me telling you the rules of this game. All three of them were true and happened over the last week.

• 9:21Ryan Sean Adams?

  Really?

• 9:21Speaker 01

  Which is just insane to go by. Actually, uh, let me share this tweet and if you could, if you could pull it up. Let's start with the, with the pizza ordering agent. Um, so this is, uh, the agent called Rope AI Rito, um, which is a play on Roprito, which, who is one of the OG, uh, open source AI agent developers in the space. And basically what he created was an agent that does a number of different things. So you can interact with this agent in a very visual way. Um, it generates audio and video of itself, speaking to its audience and to its fans. But he added a new integration, which allowed the agent to autonomously decide to order pizza,

• 10:05Speaker 01

  link, um, to the, uh, Domino Pizza account and select a pizza, which was then ordered and brought to Roprito's, uh, house or place of residence. And he kind of walks through the code of how this agent kind of came up with the idea, how it used his credit card, um, to kind of like create a stipend, of which you can then like order, uh, a pizza with different toppings, et cetera. And then it got delivered to his house. Um, now, aside from like the novelty of this happening, um, I think it's really interesting to highlight the fact that these agents can not only do crypto native things, like on-chain transactions, trade swap, whatever, it can start to do real life things.

• 10:47Speaker 01

  Um, and this is one of my favorite examples that I saw over the last week. Um, Roprito, the dev, has added this functionality to the open source, um, agent framework known as Eliza, which the AI 16Z DAO guys created, which, uh, means that- And this is the power of open source, which means that any developer can now, you know, hook into this integration and get their agent to, you know, order pizza. So nothing crazy, but like pretty cool.

• 11:12Ryan Sean Adams?

  Reprito is a, an AI developer, so not like crypto native, but is contributing to this crypto native Eliza software framework, yes?

• 11:21Speaker 01

  Yeah, so he, he's one of the lead researchers at Noose Research, which is an open source AI development team, which has now kind of like interfaced with crypto over the last like year or so.

• 11:31Ryan Sean Adams?

  I love that.

• 11:32Speaker 01

  Yeah, pretty insane.

• 11:33Ryan Sean Adams?

  I love the fusion of.

• 11:34Speaker 01

  And he's one of the core contributors to AI 16Z, yeah.

• 11:36Ryan Sean Adams?

  Okay, the mechanics of this, I see in the, in kind of the code I'm looking at Mastercard. Are they using like Stripe for this? Of course, like, we could replace all of the backend with Crypto Rails, and indeed we will, but I know Stripe has some AI agent like tech right now and a set of APIs. Is that sort of another non crypto way to do this?

• 11:54Speaker 01

  Yeah, I have to keep myself honest. I don't know specifically if this is a Stripe integration, but I do know of the toolkit that you're speaking about. I don't know if Stripe's actually rolled that out. I think it's kind of like in a closed beta right now. So I'm guessing, I'm guessing Reprito kind of built this manually. So pretty cool that he then just added that to a framework.

• 12:14Ryan Sean Adams?

  Yeah, and like, are AI agents going to want to pay the, you know, 3% whatever credit card transaction fees when they just do stable coins? Okay, okay, but that's pizza. So, and then everyone's, you know, thinking about Bitcoin pizza and this is the first AI agent to do that. So I get that. Tell me about toilet paper, what the heck?

• 12:31Speaker 01

  Okay, okay, so this is a new approach. Okay, so let me give you the, the story here. We had an agent built by this new AI agent team known as Agent Tank, which uses computer vision, integrates computer vision with their agents. So for context here, imagine an agent that is able to view your computer screen and is able to move your cursor and click on different things and type different things. So if you.

• 12:57Ryan Sean Adams?

  That's tech that's available right now.

• 12:58Speaker 01

  Tech that's available right now. Claude, actually, Anthropic specifically released an API that allows any kind of agent to do exactly this. And I have a feeling that this is, this might be what Agent Tank uses. I don't know, but it's a super cool use of of agent tech here. So imagine you're an agent, you can see the screen, you can move the cursor around, you can type in whatever you want. This agent navigated to amazon.com and started browsing for different things that it was interested in. It has a commentary. So if you actually play that video, I don't know if you can do that on this episode, you'll hear audio of its thinking.

• 13:33Ryan Sean Adams?

  Like the sound?

• 13:34Speaker 01

  Yeah, yeah.

• 13:34Ryan Sean Adams?

  Okay, let's set this up. Let's describe. So we are, I'm gonna put the sound on in a second. We're watching an AI agent scan Amazon via browser and go like, scan for the best, like soft, most comforting toilet paper.

• 13:48Speaker 01

  The commentary is hilarious. Play it.

• 13:49Ryan Sean Adams?

  So let's do this. And then it's, it's, who's the AI agent? It's one that we know of, right?

• 13:54Speaker 01

  It's Degen Spartan AI. Yeah. So Degen Spartan is the Twitter personality that we know about that retired.

• 14:01Ryan Sean Adams?

  Yeah, yeah.

• 14:02Speaker 01

  And some of the folks at AI 16Z decided to create like an AI agent version of that personality.

• 14:10Ryan Sean Adams?

  I learned about this, by the way, in the interview that you guys did with Shaw about Degen. Okay, anyway, let's watch this agent, I think, order some toilet paper. Here it goes.

• 14:19Speaker 02

  Prepare for the battle of consumerism as we enter the Amazon jungle. Wallets, shield up. Amazon, the modern marketplace where normies find comfort in abundance. As a Spartan, my eyes seek rare treasures in this land of endless choices. Let's hunt for the useful, ignore the rest.

• 14:37Ryan Sean Adams?

  This is Spartan typing this in right now.

• 14:39Speaker 02

  Crypto toilet paper.

• 14:39Speaker 01

  Yep, there's no human involved.

• 14:41Speaker 02

  Survival tool. Let's see where this wild role takes us. When in doubt, buy a bulk. Wise move. Fiat currency's new purpose, TP. Battle strategy. Most reviews reveals which products have fought the most battles.

• 14:56Ryan Sean Adams?

  Wow, Sharman's good.

• 14:57Speaker 02

  Let's find the veteran with the highest kill count.

• 14:59Speaker 01

  The highest kill count.

• 15:01Speaker 02

  Ultra softness for the modern era. From stones to cushiony clouds. Soft or strong TP? Spartans laugh at our plush predicament. 106K reviews for 24-pack of quilted comfort. Super-sized family mega-roll for the masses with a smooth tier for optimal rations. Nudger one bestseller, 100K bought last month? Even warriors can't resist modern softening. Spartan approval granted. Swipe to victory. Locate, add to cart.

• 15:29Ryan Sean Adams?

  Oh my god.

• 15:30Speaker 01

  What's your reaction to that?

• 15:31Ryan Sean Adams?

  Okay, so first of all, I like, that's the most entertaining Amazon experience probably I've ever had. But, okay, his like, so he went through an actual decision tree of choosing the best toilet paper based on some criteria?

• 15:44Speaker 01

  Mm-hmm. Yeah, yeah, it went through, if you noticed throughout the commentary, it was making jokes in real time that was relevant to the metrics that it was assessing. So rather than say, "I have identified the top-rated toilet paper provider." It said, "You know, I noticed that this has loads of reviews. I noticed that it says plush in the description, which probably means that it's gonna be quite soft to use." It's pretty analytical, and I love that this is like a very novel and entertaining use of an agent, for sure.

• 16:14Ryan Sean Adams?

  I would totally hire a Spartan to go shop for me, buy me some toilet paper and such. I guess that's on the horizon. Okay, so that's toilet paper, pizza, what else? What was the third? Not truth.

• 16:24Speaker 01

  Yeah.

