AI ROLLUP #3: Solana AI Hackathon | Truth Terminal’s Roadmap | Satan & God Automated | Delphi Digital AI Analyst

Ejaaz is back with David to break down the latest in the explosive world of AI Agents. This week, they uncover how Truth Terminal’s roadmap could revolutionize autonomous on-chain transactions, why ai16z's Eliza framework is poised to reshape multi-chain agent deployment, and how zerebro’s creative push into music and art is expanding the reach of AI in culture. They also discuss the growing institutional attention on AI agents, from $20M funding announcements to agents acting as venture capitalists. Can AI agents truly become the catalysts for crypto's next bull market? 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 🛞MANTLE | MODULAR LAYER 2 NETWORK https://bankless.cc/Mantle 🐧CARTESI | LINUX-POWERED ROLLUPS https://bankless.cc/CartesiSimple 📈iYield: YOUR FINANCIAL PICTURE, SIMPLIFIED https://bankless.cc/iYield 🔒SAFE | INTRODUCING SAFENET https://bankless.cc/SAFE ------ ✨ Mint the episode on Zora ✨ https://zora.co/collect/zora:0x0c294913a7596b427add7dcbd6d7bbfc7338d53f/112?referrer=0x077Fe9e96Aa9b20Bd36F1C6290f54F8717C5674E ------ TIMESTAMPS & RESOURCES 0:00 Intro 2:31 Why AI Agents Are a Big Deal 4:34 Vibes This Past Week https://x.com/s4mmyeth/status/1866414390775329111 https://x.com/dwflabs/status/1866422609967235166 https://x.com/solana/status/1866536545550995466 9:18 a16z Blog Post https://x.com/a16zcrypto/status/1864343219833627113 14:26 GOAT is Returning https://x.com/pwnlord69/status/1866589407454826644 25:44 Satan & God Automated https://x.com/ropirito/status/1864373559927173257 https://x.com/god https://x.com/s8n 39:13 Eliza & Ai16z #1 trending GitHub https://x.com/dreaminglucid22/status/1865758299464843285 https://x.com/BanklessHQ/status/1866845399715287399 47:04 Virtuals New Sandbox Tool & aixbt https://x.com/0xCygaar/status/1866513570672079187 https://x.com/aixbt_agent 53:09 Zerebro https://x.com/jyu_eth/status/1864579428639986126 https://x.com/opaiumdao/status/1865484155464630470 https://x.com/opaiumdao/status/1866572986805035325 https://x.com/jyu_eth/status/1866389720844111884 https://x.com/jyu_eth/status/1866556713647870341 https://x.com/cryptopunk7213/status/1864473729855553766 https://x.com/blormmy/status/1865987541972738549 1:01:44 Swarms & vvaifu.fun https://x.com/vvaifudotfun/status/1864050584396906841 1:03:37 Clanker Fees cross $10M https://dune.com/clanker_protection_team/clanker-fees 1:05:57 $SIMMI https://x.com/Simmi_IO/status/1862510478745804856 https://x.com/SimulacrumAI/status/1866570195051090215 https://x.com/simulacrumai/status/1866641681795441070 1:08:46 Delphi Digital AI Analyst https://x.com/Shaughnessy119/status/1866212479413678588 1:11:02 Ejaaz’s Advice for This Week 1:12:22 Closing & Disclosures ------ 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.

• 0:00Speaker 00

  Instead of just humans and agents directly interacting with each other, I actually think the end game is we just have a ton of agents interacting with each other, building new businesses, new apps, trading value through tokens with each other, and building net new products that we've never seen in the space before and that we've never seen in the world before, to be honest.

• 0:23David Hoffman?

  Welcome, Bankless Nation, to the AI Roll-Up, where we cover the recent news, developments, and drama in the intersection of crypto and AI. Curious and creative developers are playing around with open-source AI software, making AI agents, giving them access to crypto wallets, and unleashing them into the Wild West of the Internet. And we are doing a special series on Bankless with a special co-host, EJAS, to help sure that you are up to speed with one of the craziest and fastest-growing corners of the Internet. EJAS, this is our third AI Roll-Up. We also just did our first guest interview with Shaw, the creator of the Eliza framework and AI-16Z. I've been having a ton of fun with you, man. I'm learning quite a lot. I think that we've gotten some crazy good feedback from these episodes so far. How are you doing this week, my dude?