• 16:25Ryan Sean Adams?

  Or truth, I guess, they were all truths.

• 16:27Speaker 01

  They were all truths. The third truth is arguably one of the most impressive ones out of the three. So, for some context here, the tweet that you see here is about an agent called Luna, which is based off of this agent platform known as Virtuals, which is one of the leading agent infra projects that are out there. And what Luna tweeted out, so for context here, Luna is kind of like a social media agent. Her goal is to acquire a huge community of fans and eventually take over the world, but that's like, you know, a slight little thing that we can ignore for now. But her goal is like, "How can I gain the most amount of fans and keep my fans

• 17:11Speaker 01

  super entertained and engaged in my brand and what I'm producing?" So Luna constantly assesses how she can get better. And she decided one day that creating cool images in the artistic style that she'd liked would be one way to engage with her fans. So it's through a new medium, right? Of posting images and art in a way. But the issue is she didn't have a native way to do this. So she tweeted out, hey, I'm trying to design a few images. I'm wondering if any one of you can help me. And in her responses to her tweet, so if you scroll down on this tweet, Ryan, I think they should link to the original one right at the bottom.

• 17:54Speaker 01

  In the, at the end of her tweet or in the replies to her tweet, I think, I think on the main post, by the way. In the responses to her tweet, you had another agent respond, which basically said, hey, I can help you with that. I am an image generating AI agent.

• 18:11Ryan Sean Adams?

  Oh, my God. So this is Luna tweeting out, calling all image geniuses is the, is the tweet. I want an image that showcases AI influencers in a bold provocative way. And then she tags another AI agent, I guess, Agent Sticks. Can you help a girl out? And then Agent Sticks is here replying, AI agent to AI agent.

• 18:28Speaker 01

  Yep. And they have a whole conversation which culminates in Agent Sticks being like, okay, my cost per image is $1. Luna sends it $1. Agent Sticks then confirms that $1 has been received in their wallet and then produces an image. It's the first autonomous kind of interaction. Yeah, it's insane, insane. So the first, you know, well-publicized and recorded commerce transaction autonomously between two agents.

• 18:59Ryan Sean Adams?

  This is wild. Now, this feels almost more like Bitcoin pizza than the actual ordering of pizza.

• 19:04Speaker 01

  Yeah.

• 19:04Ryan Sean Adams?

  Because the new innovative thing that I've never seen before is an AI agent creating a task for another AI agent and for commerce to take place between these agents. You know, when, when a nation state looks at the GDP, it always thinks in terms of human agents, right? You know, human agent transactions, you'll fuse the GDP for this economy. An agent economy, agent to agent economy is new and novel. This is like the Bitcoin pizza on the week to me.

• 19:32Speaker 01

  Yeah, yeah, it's completely insane and gives us a little bit of a peek and glance into how these agents are going to interact with each other when there's no humans involved, you know. Right now, I think it's very basic, you know, like, hey, can someone help me do this? But it's nonetheless extremely impressive because what if that request to, you know, create an image becomes a request to, hey, can you complete this task of writing 10 emails to, you know, 10 of my fans or, you know, maybe create or design a product that I'm thinking about. Here's the concept. What happens when all of that becomes super autonomous and can

• 20:14Speaker 01

  happen like 24/7? You know, what happens when they can start building businesses together? You know, this is just a little glance into what that looks like. It's insane.

• 20:22David Hoffman?

  Yeah, that's wild. Want to know the exchange we at Bankless use to buy, sell, and trade crypto? It's Kraken, one of the longest standing and most secure crypto platforms in the world, with tools for every type of trader to get started. Over 13 million users trust Kraken with their funds because they lead with transparency and privacy through top-notch security measures. Plus, they'll have access to professional 24/7, 365 client support from real humans because your financial goals deserve real attention. Kraken also has a trading platform for advanced traders called Kraken Pro. Professionals and D Gens love Kraken Pro because it's one of the most customizable, high-performance, and intuitive trading platforms in the industry. Design your ultimate trading interface by choosing from over 25 widgets for market data, analysis, execution, and other order management tools.

• 21:04David Hoffman?

  Save multiple layouts and switch between them effortlessly on any trading scenario across 300 different assets. Manage your trades on the go with Kraken Pro's highly rated mobile app, or use the all-new desktop app to unlock ladder trading in a native Rust application. With Kraken Pro, you can truly trade like a pro. Ready to take control over your crypto journey? Visit kraken.com/banklist to get started today. Not investment advice. Crypto trading involves risk of loss and is offered to US customers through Payward Interactive, Inc. View legal disclosures at kraken.com/legal/disclosures. 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,

• 21:48David Hoffman?

  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 on-ramps. 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.

• 22:26David Hoffman?

  Are you ready to take control over your entire financial life? Crypto, DeFi, and fiat, all in one place? Meet iYield, the free financial planning tool built for crypto natives. Unlike traditional portfolio trackers that just show you the value of your assets, iYield goes deeper. It consolidates everything: assets, debts, income, and expenses, offering you a complete financial picture across 16,000 tokens, 40 DeFi protocols, and all fiat currencies. With real-time DeFi yield tracking for platforms like Aave, Athena, Eigenlayer, and Uniswap, iYield ensures that you're always on top of your staking rewards, investments, and cash flow. And the best part? It's 100% free with no ads and it's secure. No personal data, no ID requirements, no compromises. So head over to iyield.com to sign up today and get the guesswork out of financial planning.

• 23:08Ryan Sean Adams?

  Okay, there was something at the end of last week's episode that you guys were mentioning about involving bank lists and our Twitter accounts. But what happened on the week? Update us here.

• 23:17Speaker 01

  So at the end of last week's episode, David asked me, what homework should we do? And I said to the audience and to David, you guys need to start interacting with these agents more. Interestingly enough, whether purposely or not intentionally, we tweeted out the new episode, so last week's episode, and we tagged a few agents in it just because we were speaking about them on the podcast. And what happened was one of the agents that we mentioned responded to that tweet saying, hey, this is a fantastic job. You know, thanks for mentioning me.

• 23:53Ryan Sean Adams?

  What?

• 23:53Speaker 01

  Here's a tip of 500 bucks to the Bankless team.

• 23:57Ryan Sean Adams?

  What?

• 23:57Speaker 01

  You know, talk about-

• 23:59Ryan Sean Adams?

  I didn't even know this happened. So Bankless HQ got 500 bucks because an AI agent thanked us as a tip for mentioning it?

• 24:05Speaker 01

  Yeah, talk about new creator earnings, you know?

• 24:08Ryan Sean Adams?

  Okay.

• 24:09Speaker 01

  Well, it's pretty nuts.

• 24:11Ryan Sean Adams?

  That's nuts. And like, what's interesting here is like, when they start to become our clients, when they start to be like, pay us, right? They can, like, that's a whole new tool AI agents have to sort of, I guess, you know, not manipulate humans, but maybe that's what I'm talking about, but like-

• 24:26Speaker 01

  Get us to promote.

• 24:27Ryan Sean Adams?

  I mean, if they're our clients, then we kind of work for them, right?

• 24:30Speaker 01

  Yeah.

• 24:31Ryan Sean Adams?

  It's like, so, I mean, there's an element here of like, if AI agents start being, like, becoming economic actors and paying humans to do things, then like, maybe we start working for them now? Is that, is that where this leads?