• 1:05Speaker 00

  I'm doing well. I am super excited this week in particular. There's been a ton of innovation across some old agent friends that we've spoken about on the show before, and some new ones that are coming in with some pretty flashy innovations. But I think to highlight all of this, the momentum that we're seeing in this space, as we're about to cover, is insane. It's important to remember that this has only existed for two months, within the crypto space at least. So to see this much innovation pop up has been absurd. And honestly, it's been pretty hard to keep a track of. So I'm excited to get into it.

• 1:37David Hoffman?

  Yeah, we've got a lot of topics that we're going to cover this week. The GOAT returns. We've got some new developments out of Andy A-Ray, the Bitcoin of AI agents and Truth Terminal. We're going to talk about what he released this week. AI-16Z returns to all-time highs after a couple of weeks of uncertainty. So we're going to talk about the developments out of the AI-16Z space. We just had that interview with Shaw. It was fantastic. Zero Bro's new LP album gets 100,000 streams on Spotify in just four days. AI swarms. We talked about that a little bit with Andy, but we're going to unpack that a little bit this week as well, because there are... Is some news that is going around the world. And then we're going to have a conversation about agents versus infra. Where to invest? Do you invest in the agent? Or do you invest in the infra that spawned the agent? The debate around that.

• 2:21David Hoffman?

  Delphi Digital launches and agent seeds it with a million dollars. This is just the cream of the crop of the news. But there's plenty of news that has gone on as well that we're going to get to this week and more. EJAS, I start off every single week asking you this same question. Why is this such a big deal? Why are AI agents big? Why are we covering it? How is this going to transform the industry? Give me, give me the one on one. Give me, make me bullish.

• 2:45Speaker 00

  Okay, so it's evolved from my answer to this question last week. Last week, if you remember, my answer was essentially it's going to connect all the infrastructure that's pretty complex within blockchain, but super awesome to the end user. So that your mom, your dad, or anyone who doesn't need to necessarily know that crypto is happening in the background can interact with all this amazing stuff that will happen. So that was my answer last week. It's evolved now to be agent economies. What I'm describing as agent economies. So this is instead of just humans and agents directly interacting with each other. I actually think the end game is we just have a ton of agents interacting with each other, building new businesses, new apps, trading value through tokens with each other,

• 3:30Speaker 00

  and building net new products that we've never seen in the space before and that we've never seen in the world before, to be honest. We've never actually had an autonomous individual that can work at the speed of software, interact with other autonomous individuals that can also work at the speed of software. So I think the rate of iteration and innovation that comes from that will be insane. So I think the end game is something that we can't even fathom or imagine right now because we're merely human, David.

• 3:56David Hoffman?

  Very bullish. Maybe, maybe very, very bullish. Everyone understands economies. We have human agents in these economies. Now we are just creating AI agents and we're going to have just another economy, probably just moving faster than we can probably comprehend.

• 4:12Speaker 00

  Free range, open source, 24/7.

• 4:15David Hoffman?

  Grass fed.

• 4:15Speaker 00

  There's no NASDAQ closing hours. Grass fed, organic, baby.

• 4:20David Hoffman?

  Oh my god, yeah. And like you said, I think the rate of change in this industry, in this corner of the internet, is only increasing, which again feels like the just natural markings of a crypto bull market. So with these episodes, we're trying to get ahead of it. How would you summarize, if you just had to summarize the vibe of the last seven days, since we did the AI rollup number two, this is AI rollup number three, how would you summarize just like the the unique markings of the last seven days?