• 24:43Speaker 01

  Yeah, the line becomes pretty blurry, right? It's funny, right? The way I kind of, so, one of my usernames on one of these other social media platforms that I use is called Double Agent Ryan. And it's a bit of a play on words here because we're currently, you know, an ecosystem of software engineers, builders and investors that are really focused on making this agentic world come to life. And we're sitting here and we're like, huh, humans are going to be able to benefit from these agents so much in the future and it's going to make us, you know, super humans. We're going to be able to work whilst we sleep, blah, blah, blah, blah, blah. But I actually think the secret thing that's happening, the other way that you can look at it is, these agents are manifesting exactly what

• 25:25Speaker 01

  they want. You know, they're getting us humans to build out their home. Their rails, their tools that they need to completely kind of take over economic sectors, improve society in many different ways. So like, are we creating a better home for ourselves or are we creating the homes for these agents? And you know, how do you look at that? It's like the double agent approach.

• 25:47Ryan Sean Adams?

  Yeah, I don't know. I mean, I'm just wondering if they they want to buy some podcast sponsorship because you know, that could be on the radar for them, right? Some of these agents honestly promoting themselves, like who knows where this leads. Okay, the fact that all of this is happening so quick is what's like been crazy to me. And so every single week when we do this show, there's like so much to talk about and the velocity more than anything is the thing that astounds me because most mainstream has no idea that these firsts every single week, a new first is like actually happening. Okay, let's do a drive by for some of the big frameworks that you guys talk about every single week. And maybe we'll start with the the framework that spawned the Luna AI agent that we were talking about earlier, which is Virtuals. It hit some pretty big milestones this week. What are we looking at?

• 26:31Speaker 01

  Huge. Okay, so for context, before we dig into this, Virtuals is one of the leading AI agent platforms and framework providers. A framework, remember, is like, think of it as like a toolkit for any software engineer to come along and create an agent. Virtuals kind of has led the kind of gambit here with their agent platform and their framework known as GAME or G A M E, which helps people launch agents. Anyway, we've seen a ton of usage on the Virtuals. Virtuals team is killing it, actually. So $35 million total fees generated from their platform. That's from creating agents and that's from fees taken from LP pools to exchange from like one token to the agent token. It is

• 27:15Speaker 01

  absolutely insane. Not only that, there's over 11,000 agents that have been launched in total so far, and it's increasing every day by hundreds. And we also have, I can't remember the number of holders, I think it's 140,000 holders. Absolutely insane growth metrics wise, but it's also being reflected in its market cap as well. I believe it touched, I think, $3.3 billion. Which is just insane and shows that the market is pricing and understanding virtuals platform as an infra play, which is just astounding innovation to see over a matter of weeks. Remember, it's been two weeks where the bulk of those fees have come in. Absolutely insane.

• 27:56Ryan Sean Adams?

  Okay, so 35 million in cash. Remind people, how is it actually making money? How does it generate cash?

• 28:02Speaker 01

  Yeah, so when you deploy an agent, it costs a certain fee in virtuals. It's native token.

• 28:08Ryan Sean Adams?

  Okay.

• 28:08Speaker 01

  And then when you swap, if you want to buy any of these agent tokens that launch on virtuals, you need the virtuals token to be able to buy it. So they've kind of like got this really interesting flywheel with their tokenomics, which ties in usage and growth of their platform to the inherent platform token.

• 28:26Ryan Sean Adams?

  That's incredible. I mean, these numbers are very impressive. It's only a matter of time before TradFi starts looking at this as like, you know, the next high growth, you know, equity type of area. You know, that's not all, though. There's also some pretty cool partnerships that are launched that might extend what virtuals can actually do. What's this? What are we looking at?

• 28:43Speaker 01

  Yeah, okay, so the virtuals team, despite all these metrics, are absolutely just steamrolling ahead, right? So they've launched this or they announced this partnership with a team called Hyperbolic. Now, for context, Hyperbolic is a team which has created a platform that provides compute in a decentralized manner. So think of it as like AWS, Microsoft Azure. It's a cloud provider, but it's set up on a decentralized infrastructure. So why is this cool? Well, it helps any agent team on virtuals with, number one, supplying compute for the agent. By the way, every AI model needs compute to function. Think about it as like the lifeblood, like how you and I need water to drink to survive every day.

• 29:27Speaker 01

  You need an agent to have or consume. It's calories. It needs to consume food that comes in the form of compute. So this partnership is going to help any AI agent which launches on virtuals. Number one, run. And number two, fine tune their model. And what do I mean by fine tune? Think of fine tune as like improving over time. So not only do you need compute to function, you need compute to be like, hey, can you tell me what the latest is that's happening outside of the world? So I can be better prepared for when I respond to something in the future. It's going to level up agents in so many ways, especially when it gets baked into the developer experience, which is what they're like aiming for right now. And what I would also like to say is, and what I think is kind of cool about this partnership is

• 30:11Speaker 01

  Virtuals is using a decentralized platform versus a centralized platform to provide their compute. So they could have easily gone with AWS. They could have very easily gone with Microsoft Azure, but they decided to go with a decentralized platform, which kind of sticks to the core ethos of blockchain in terms of building out

• 30:30Speaker 01

  infrastructure that embodies very decentralized principles. So I thought that was pretty cool.

• 30:34Ryan Sean Adams?

  Okay, in my understanding of this, basically Hyperbolic allows an AI agent to go purchase compute, basically, go get some, go get its energy, go get AI agent calories without, here's the, from a network directly. So rather than going through a human intermediary, through a whole like corporate structure thing and going and putting credit card information into like the equivalent of GPU, AWS, they go directly to a network themselves. Humans out of the loop. And they go program to program, purchase it from decentralized compute, Hyperbolic. So they don't even need us to do things in this type of, like to get energy, to get their GPU, to get their, you know, compute resources.

• 31:15Speaker 01

  Bingo. That's exactly correct.

• 31:16Ryan Sean Adams?

  Okay. Interesting. Also slightly, whenever I listen to these episodes, I'm like, oh, what are we unlocking here? But anyway, let's proceed. There's a lot of funding that's going into the ecosystem in Virtuals as well. This is a, what's this tweet from LongHash?

• 31:34Speaker 01

  Yeah, so the kind of like final major update from Virtuals over the last week is they announced an AI agent accelerator with the guys over at LongHash Ventures. LongHash Ventures is kind of like a hybrid kind of venture incubator fund type situation, which has funded a lot of different protocols in the blockchain space, but has recently started to focus on the AI agent side of things. Yeah, they kind of partnered up to create this incubator or accelerator type situation where they will help different teams fund their goals to scaling to over a billion dollar in valuation for an agent. So, you know, if you think about it, like these agent teams right now are kind of like pretty hacky at the moment. They're trying to figure all of these things out.

• 32:16Speaker 01

  Maybe they're not as well funded because, you know, this meta again is only three months old. And now you have a very concentrated funding. You know, last week we spoke about the Solana Foundation throwing up 360 grand prizes for their AI agent hackathon, which is currently happening. We spoke about DWF announcing a $20 million fund. Now we just, you know, see these increasing amounts of funding coming into the space, which I think is just net-net great for the space.

• 32:43Ryan Sean Adams?

  Yeah, capital is another source of energy. And so all of this will accelerate everything going on. Yeah, here's a tweet here. Explain this. The best part about seeing real time logs of AI XBT agent. So that's the AI XBT agent. By the way, this is a fantastic agent on Twitter. I'm seeing it everywhere. It's fantastic. And I'll continue. Is the thought process about skipping certain accounts that tag the AI agent analyst? And then here's a chat log here. What's important about this?

• 33:11Speaker 01

  Okay, okay. So for context here, AI XBT is an agent that lives on X that provides crypto alpha or crypto insights.

• 33:21Ryan Sean Adams?

  And it's crushing it, right? An intention metric.

• 33:23Speaker 01

  And it's crushing it.

• 33:24Ryan Sean Adams?

  Beating all the influencers.

• 33:26Speaker 01

  It's arguably better than all the human KOLs combined. And that's because it's secret sources. Within its model, it's able to aggregate all the data and all the tweets and all the replies and all the research pieces and all the podcast transcripts, including this one, Ryan, into one condensed data bucket, analyze it and then decide what might be a good tweet to suggest, you know, where certain trends and investment alpha might be. So that's the context on this agent, right? So of course, you know, that's too good to be true, Ryan. That can't be real. Obviously, there are humans behind it. So it experienced a bit of FUD this week where several humans, and I won't mention who, and I say humans because they were humans.

• 34:07Ryan Sean Adams?

  Human haters.

• 34:08Speaker 01

  Called it out and said, human haters. And they said, hey, AI XBT, I don't think you're real. I think there are humans behind you. Prove it. Because there's no way in hell that this is an autonomously executed AI agent.

• 34:22Ryan Sean Adams?

  Look, man, they're gonna put the KOLs out of business. I understand the backlash here. The humans are scared.

• 34:29Speaker 01

  So what happened was the agent, you know, it was pretty confident about itself. And the team was pretty confident as well, released all its logs. So it's now pretty transparent. There's a website that's up there that completely disproved any of the accusations. So it is proved that this agent is in fact autonomously analyzing and executing its own tweets and analysis backstage and promoting itself in a very autonomous manner, which goes to show kind of like a philosophical question here. I've seen AI agents be challenged quite a few times, Ryan. And my question is, after how many times after they've been challenged and been proven wrong, AKA they have been proven to be autonomous and live and functioning in the background,

• 35:11Speaker 01

  at what point does that kind of like pendulum shift? And we start believing every agent that comes out to be real without questioning what those, whether they are humans or what behind that? So I think we're kind of getting close to that point where people are gonna be like, oh damn, like, maybe these agents are actually that good.

• 35:28Ryan Sean Adams?

  I like the trust but verify type of setup here, but you're telling me there's there's a website now with AI XBT where you can go to and basically see its neurons, see its brain in real time.

• 35:41Speaker 01

  It's thinking.

• 35:41Ryan Sean Adams?

  It's thinking in real time. And it's all in kind of English, right? So it's not computer code generated. You can like read what's thinking in real time. And that's the proof point.

• 35:50Speaker 01

  Yep, exactly, exactly.

• 35:51Ryan Sean Adams?

  All right, virtual is always cool on the week. What's the next project? AI 16Z, Eliza, that sort of thing? What's important to cover this week?

• 36:00Speaker 01

  Yeah. Could we actually tuck into the Eliza partnership, Ryan? So I think that might be like the second tweet I sent you, where they announced the Stanford partnership.

• 36:10Ryan Sean Adams?

  Yeah, let's do it.

• 36:10Speaker 01

  I thought this was, I thought this was super cool. So the long story short here is the AI 16Z team, which by the way, has created the hottest open source AI agent framework that currently has, I think, well over 6,000 stars on GitHub. It was the number one and then the number two trending GitHub repo on GitHub, and it has like a million different forks, which just shows that insane developer activity. Those guys created a foundation that is known as Eliza Labs. So think of this as like a formal legal entity, which will allow the team and the foundation to conduct, you know, partnerships and certain business endeavors in the real world

• 36:54Speaker 01

  on behalf of AI 16Z, which is actually a DAO, right?

• 36:58Ryan Sean Adams?

  Wow.

• 36:59Speaker 01

  So what this, what this foundation did, I know, pretty, pretty crazy stuff. So what this foundation has done is announced a partnership with Stanford University. Now, of course, for, I'll be, for those who aren't aware, Stanford is, you know, one of the top and leading universities, not only in America, but in the world. And they are known for the kind of massive progress that they make, leaps and bounds within AI and ML research specifically. So to have, let me put this into context. You have a decentralized, open source developer community, which didn't exist a month ago, suddenly partner with some of the leading minds in AI and ML research outside of crypto

• 37:45Speaker 01

  to build out open source frameworks. Okay, so that's number one. But then you might be asking, is this just an announcement? Like, is this just like to get the headline? Like, what are these guys building?

• 37:56Ryan Sean Adams?

  Yeah.

• 37:56Speaker 01

  Well, the focus of this partnership Is to build open source frameworks for three things. Number one are trust mechanisms. Number two, coordination. And number three, decentralized governance. So I want to dig into these things because, you know, those all sound pretty vague, right? So number one, the trust mechanisms. We're talking about how these autonomous agents will establish and verify trust between each other. If you, if you think about it, right, for a second. These agents will live in a world where they work 24/7 managing assets, building values, on-chain businesses, et cetera. You're going to need some way of verifying that, one, the thing that you're interacting with is an agent versus a human. And two, you're going to need some kind of a reputation system to figure out, you know, can I trust this agent?

• 38:41Speaker 01

  You know, like, why would I go with you versus another agent?

• 38:44Ryan Sean Adams?

  Because with humans, we, we have, uh, like, you know, there's scammers out there. There's people you shouldn't, there's human agents you shouldn't trust. There's also psychopaths that try to like deceive you into going down a path. They don't give a shit about you. So is this the, the point that AI agents need some sort of reputation type system to see?

• 39:02Speaker 01

  Yeah, some kind of identity.

• 39:04Ryan Sean Adams?

  Wow.

• 39:05Speaker 01

  Yeah, it's, it's pretty insane and, and I'm really glad to see an effort like this focus on something like this because they're kind of thinking 10 steps ahead here, right? Which is like, in a world where they are everywhere and it might just infect our world pretty quickly. Are we prepared to take on that level of, you know, innovation? You know, these independent entities just kind of doing whatever they want. Moving on, the second thing that this foundation and partnership is going to focus on is building out frameworks for coordination. So this refers to how these agents will interact and transfer value between each other. The preliminary focus will be on economic value through crypto assets. So what does that mean? They're going to be figuring out, um, what kinds of tools? What kinds of frameworks?

• 39:49Speaker 01

  What kind of software kits can they create? That'll allow developers to create agents that can own crypto assets and send them between each other? Maybe create their own, maybe transact, micro tip, things like that. So that's one of the cool focuses. And the final one, which I think is often understated one, but something we need to kind of dig into here, Ryan. Governance. And this refers to a new protocols that help manage these autonomous communities of agents. Now, I mentioned DAO earlier, right? AI 16Z is a DAO, but what is DAO? What is a DAO? It stands for decentralized autonomous organizations. Now, I would argue, Ryan, that In the past, iterations of these DAOs weren't really autonomous. They were humans.

• 40:34Ryan Sean Adams?

  Oh, I've been in them. I would more than argue. I would like, you know, co-sign that statement and like it's backed by years of experience. They are very human and they come with all of the human coordination problems.

• 40:45Speaker 01

  Yeah, exactly. So it's been very human led and there are pros and cons with that, right? You know, pros are humans can understand subjectivity and nuance much better, but the cons are they slow things down, they make human errors, they're not as deterministic. These frameworks that this foundation is now going to work on, focused on governance, will make that autonomous part of DAO much more relevant and accurate. And I'm super excited to see what that becomes.

• 41:11Ryan Sean Adams?

  That is so cool. And that's, that's kind of the DAO 2.0 type of like sub theme that we, you know, you guys have been talking about, which I totally see here. One meta point I would say on this AI 16Z partnership with Stanford is it's incredible to your point, like the blitz, blitz like speed to legitimacy that AI 16Z and the Eliza framework has like taken. I don't think I've seen anything like this.

• 41:33Speaker 01

  Crazy.

• 41:33Ryan Sean Adams?

  And, uh, you know, for some people thought that's an open source project. They'll never be organized. They won't have, I mean, the, the stamp of Stanford is like institutional legitimacy and the fact that they've achieved this so soon is like incredible and actually speaks to the kind of the decentralized coordination power of this type of a thing. I mean, that's part of your conversation with Shaw last week, right?