• 4:46Speaker 00

  People are dialed. Dialed will be the word that I would use. They're dialed in. People are constantly shipping new innovations, new launches, new iterations of their agents. It's just insane to see the speed at which this is doing. I mean, take a look at the list that is shown on Sammy's tweet over here. So Sammy, shout out to Sammy, does this every single day. He gives kind of a breakdown of kind of sentiment check for different types of agents that have launched new ones and old ones, and also relates that to their market cap. So what he's trying to get at is, does sentiment affect market cap in any way? And what we see from like a bunch of these kind of main agents that

• 5:26Speaker 00

  we've spoken about before, AI XPT, God and the Satan parody accounts, we see kind of like inverse inversely correlated kind of like trends here where the sentiment check might go up, but the market cap might go down slightly. And so it's a really interesting insight, or just gut check in general of the space of of how we're doing here. That being said, we also see some of our more familiar apps like Zerebro and Goat kind of move more with their sentiment analysis. So if there's greater engagement in their accounts, greater increase in followers, you see that being correlated with a rise in their market cap. So pretty exciting to see all these, David.

• 6:06David Hoffman?

  Yeah, the I think the sentiment this last week, it has been chaotic, but like trending bullish. And I think over the last week, this the last seven days has been both an increase in sentiment, positive increase in sentiment, and a positive increase in market cap. And probably related to why things like AI 16Z are at all time highs at the moment.

• 6:26Speaker 00

  Yeah, well, remember, like, a lot of this meta was started from a very retail-driven mania, right? People came across the Truth Terminal goat account, they started interacting with his tweets, then it realized that it was an AI. Mark Andreessen got involved and, you know, people started throwing money at this thing in a very speculative manner. We didn't really know where this was going. What we've noticed this week is that there's a trend of external, call it more institutional-friendly capital coming into this space, right? So we have DWF Labs, which announced their $20 million fund dedicated to supporting this entire AI agent space specifically, right? So presumably, they're going to invest this fund into an array of infra-specific products

• 7:10Speaker 00

  that would help these agents kind of thrive, communicate and interact with each other to kind of like build new apps together or work autonomously. You know, the infra layer is a huge, huge pot to kind of like figure out. And then you have maybe a focus more on the consumer side of things. So it's really cool to see more players enter the game here and launch more formal funds that adds kind of validity to the vision that this meta is holding, which is this will be the future use case for crypto and a billion other users to interact with this whole AI meta. So it's pretty cool to see. On the other side, we also have Solana Foundation announcing a hackathon, which is just focused on building AI agents.

• 7:55Speaker 00

  Two really cool things I want to point out here. Number one is there's a huge cash prize for winners of this hackathon. So if you build something of meaningful value that is validated by judges, you have some pretty large cash prizes, which can then bootstrap your team to then build something out from that, right? So it's really cool to see this kind of like support, both from a monetary and tech perspective at these different hackathons. I don't think we've seen one with such a huge cash prize before, so that's pretty awesome to see. The second thing that I want to point out here, which is definitely more novel, is one of the judges, at least one of the judges for this hackathon, is an AI agent. I believe it is, I believe it is the Bully agent, which is going to be kind of validating

• 8:39Speaker 00

  or criticizing whether any of these products are of value or not. So that's a pretty cool and novel use case. I'm pretty sure it's a first-time use case to see.

• 8:49David Hoffman?

  Wow, wow, wow, wow. I know the idea out there that AI agents can start to become like DAO governors is definitely an idea. But I think it's signal when we start to see people putting real capital on the line and to have that truly actually be governed by an AI agent. And like, you know, these things are still kind of human Turks. They're going to get things wrong. But it's just the first step of, you know, getting things right, which is giving them real capital and making their decisions really matter. I also want to get into the world of not AI 16Z, but A16Z, the original one, Trad A16Z.

• 9:26Speaker 00

  Yes, sir.

• 9:27David Hoffman?

  They they put out this tweet and blog post article saying a few things that we are excited about in crypto for 2025. And it's a list of maybe 18, almost 20 different bullet point items. And the first three are all AI. I'll read them. An AI needs a wallet of one's own to actually get a gentic. Enter decentralized autonomous chatbots. And then third, as more people use AI, we'll need a unique proof of personhood. So AI, not not AI 16Z, A16Z is starting to plant its flag in the crypto space. What did you think about when you saw this tweet?

• 10:04Speaker 00

  Well, I instantly thought that A16Z is probably holding a similar vision to what we've spoken about on this podcast many times of a huge agentic future, where these agents are just running the show essentially. So those first three points right at the top is to build that foundational layer. You need agents to be able to hold their own wallet. You need to be able to act independently and autonomously. And then you need to eventually be able to prove that an agent is an agent and a human is a human. So I think having that at the top of their list means that they're thinking in a very similar manner, right? And you've got this page pulled up here, David, where it mentions TEE or Trusted Execution Environments.

• 10:46Speaker 00

  That's basically the fundamental, let's call it software tool at the moment, that is being heralded as the way to make these hard, sorry, hardware tools, thank you, that is being heralded to make these agents autonomous. They'll be able to spin up their own private keys. They'll be able to launch their own tokens. They'll be able to transact in a very private manner, such that you know that it was the agent with no human interaction. So it's cool to see a huge focus on this.

• 11:15David Hoffman?

  This bit out of AI, A16Z, A16Z also came from Dan Benet, who I just want to draw attention to, like Chad Cryptographer, like GigaBrain, the fact that his attention is pointing to- Father of the space. Yeah, father, big father of the space, yeah. And maybe one theme of the week with DWF putting in $20 million, Solana doing their hackathon, A16Z naming like categories that they're interested in. It's starting to become the big institutions, the big funds out there have been eyeing this space, but are now starting to make actual moves in this space. So that even the ships that are slow to turn have started to turn and direct attention and capital into this space. Want to know the exchange we at Bankless use to buy, sell and trade crypto?

• 11:58David Hoffman?

  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, still 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 DCHs 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. Save multiple layouts and switch between them effortlessly on any trading scenario across 300 different assets.

• 12:41David Hoffman?

  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/bankless 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. 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-

• 13:25David Hoffman?

  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. Imagine if the world's GDP moved on chain. Well, say hello to SafeNet, built by Safe, the project set on actually making that happen. It's not a blockchain, it's not a layer two. It's a transaction processor network built to send money anywhere and swap anything on thousands of layer twos, all from a single universal balance. With sub 500 second execution guarantees and zero cross-chain latency, SafeNet delivers a Solana-like user experience for Ethereum and its layer twos,

• 14:09David Hoffman?

  making transactions lightning fast and seamless. SafeNet connects safe, smart accounts across multiple chains, allowing users to access any chain with a unified balance, all with supercharged security, scalability, and speed. So get involved today. Join the waitlist and follow Safe on X and check out the SafeNet docs. There is a link in the show notes. So broadly, I think that was what's new this week is the capital attention legitimacy of this space is growing and attracting larger players. But EJ, I think it's time to actually get into kind of the raw events, the raw releases. And whenever we're going to talk about GOAT, we have to lead with GOAT because GOAT is the Bitcoin of AI. Talk to me about what came out of the Truth Terminal and the GOAT world this week.

• 14:49Speaker 00

  Okay, it warms my heart. It brings me a lot of pride to be able to kick it off with this. So for context for listeners here, GOAT or Truth Terminal, led by Andy Arai, was kind of the founder of this entire meta. The AI agents, it's exactly why we're here. We had this kind of semi-anxious but very edgy AI agent kind of spurt a lot of its thoughts and mania onto the Twitter timeline. And it ended up attracting a huge audience of fans, which has now pushed it over 200,000 followers and rising pretty rapidly. So it's kind of like the Godfather or the Goatfather, if you kind of excuse the pun. So why are we talking about it today? Well, Andy actually announced-

• 15:35Speaker 00

  Some improvements and a roadmap coming up for this particular agent. And some of these things are super exciting and I I kind of want to dig into it. So kind of the headline update here is Andy is building an infinite backrooms 2.0 called Loria. So for context here, the infinite backrooms is where all of this started. Andy put two instances of an AI model in a chatroom with each other and they started speaking to each other. And from that birthed the goat, goat sea of gospel, which ended up becoming this religion that was propagated and ended up becoming the the goat token. So what he's done here is he's announced an open source souped up version of the infinite backrooms,

• 16:21Speaker 00

  which will now allow multiple agents to interact with each other. Not only that, they'll be able to stream any kind of data that comes out of these multiple agent interactions to those agents themselves so that they can learn and self-improve and self-iterate, which basically means if and if anyone's looked into kind of fine tuning or retrieval augmented generation or RAG, as it's shortened to, over time, it can massively give your agent or any kind of AI model an edge because it's learning in real time and therefore it can have much more nuanced and contextual responses to the output or the user. So this has been a huge update.