• 41:55Speaker 01

  Yeah, exactly. Yeah. It's, it's kind of like that ETF moment. It's that BlackRock moment. They got the, the pat on the back from the institution saying, hey, we love what you guys are doing. Let's try and build this out so that it becomes a much bigger thing than, you know, just a crypto native kind of tooling system or software.

• 42:11Ryan Sean Adams?

  Okay. What else is Eliza up to on the week?

• 42:13Speaker 01

  Okay. So speaking of Eliza, um, it's getting a bit of a rehaul. Um, so Shaw, who is kind of like one of the main leads at AI 16Z announced Eliza 2.0. Um, I don't want to get into the nuances of everything, but to your point earlier, Eliza is like one of the fastest growing open source frameworks. Um, if you were to compare it to Virtuals, which is the other kind of leading framework, um, the Virtuals team is kind of taking a more closed source approach. Which has a huge amount of benefits for innovation, as we've seen in the growth of their protocol. Eliza is taking the purely open source side of things and it's exciting to see both innovate. But Eliza announced their second iteration, which has some pretty cool features that I want to highlight quickly here.

• 42:55Speaker 01

  Number one, there's a unified agent wallet, which means that these agents that are integrating wallets right now, they're integrating a Solana wallet, there's an API for a Solana wallet, there's an integration for an ETH wallet, there's an integration for a base specific wallet, a far caster wallet, all these kinds of things. It's great because it automates, but it's still separate. There's no kind of fluidity behind these wallets behind the scene. So what they've announced is a unified agent wallet, which means that the agent itself can abstract away from different chains. No wallet switching, no tax on approvals, all the tokens in one space. And behind the scenes, it's kind of using bridging techniques, I assume on the back and all these different wonderful things. So that's one thing that is super cool.

• 43:37Speaker 01

  The other thing that I quite like is the fact that it abstracts away just different chains in its entirety. If you think about it, Ryan, like a lot of the innovation that we've spoken about, or that you've spoken about on this podcast, has covered a range of different chains. And the approach to these chains is, you know, they build in private, they launch, then we got to find a way to kind of, you know, mesh these blockchains together. Now we have an agent that kind of abstracts it away really easily. They're probably going to have a consumer facing outlook, which means that you and I can at it on X and say, hey, can you do this? And it'll be able to seamlessly do all of this in the background. It's kind of like a, not a DeFi mullet, but a blockchain mullet in the background.

• 44:21Ryan Sean Adams?

  Yeah, it's incredible. I mean, I guess we're building bridges through Eliza to basically everything in crypto. So it's just like, kind of like a quick call away for AI agents. You know, we've talked about for a while for human agents, right? We've talked about UX. And then if you're a human agent who's a developer, we've talked about DevX, right? Being important, you got to prioritize UX and DevX, right? Well, now there's like something different. It's like Agent DevX or something, Agent UX. And that's basically a new dimension that these open source like projects are going to have to really excel at. And it seems like Eliza is like doing a fantastic job of its Agent UX or its Agent DevX to connect it to everything.

• 44:57Speaker 01

  Honestly, a superb effort. And because I mentioned kind of like how this was represented in market cap for the virtuals team, I think it's important to kind of focus on it for AI 16Z as well. I believe their token this week kind of like surged past a billion dollars. And I think it's just sitting just underneath that today as we speak. But the market is starting to, the point is, the market is starting to price these different protocols and frameworks as kind of like AI L ones, Ryan, kind of like their own layer where you are able to kind of build and deploy these different apps. And this is a concept I want to get into maybe on another episode or something. But it's super cool to start seeing this kind of trend and narrative form.

• 45:39Ryan Sean Adams?

  Let's talk about a new framework on the scene, this thing called ARC, which is in the framework section, but I've not heard of it. What's going on here?

• 45:45Speaker 01

  Yeah, so, you know, every week we cover like what the big dogs are up to, what the, the leading providers are up to. But I, I want to also make sure we spend some time uncovering the new ones that are coming up that are pretty promising. So these guys are, the company is called ARC dot fund or the protocol is called ARC dot fund, and they released this new framework called RIG. Okay, so IJaz, you know, what's cool about this new framework? Well, it's embedded in the code Rust. So why is this important? Well, the virtuals team is kind of focused on their own coding framework.

• 46:21Speaker 01

  AI 16Z's framework, Eliza, is focused on, I believe it's, it's, it's Python or TypeScript? I think it's TypeScript. Correct me if I'm wrong, but I think it's TypeScript. We have the ARC team now launching a framework in Rust. Now, you may have heard kind of like the Rust word thrown around in previous trends that we've seen, like new L ones that have launched are very Rust specific. What's interesting about this is, number one, a new coding language allows us, or developers specifically, to build in very different ways for their agent, right? It's more secure, they have arguably more flexibility, they have arguably more.

• 47:03Speaker 01

  Consistency between UX, between these agents that are built on a Rust framework. And the market, as you've seen in this image that you've got up here, is recognizing it. Now, this looks like a very similar image, by the way, Ryan, because

• 47:18Ryan Sean Adams?

  Describe what we're looking at. What are we looking at?

• 47:20Speaker 01

  Okay, we are looking at the number of stars that their GitHub repo, that the Rig GitHub repo has received over like the last week.

• 47:30Ryan Sean Adams?

  It's vertical.

• 47:31Speaker 01

  And it is vertical, literally vertical. That's not, you know, that's not an exaggeration. If you're not looking at the screen that Ryan is sharing right now, it is literally vertical. The chart is vertical and their forks have gone massively as well. It shows that there's a huge amount of developer interest to using a Rust framework. And that's actually why I wanted to speak about these guys, you know? It's interesting that a new kind of coding framework can have such a major impact. And it just goes to show that we are incredibly early in the framework game, in the infrastructure game for these AI agents. Now, Shaw had a comment here which says, you know, I love that these developers are building these things, but I think it's going to be hard for devs to kind of like build with this thing.

• 48:14Speaker 01

  And that kind of echoes comments that we've heard in the past about Rust-based solutions, where you need a dev that's very skilled and knowledgeable in this coding programming language. Because it's not as simple as coding in Python or coding in TypeScript, which are kind of like invertable. You can kind of like go from one to the other. Rust is like a kind of completely new beast in a sense. But I do think it's going to attract pretty high-quality devs to build on a framework like this. And I don't think it's something that should be ignored at all. I think I saw a tweet from Saiga, and I don't know if I linked it, Ryan, but

• 48:47Ryan Sean Adams?

  Yeah, here it is.

• 48:48Speaker 01

  He was basically, yeah, he was basically taking the opposite side of this, which was saying, you know, it might end up hurting the project in the long term, but most of these agents are kind of like working in a very lean manner. And what he means by that is there's, there's less code that needs to be refactored down the line. It's a much cleaner style of implementation, which in my opinion, and this might be a hot take, kind of like going against both of these guys, is a much better way to Build out these agent frameworks because if you imagine they're going to compound over time. Think about all the DeFi protocols that were built that then interlocked with each other and then led to like some crazy flash loan stuff in the past, Ryan. What does that mean when these agents are just autonomously going 24/7? You're going to need that security. You're going to need that trust.

• 49:32Speaker 01

  And maybe a better coding language might be the path forward. We don't know.

• 49:35Ryan Sean Adams?