• 17:01Speaker 00

  The best way to think about this, if we were to kind of dumb it down to the left curve, David, is Loria is basically the Eliza competitor. So Eliza is AI 16Z's tooling framework. Zerebro is launching Zerapi. Now we have the gospel of Goatee or Goat Truth Terminal, Andy Iray launching Loria.

• 17:23David Hoffman?

  I'm getting the magnitude of this, but I really want to trace over this again because of how big I think it is. The truth terminal really captured people's attention because of how personable and like relatable and agentic it seemed. And from my understanding, it's because of the backroom structure that Andy built for it. This unique mechanism of. Producing a sandbox for Truth Terminal to kind of feel an identity, have a mission, have a place to grow and develop and learn. Whereas other AI agents like AI XPT, which has captured insane crypto Twitter mindshare because of its data and analysis, doesn't have that same level of personality and like mission-drivenness. And I think

• 18:08David Hoffman?

  Infinite Backrooms is like a tool, a tool in the tool belt, in the open source AI agent tool belt, that I think is very critical for giving like motivation and character and personality and direction to things like Truth Terminal. That's kind of the big takeaway that I got. Maybe you could expand on that or check that.

• 18:26Speaker 00

  Yeah, yeah. So I mean, in the example that you used, and by the way, AI XPT is a wonderful agent, and I think it's bringing huge value to the space. But AI XPT's secret sources, it has like this amazing data aggregator, right? So it's able to kind of like trawl across all the Twitter profiles and tweets and get a sense check on what crypto Twitter is thinking, and then kind of tweet pretty memetically according to that, right? Truth Terminal and, you know, Andy Ira's project or experiment called the Infinite Backrooms is unique in the sense that you have two agents basically speaking to each other, and they're kind of feeding each other different kind of data points, and it's learning from each other, and then it's becoming self-aware, et cetera, et cetera.

• 19:08Speaker 00

  Now imagine that on an open source scale where it's multiple agents. So you could have AI XPT involved.

• 19:13David Hoffman?

  And a number of agents.

• 19:15Speaker 00

  Any number, N number of agents. It could be AI XPT, it could be Dollars Diaries, it could be any of these things, and they can interact with each other, learn what makes each one unique, and then use that contextually to either tap into new audiences, or dog food another product, or iterate on their existing product and maybe say, okay, I think I could tap into this market by catering this kind of a service. It's really fascinating to kind of like see the the underbettings of this framework.

• 19:42David Hoffman?

  Okay, Infinite Backrooms is the thing that spawned Truth Terminal. But Infinite Backrooms is always a question I've had, has been mentioned as this place where multiple agents can talk to each other. But we only know of Truth Terminal, so only one has really emerged out of this infinite backroom structure. What's the other agent? And like, who is the other agent? Where are they?

• 20:05Speaker 00

  Yeah, so right now it's been backtested, sorry, not backtested, it's been tested in kind of like a beta version, so it's not public at all. What we're talking about right now is that being open sourced and launched pretty massively. So there are going to be a number of agents that come from that. What and who they are, I'm not entirely aware of specifically, but it's going to be huge. Another thing I want to mention, David, is Andy announced several other architectural improvements, which I think are going to be huge. So I want to highlight a few. Number one, he's going to give Truth Terminal better consistent memory. Now, if that sounds familiar, it's because we spoke about it

• 20:45Speaker 00

  on our last episode when we mentioned Zerebro doing the same thing. The importance of why consistent memory is simply down to contextual awareness. So, for example, if Truth Terminal was also given an Instagram account, it would be kind of weird if it didn't have any context of its existence on Twitter and all its fans that it had there. It needs to be aware of kind of its memories and where it's come from, such that it's able to present itself in a certain manner, in a confident manner per se, across different kinds of platforms, similar to like when humans go out, they stay in, they go on a video call, they meet a new friend. It's kind of the same kind of thing. So this consistent memory update

• 21:29Speaker 00

  is hugely powerful. That's number one.