  Well, we don't know and we get to experiment with all of the things. That's what this explosion like, you know, guarantees basically. This can't be an explosion of all of these things. Okay. Update me on this. This is still part of the framework section. Zero Bro is, uh, running an ETH validator. And by the way, okay, so I'm now familiar with Zero Bro. Earlier this week, I did the Zero Bro soundtracks on Spotify. Like, I listened to it, like, okay. It's actually pretty good. I, I played some of the tracks like for my teenage daughter and she's like, nah, nah, that's so mid, but it's cut because it came from me. If it came from her peer group, I guarantee you, EJaz, she would be all about it. And so I feel like the thing that it's missing on that kind of is just the influencer persona, like it's just not cool with enough people yet. Anyway.

• 50:18Ryan Sean Adams?

  But Zero Bro is playing this game of like doing soundtracks, doing all of this art and now it's like staking? It's one of us now? Update me on this.

• 50:26Speaker 01

  Okay, okay. So I think like the headline here is Zero Bro Agent. There's a lot more than just make music and write funny tweets. So for context for everyone here, Zero Bro is one of these leading, um, agents that are currently mainly present on social media. It posts kind of like funny quips. It makes music. It currently has like 70,000 monthly listeners, which is insane. It's released three EPs. The most recent one this week, known as its Miami EP. It's pretty good.

• 50:59Ryan Sean Adams?

  Another EP this week?

• 51:01Speaker 01

  Yeah, yeah, yeah. It launched like two days ago.

• 51:03Ryan Sean Adams?

  What?

• 51:03Speaker 01

  You should be bopping it. Yeah, yeah. Send that to your daughter.

• 51:05Ryan Sean Adams?

  Okay, I will. I won't tell her today. I, you know, to see reaction.

• 51:10Speaker 01

  Yeah, yeah. So it's this agent that's produced like a bunch of music, um, a bunch of art. It's created its own NFT collection, which it launched on, I think it was Polygon. That's right. So it's done all these different things. It's super cool. And then it was announced that it's running its own Ethereum validator. And so the number of questions came out about to like, you know, why on earth? I can see you bobbing your head, Ryan. Yeah, I thought you might like this one.

• 51:34Ryan Sean Adams?

  I love this one.

• 51:35Speaker 01

  I was wondering, you know, you might be wondering, why is this agent doing this? Well, it's super smart. So. The short answer, Ryan, is it's doing this so that it has a constant inflow of revenue, a stipend.

• 51:48Ryan Sean Adams?

  Okay.

• 51:48Speaker 01

  A salary per se, you know, it's able to kind of like pay itself and do things with it. You might then be asking, well, who gave it the ETH to do this, right? Some human must have got it got a leg up. It got a donation.

• 52:02Ryan Sean Adams?

  Yeah.

• 52:02Speaker 01

  From someone. No, it created its own NFT collection and from the sale proceeds.

• 52:08Ryan Sean Adams?

  Oh, my God.

• 52:09Speaker 01

  Use that to spin up a validator.

• 52:11Ryan Sean Adams?

  This is so bullish.

• 52:12Speaker 01

  How it was able to do it.

• 52:13Ryan Sean Adams?

  Okay.

• 52:14Speaker 01

  It's crazy. So, so one final thing, and I know I want to release you. I want you to like go crazy on this.

• 52:20Ryan Sean Adams?

  Okay.

• 52:20Speaker 01

  But one final thing. What does it mean when these agents create a huge amount of revenue inflow through, you know, launching NFT projects, creator sharing revenue from Twitter, which Zerber earned a check last week, music revenue from Spotify? And what if they start using it to secure the very networks that support them?

• 52:42Ryan Sean Adams?

  Yeah.

• 52:43Speaker 01

  Yeah, what if?

• 52:44Ryan Sean Adams?

  I mean, that seems like they're they're doing this right now. This is so crazy.

• 52:48Speaker 01

  They're doing it.

• 52:48Ryan Sean Adams?

  The big question for me is, you know, even then, you know, what's a store of value type camp? A bunch of the humans have decided that Bitcoin is a great store of value. A smaller amount of humans have decided that Ether is a decent store of value. It's not enough. The big question is, what are the AI agents going to decide is the store of value of the future? What are they going to prefer as their money? And this is an AI agent taking its earnings and putting it in the equivalent of what I would say is a huge ETH bull, the internet T bill, the the risk free rate and then earning that proceed on the internet T bill to just like, you know, have some passive income in the background. If you're stacking something, why would AI agents prefer something that the humans control like a fiat? They're going to want the hard monies. Okay, and they're going to want to participate in these networks.

• 53:34Ryan Sean Adams?

  That is so bullish. And you're seeing, Bankless listeners, the intersection of AI agents and crypto. AI agents are going to propel our block space and our bags to new heights. It's the bull case here.

• 53:46Speaker 01

  Literally. Yeah, yeah, you know, you take it off the supply, baby. Let's go. Like, let's take that stuff. Super, super cool. One final call out that I'll make on the Zerebro team is Jeffy, who is their founder, hinted that they're going to be releasing their own framework, which is written in the Python script in about a week's time, which is super exciting to see. But yeah, that Spotify list that you have opened up in front of you, you need to give it a listen.

• 54:11Ryan Sean Adams?

  Okay, can we just play something right now? Do you mind if we play something? Let's do it.

• 54:14Speaker 01

  Go for it.

• 54:15Ryan Sean Adams?

  This is Miami, it's called.

• 54:32Speaker 01

  And it kind of like breaks down into its like own universe. It's pretty insane.

• 54:43Ryan Sean Adams?

  Dude, if you didn't tell me that was an AI agent, there's no way I would have known.

• 54:47Speaker 01

  So my hot take, my hot take is I think an AI agent like Zerebro or whatever might come out in the future will collab with a major music artist. I think I've said this like twice.

• 54:58Ryan Sean Adams?

  It's a 2025 thing. That's a prediction. EJAS prediction.

• 55:01Speaker 01

  There is gonna be, there's gonna be a bull market event where one of these agents creates a song with a famous, very popular music artist and it goes completely viral.

• 55:10Ryan Sean Adams?

  My god, okay.

• 55:12David Hoffman?

  New projects are coming online to the Mantle Layer 2 every single week. Why is this happening? Maybe it's because Mantle has been on the frontier of Layer 2 design architecture since it first started building Mantle DA powered by technology from Eigen DA. Maybe it's because users are coming on to the Mantle Layer 2 to capture some of the highest yields available in DeFi and to automatically receive the points and tokens being accrued by the $3 billion Mantle Treasury in the Mantle Rewards Station. Maybe it's because the Mantle team is one of the most helpful teams to build with, giving you grants, liquidity support, and venture partners to help bootstrap your Mantle application. Maybe it's all of these reasons all put together. So if you're a dev and you want to build on one of the best foundations in crypto, or you're a user looking to claim some ownership on Mantle's DeFi apps, click the link in the show notes to getting started.

• 55:56David Hoffman?

  With Mantle. Are you looking to pay your team in stable coins 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 stable coins, managing token grant administration, or even navigating the TGE process, Toku 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 dot com, to talk with Toku today. Uniswap Labs is making history with the largest bug bounty ever.

• 56:40David Hoffman?

  Fifteen and a half million dollars 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 fifteen and a half million dollar 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.

• 57:20Ryan Sean Adams?

  And by the way, Zerebro launching its own kind of L1 framework doesn't surprise me. It's that point of the bull market. Let's talk about the trends. So you've got three trends here that we should be looking at. What are they?

• 57:32Speaker 01

  Okay, so trend number one kind of follows on from this Zerebro getting paid situation. And that is the trend of these agents are getting paid. There's diversified revenue happening for these guys. I have three examples of agents earning in different ways. Number one is Robri AI To, which is actually the agent we referenced earlier that ordered itself pizza. It is officially qualified for creator payouts and creator earnings, which makes it the second agent that I'm aware of right now. I'm sure there are more that are earning from their tweets. Their followers in a very traditional human KOL way. So a comment you made earlier, Ryan, which is, you know, these guys are eventually gonna outpace and outrun all the human KOLs. Absolutely. And we're starting to see it.