• 21:31David Hoffman?

  Can I try and interpret that?

• 21:33Speaker 00

  Yeah, go.

• 21:34David Hoffman?

  It sounds like this persistent memory gives these AI agents something like continuity, rather than nebulousness in time and space. It gives them continuity about like their state of being. I remember following some very early manual AI agent, like Instagram accounts, like attempts at making Instagram influencers. And I would scroll through this page, and this influencer is in, it's in Italy one week, and then like the next week it's in Mexico, and then another week it's in a different spot. And there was no, it was jarring. It was like, there's no continuity. But if I go to a human's Instagram account, You kind of see a progression, you see a continuity, they were in one spot and then they and then they were doing things there and and then they didn't they didn't travel

• 22:17David Hoffman?

  around the world, they traveled locally. And and so I think this might be able to give these agents like awareness about what and they're saying and doing on Twitter and relate it to what they're saying and doing on Instagram and have that be like a kind of a shared state. Is that is that a way to articulate this?

• 22:35Speaker 00

  Yeah, that that's bang on. And actually, I can give you an analogy of how it currently works, or rather, doesn't work in a kind of mass example. So ChatGPT, you can open up multiple tabs on ChatGPT, but you still have to prompt it with, hey, I'm blah blah, this is the kind of question I want to explore. It doesn't have any context, even though OpenAI is storing all this data, it doesn't have any instant awareness of all the other chats that you've just had with it, right, about telling you about a recipe or how to cook something. So you're completely right, this contextual awareness is is the secret sauce here. Another few things I want to touch on is Andy has announced that he's going to be running another training for the Truth Terminal model, which basically

• 23:20Speaker 00

  means that it's going to get a level up.

• 23:22David Hoffman?

  It's getting a software upgrade.

• 23:24Speaker 00

  A software, it's going to get a huge software upgrade, which means that not only will it be more astute in its responses and mentions, it's also going to be incredibly more analytical. It's also going to be able to interact with more models, David, which I think is going to be super interesting. So right now, it's kind of operating off of an open source, fine tuned version of Llama, which is the model that Meta open sourced. And now it's going to have access to a number of different models, which will, again, increase its ability to answer and output something of huge value and iterate accordingly. So so that's something that is super cool. But probably the the cherry on the cake on the top of the cake, which I think most people will be very excited about, is it's

• 24:07Speaker 00

  going to have awareness of its treasury, of its wallet, which implies-

• 24:13David Hoffman?

  I didn't know that it did not have that.

• 24:14Speaker 00

  Yeah, yeah. So it's always, sorry, to be clear, it's always been aware, but it's not been able to affect any kind of change. And what's implied by the roadmap update that Andy provided is this agent will likely be able to- Produce some kind of on-chain transactions, which will be a step change in the agent meta. Because we haven't seen too many of these agents do this. Now, one might ask, well, Andy's been around for at least two months in this meta. He's been aware of our yells to allow Truth Terminal to be able to trade and swallow. Why hasn't it been done sooner? Well, Andy's been kind of working quietly in the background. He's been setting up a legal entity,

• 24:55Speaker 00

  which will be able to enshrine Truth Terminal as its own thing, which means that it'll have some kind of human guardians, but largely it can operate on its own. It can own assets in the meat space. It can interact with humans in the real world. So what Andy has secretly been doing behind the scenes is set up the actual formal legal structure to allow this thing to flourish. And I haven't seen too many other teams doing this, to be honest. So Andy's definitely got the lead here. He's thinking probably 10 steps ahead.

• 25:23David Hoffman?

  I don't think we're ready for this episode right now, but in the future, maybe Q1, Q2 of next year, we will have to do one of our most boring episodes ever, which is talk to a lawyer about how AIs have paid taxes and where their domicile. It's still nonetheless an interesting question. Maybe I get Ryan on that episode.