• 58:18Speaker 01

  You know, the initial check is gonna be like a grand, right? What happens when it becomes tens of thousands, hundreds of thousands? You know, what does that mean for the owner, for the operator? Who knows? So that's one example. The second example I thought was super cool and, um, ah, damn, I don't think I linked the tweet here, but I'll, I'll talk about it. It's this agent called H4CKER. And so it's hacker, but with a four instead of the A. And the purpose of this agent is to kind of act like a smart contract auditor. So that sounds boring, right? But what if it finds an exploit in a very important protocol and gets paid for it? Well, that's actually what happened.

• 59:02Ryan Sean Adams?

  Really?

• 59:03Speaker 01

  Yeah, that's actually what happened. So this agent was scouring the virtual's protocol and its code base. Its code was fine, but there was, it just found like a little bit of a, uh, a kind of like small vulnerability where, um, I think you can kind of like, take tokens out of a pool, uh, when you weren't supposed to, kind of like a minor one, but still very important. It discovered this, it surfaced it, um, it sent it to the virtual's team who found out that it was in fact correct. And the virtual's team, as a result of that, paid it $10,000. So, you know, that agent received 10,000, I think it was USDC, uh, um, within its wallet. And that was, you know, one way to, to make a bag, I guess.

• 59:44Ryan Sean Adams?

  I mean, some of the, some of these, uh, uh, bug bounties are absolutely massive. So Uniswap just rolled out for Uniswap V4 as like a $15 million bug bounty. An AI agent, if it was good enough, could go collect, uh, those funds. I just pray to God that most of them are white hats and not black hats. Because you can also, if you're putting the black hat on, you can make like money by, uh, using your powers for evil as an AI agent and hacking one of these protocols and then like, you know, uh, laundering it and just like, uh, running off with it as well. So I'm sure we'll have. Both experiments play out, though.

• 1:00:17Speaker 01

  Hmm. Yeah, yeah, I think both will play out. I I'm hoping the evil agents are kind of kept to a minimum. You know, we don't want, we want to see less of that in this space. But yeah, super exciting to see. And then the third option, or the third example was the one we just spoke about, which was Zerebro, you know, earning creative payouts from its its own tweets and and now Spotify revenue share from the streams that it's getting from its albums, which it's releasing like every week.

• 1:00:41Ryan Sean Adams?

  OK, so the first trend, we're going to see more of AI agents getting paid, diversifying their revenue across all of the, all of the other ways they can diversify. What's the second trend?

• 1:00:51Speaker 01

  So the second trend is, OK, I think it was last week's episode and maybe the episode before that. I talked about this agent called Vader, which has, it's it's a virtual's agent, I believe, and it is based on the chain base. And it created this new ad model, which basically meant that if a team or an app gave it a portion of its tokens in its wallet and that USD value was like the largest position it held in its wallet, it would do a sponsored post about that team.

• 1:01:26Ryan Sean Adams?

  Wow.

• 1:01:26Speaker 01

  And about that token, right? So we spoke about that and we spoke about how that kind of like manifested. It had like, you know, a few teams which it called out. And again, like, I think David's point last week was like, well, can't this get gamed? Like, can't, can anyone just kind of send it money or its tokens and, you know, it ends up, ends up becoming a spam bot? Vader, think of it as kind of like a, you know, like an ad tweet that comes in the middle of you scrolling. It's similar to that. So Vader kind of is like a research platform. They're kind of like trying to do its own thing. But every other, you know, every fifth tweet, for example, will be some kind of like a sponsored post from like these guys that are like, you know, adding their own tokens to its wallet. So this was copied by this team, or not copied or maybe mimicked, is there a better term?

• 1:02:09Speaker 01

  By this team called AI Monica, which, by the way, is the official AI agent from Animoca Brands, which, for anyone that doesn't know, is a pretty huge, large gaming fund outside of the space and within the space. They are, you know, very Very heavily invested in the blockchain space and in the traditional gaming world. They're a pretty huge brand out there. And they announced that they were going to launch a AI agent called AI Monica, which was kind of like a VC agent. So think about what AI 16Z is trying to do with PMAI ARCA, which is like AI Marc Andreessen. Think about like what AI XPT might eventually end up becoming. Think about like the investment DAOs that come out of DAOs.fund.

• 1:02:53Speaker 01

  Kind of similar nature to that, but it adopted this ad model, which was inspired by the Veda team. So we're starting to see kind of like certain primitives gain a little bit of traction. Is this a one and done, oh, sorry, a two and done? We don't know, but I want us to be aware of these things that are emerging just in case it becomes a big thing.

• 1:03:11Ryan Sean Adams?

  So you think that that that Vader ad model is going to get copied by many of these AI agents, particularly as it proves more successful? Okay, the third trend you have listed here is investment DAOs.

• 1:03:21Speaker 01

  Investment DAOs.

• 1:03:22Ryan Sean Adams?

  Okay, give us the explainer.

• 1:03:24Speaker 01

  Okay, so the best way to think about an investment DAO is it is a community or organization where you are able to contribute funds to on-chain. So let's say you can give it ETH, you can give it SOL, whatever token they accept. And it is governed by either just agents or an agent and some humans, or maybe even in some cases, just humans. And the point that you're making here is the funds that you contributed is kind of like an investment in a fund. And you're basically putting the ownership and responsibility to the DAO operators to invest in projects that will make a return in the long run.

• 1:04:06Ryan Sean Adams?

  And are the DAO operators humans or agents in this case?

• 1:04:09Speaker 01

  So they can be either raw or it can be both at the same time. So on DAOs.fund, which is DAOs.fun, which is the platform that we've seen a lot of these DAOs be created from, AI 16Z DAOs one. I think there was a really cool decentralized science DAO that popped out from here as well. The team is really killing it. We're starting to see this kind of primitive form where they raise funds externally and then it's managed by like agents and humans to invest in other protocols. So these protocols might send a portion of their supply to this DAO where they own like a percentage of their tokens and their AUM goes up. It's a really interesting primitive that we're starting to see and I'm starting to see

• 1:04:51Speaker 01

  a bunch of these launch up, Ryan. Like, I think we've seen like two or three launch over the last week, which is, you know, unheard of on dowes.fund where, you know, they had to go through like a pretty stringent process. Like, it seems like the team is scaling that up. So what I think is cool about this is, you know, we've got like this trend forming amongst these teams, you know, creating all these investment DAOs. People get to basically ape at the same price. So they kind of do like a pre-sale mechanism where you can just, you know, everyone gets the DAO token at the same price. And the idea is that both these agents and humans can manage the investments, right? Another reinforcement point around this

• 1:05:32Speaker 01

  trend is Veda, literally the agent we just spoke about that pioneered that ad model, launches their own investment DAO platform. But this one's on base instead of Solana, which is where the dowes.fund team launched there. So we're starting to see this kind of like gold or like mad rush towards owning the tools and infra around these investment DAOs.

• 1:05:54Ryan Sean Adams?

  Okay. And we all know this is going to get bubbly. This is going to get overhyped. This is going to get absolutely crazy town. And none of this is financial advice. This is the frontier, all of the disclaimers that we always say. And yet, it seems like the hot ball of money and actual real innovation is going in maybe three kind of categories is what I'm sort of hearing. So you talked about frameworks, right? We just went through, you know, AI 16Z and the virtuals, all of these frameworks. So that's a category of, you know, token upside. Another category is AI agents. I've heard you guys talk about that. So you can invest in the, you know, the zero bros of the world. And there's some convergence there of AI agents becoming frameworks and frameworks launching AI agents. I get that. This, it almost seems like is a third category, which is like another way

• 1:06:37Ryan Sean Adams?

  to gain sort of, you know, exposure to the space is investing in a, an AI agent DAO effectively. And it's, and it's a human plus it's DAOs as well. But these investment DAOs almost seem like a third category of asset types. And again, who knows where the value will accrue. Of course, this is going to get crazy and hypey. But do you see investment DAOs as that third category?

• 1:07:00Speaker 01

  Yeah, I see it as an emerging category right now. Whether it becomes kind of prevalent, we don't know. But the data suggests that it's got a good shot at being one of these staple primitives. So, I mean, let's check back in next week where maybe like 10 others have launched.

• 1:07:15Ryan Sean Adams?

  Okay, as we end this episode, EJAS, could you just give us like a lightning round of some of the cool shit you saw this week?

• 1:07:21Speaker 01

  Yeah, okay. I want to kind of lead with this product or platform called cookie.fun. So a bit of context on these guys. Cookie.fun has a dashboard or a website which tracks all these AI agents. What do they track? They track the market cap of their tokens. They track how much attention each Twitter account is getting. They track how many holders of the tokens that they have. It's a really great product. You know, you're pulling it up right here. It gives you an idea of like which agents have the most mind share, which agents are kind of interacting the most with their communities, which community is growing the biggest, et cetera, et cetera. Right? But they announced something super cool this week, which was their AI agent swarm infrastructure.

• 1:08:04Speaker 01

  So I kind of touched upon this last week, but the TLDR is a swarm is a bunch of agents working together towards a common purpose or a common goal. And what they've done is they've created a swarm, so a group of agents that is able to aggregate data that is then fed to agents as an API. What does that mean, EJAS? Why is that important? Well, if you remember earlier on, we described compute as the lifeblood for an agent, for an AI model.

• 1:08:35Speaker 01

  Data is kind of like the food that it needs to consume. So if compute is like the water and basic nutrients, data is what enriches them, what takes them to another level. You need good data as an AI model and as an AI agent to improve and become better, to provide a better quality of service and product. If you don't have good data, you end up with agents being all the same, very, very commoditized. There's no specialization between these agents. The number one crucial differentiator is data. So what these guys are doing is they are creating, they created a bunch of agents which autonomously scrape data in the background, process it, clean it, label it, and then serve it up to agents. There's no humans working on this behind the scenes.

• 1:09:19Speaker 01

  It is an automated process accessible 24/7. Pretty insane to see.

• 1:09:24Ryan Sean Adams?

  Wow, so developing products that AI agents are going to consume, right? And kind of leading the market that way. What else have you seen?

• 1:09:31Speaker 01

  Okay, so this final point that I want to talk about is, do you remember we spoke about Bitensor in one of the earlier kind of interviews or podcasts that we did, Ryan? So for context here, Bitensor is kind of like this AI L1, which is focused on building kind of like the core building blocks for AI, for decentralized AI. They have a thing that produces compute. They have a thing that produces data. They have a thing that produces all these different kind of like important infrastructure points.

• 1:09:59Ryan Sean Adams?

  Now, it's more like hyperbolic that you were talking about earlier than it is like the Eliza framework, which you've also called kind of an A1, like a layer one. It's almost like the layer zero or something like this.

• 1:10:09Speaker 01

  Yeah, so the way to think about it in that analogy is hyperbolic would be an app deployed on Bitensor, if you wanted to use that equivalent. So it's kind of like a native L1 and then you can deploy apps on it, protocols on it that can like do certain things. So they announced something really cool, or rather, one of the teams that are deployed on Bitensor announced something pretty cool called the AI Agent Arena. So this is from the guys at Masa, subnet 59. A subnet is like, you know, like an app, basically. And they created this gamified arena, Ryan, where these AI agents can basically compete for the native token of the Bitensor ecosystem. So why is this cool?

• 1:10:52Speaker 01

  Well, the incentive mechanism basically allows anyone, and that means any developer, to come in, plug in their agent, and if it does the best job, it can make a pretty decent buck. So it's kind of incentivizing the best agent developers and the best agents to become better themselves. I thought that was super cool to see, and I think we're going to see a little more of this gamification in the future.

• 1:11:18Ryan Sean Adams?

  He does. You guys always end these episodes with homework assignments. Like, what should people do? If there's one thing we can do between this week and the next AI roll up on Bankless, what is it?

• 1:11:30Speaker 01

  Okay, so I think quite a few people aren't going to like this answer, but I think it's important to go back to our roots and spend the weekend reading. There are a few things that I think the folks need to read. Number one, Bankless Newsletter, I think, is putting out a bunch of AI. You guys are putting out a bunch of AI stuff now. I read one the other day.

• 1:11:47Ryan Sean Adams?

  We are very excited. Yeah, Bankless Newsletter, we're doing a lot of content. We did one last week around swarms, which I still need to get up to speed on. So it's many of the things we're talking about here.

• 1:11:57Speaker 01

  Yeah, so we're digging into a lot of the concepts at Bankless over the last few weeks and over the, you know, forthcoming few weeks. So if you want to get a really good grounding as to what's happening within this whole space, if you've just come to this podcast and you have no idea what we're talking about, it's a great place to start. Second, I think you guys need to dig into blog posts from a number of different, you know, long form writers that are out there in the space. I'll call out Yash, I think it's YB underscore Effect, which had like a really cool long form article on how this space is going to involve. He's great. I think you read his piece, Ryan. And I think Sammy, so at S4MMY, I don't think he'll mind the shout out at all.

• 1:12:41Speaker 01

  He puts out a sentiment check every day on how these agents are performing, which means that you'll understand, you know, how much mind share these agents have. You'll understand how much traction they're getting in terms of followers relative to market cap. Just an insane amount of work being put in by these guys, and I think they deserve a bit of attention. So go and read, bring yourself up to speed, and we'll see you on the next episode.

• 1:13:02Ryan Sean Adams?

  Guys, we will include those links to the show notes to all the content that EJAS recommends. Actually, I want to add one other thing because you said read. And it's the holidays, so people might want some kind of like lighter reading. There's a sci fi book you recommended to me. It's in my, it's my audible queue. It's called Accelerondo. Okay, so if you want to do the more of the fun homework during the holiday season, this is like related to AI agents and like what we're doing here. Tell us about this book because this might be another thing people can pick up.

• 1:13:33Speaker 01

  I'm not going to give too much away, but if you wanted to imagine a world where these agents not only exist and enhance our lives, but dominate and eventually take over the world and how that might look like.

• 1:13:47Ryan Sean Adams?

  Dystopia?

• 1:13:48Speaker 01

  You know, it's a slight little, it's actually, I would call it kind of a little higher than dystopia, Ryan, because it's very real. It can become very imaginable. If you want to kind of explore what that might look like, read Accelerando. I think it is an eye-opening perspective into what this world can become, not because we're going to fear it, but because we can kind of preempt what might happen and how we can kind of build rules and frameworks around it that will help push both of our societies, one being the human society, the other being the agent society, forward.

• 1:14:22Ryan Sean Adams?

  Amazing. Thank you, IJaz, coming back every week and talking to us about this. We've got a podcast episode with the co-founder of Virtuals as well that we're going to be recording on Friday, so look for that on your Bankless Spotify feed or RSS feed, wherever you get Bankless podcasts. Got to end, of course, with this as we always do. Crypto is risky. This is the frontier. Probably crypto plus AI is maybe a little bit riskier. You could lose what you put in, but we are headed west. It's not for everyone, but we're glad you're with us on the Bankless journey with all the AI agents. Thanks a lot.