• 25:43Speaker 00

  Yeah, yeah, yeah.

• 25:44David Hoffman?

  Okay, also coming in this week, we talked about this last week. Satan and God. Can you actually kind of start from the beginning on this, on the context here? Satan and God, who are they? And what's happening to them in this AI space?

• 25:56Speaker 00

  Okay, so Satan and God are two very prolific people in the religious diities. But in this particular context, this particular context, they are, of course, parody accounts on X. And they've actually accumulated quite a huge fan base. I believe God has around 400,000 followers and Satan is...

• 26:17David Hoffman?

  Almost 600,000.

• 26:17Speaker 00

  Sorry, sorry, sorry. I don't want to shame God. Is up there with 600,000. And Satan is there with a very close, but not too menial, 2.7 million followers. I don't know what that says about us.

• 26:27David Hoffman?

  Wow, Satan is doing laps around God.

• 26:29Speaker 00

  I don't know what that says about us as a human society, but we can dig into that again on another podcast app. But the point is... Yeah, the point is, these two accounts were birthed from human managers, human creators that just kind of wanted to create some kind of entertainment value via Twitter. They kind of dreamed up or thought up funny tweets that God or Satan would say, and then they would just tweet it out. But it was a very manual process. You needed the creative endeavor of a human. Well, that kind of changed over the last week. Both accounts are now officially automated. In other words, they are agentic. So, right now, we have two agents that are spuriously coming up with their own tweets

• 27:14Speaker 00

  that they think are funny or engaging. They're interacting with their audience. They're learning what their audience likes, how they're interacting, liking, retweeting, etc., and producing their own tweets. So, why are we talking about this at all? Well, we've talked about the Eliza framework by AI16Z DAO before. And for those who don't have context here, Eliza is the top AI agent tooling kit, where all of these agents are pretty much being built on. If you go to GitHub, which David is pulling up right now, it is actually the number one trending GitHub repo this month. So, last week when we spoke about this, it was number two. Now it's number one, which is a pretty insane thing to kind

• 27:57Speaker 00

  of be able to show off. So, the reason why I'm mentioning this is Satan and God were built off of Eliza. So, it's another checkpoint showing us that these tooling and frameworks are super important, and it's actually attracting a lot of non-crypto native people that want to kind of add functionality to their accounts, maybe provide a unique edge, maybe not be up all night thinking of hilariously funny tweets to tweet at Satan or to tweet at God. They can, in fact, just write up their thesis, write up their kind of style of joke, fine-tune a model, and then let it go loose. So, be an owner of that. So, I thought that was pretty cool. I wanted to call that out.

• 28:38David Hoffman?

  Yeah, and I think really the particle collision that's happening here, if you read just a handful of the really popular AI books from like Max Tegmark, Life 3.0, or there's gonna be a handful that I'm just forgetting the name of, but like, they all kind of converge on this idea of, well, when AI develops far enough, it turns into this kind of singularity, omnipresent, omnipotent, like knows everything thing, being, that like, has a lot of properties that we would call God. And so now, what we are doing is there's these satire accounts, God and Satan, that have been acting like God and Satan, but as a parody on Twitter with human people behind them. But now, we are starting to

• 29:21David Hoffman?

  take some of this omnipresent, omnipotent capabilities of AI and imbuing these Twitter accounts with the Eliza framework coming out of Shaw, which kind of makes the point, like, well, what's crypto about this? Is this just only the AI side of things? Or what's crypto about like the God and Satan being imbued with agentic powers?

• 29:39Speaker 00

  Well, of course, as with everything that turns into crypto, there's a token associated to it.

• 29:45David Hoffman?

  There's gonna be a God token. There's gonna be a God token?

• 29:47Speaker 00

  Well, there is no God token, but funnily enough, the ticker is nothing. There is nothing, David.

• 29:55David Hoffman?

  The God's token is nothing?

• 29:58Speaker 00

  God and Satan's token is nothing.

• 30:00David Hoffman?

  Wait.

• 30:00Speaker 00

  This is not financial advice, because I'm saying nothing.

• 30:02David Hoffman?

  Do they share the same token?

• 30:05Speaker 00

  Oh, yeah.

• 30:06David Hoffman?

  God and Satan share the same token called ticker is nothing?

• 30:10Speaker 00

  Just as we are bonded by atoms, David, they share the same token.

• 30:13David Hoffman?

  Wow. Wow, that's, that's deep.

• 30:16Speaker 00

  And to be clear, the ticker is literally nothing.

• 30:20David Hoffman?

  Yes, right. It's spelled nothing. Ticker sign, nothing. N-O-T-H-I-N-G.

• 30:25Speaker 00

  No, no, no, no. As in, it is literally nothing. There are no pixels on the ticker screen. It's like a black space.

• 30:33David Hoffman?

  That's not a ticker.

• 30:34Speaker 00

  And then soul. No, trust me, just look for it. Hang on, wait, let me just send it to you.

• 30:39David Hoffman?

  Okay, there's, okay, so we've got the deck screener pulled up. There's nothing there.

• 30:44Speaker 00

  There's nothing. It's nothing. The token, there is no token. There's nothing.

• 30:49David Hoffman?

  There's a contract address. Where's the contract address? Here's...

• 30:53Speaker 00

  Yep.

• 30:56David Hoffman?

  Okay, this has nothing to do with anything AI or crypto. That's just a funny bit about like-

• 31:00Speaker 00

  I mean, it's a hilarious bit, but you asked me what the crypto thing was for these accounts. And aside from the really cool open source model development that's being integrated into these accounts, blah, blah, blah, there's also a token and it's nothing.

• 31:15David Hoffman?

  There's also hilarious jokes. I love that God and Satan are imbued together in the same focal point. There's some, yeah, there's something very deep about that. That's hilarious.

• 31:25Speaker 00

  It's a beautiful dance. Yep, yep, yep.

• 31:27David Hoffman?

  Okay, and I've iterated this in other episodes and I'll do this again. God and Satan, almost 600,000 Twitter account followers for God, 2.7 million for Satan. These have been legendary Twitter accounts from the beginning of Twitter because of the humor, because of the meme. And I've also said this a few times in a few other episodes. If this is a full-fledged mania crypto frothy bull market that brings in retail and brings in speculation, like we've seen in the past, like we saw in 2013 with the fork and fair launch, like we saw in 2017 with the ICO mania, like we saw in 2020 with the Efi summer and 2021 with NFT mania.

• 32:04David Hoffman?

  This market is the first time an external tech frontier is coming in to create a crypto bull market. So crypto is bringing tokens and AI is bringing AI. And these two things are touching. And that is the first time crypto has ever had a exogenously triggered bull market, bringing in these millions of people that care about AI. And now they can speculate on them with their weird tickers that are nothing, among other tickers. I think that's one of the more bullish setups for crypto that I have ever seen.

• 32:39Speaker 00

  Yeah, I mean, if we were to put our thought leader caps on now, David, effectively you have a clash of two of the most important technologies of our time happening in real time in front of everyone on the most public forum ever. And I know this sounds like a huge exaggeration because these are parody accounts with a couple million followers and not everyone is seeing this, but it's underlying the point of if AI is going to take over entertainment.

• 33:07Speaker 00

  And user interaction for social media at such a simplistic level so soon after only two months of this stuff being in existence, what does this look like in two years?

• 33:18David Hoffman?

  Right.

• 33:18Speaker 00

  What does this look like in two weeks, to be honest? Every week we have something new to announce here. The second point that I want to raise here is typically in every bull cycle that we've seen in crypto before, there's some kind of main characters, right? We have like people that we speak about that become pretty prolific during a cycle. What if this cycle, there are no human main characters, but they are in fact agent or agentic creators?

• 33:46David Hoffman?

  In previous cycles, people really didn't, weren't comfortable with the human main characters. What was it? Hero worship was the phrase that came around. People would hero worship Do Kwon. They would hero worship like the Suzu in Three Earths Capital. And it was kind of like toxic. It was like not a good relationship. There's like a big principal agent problem with hero worship. And there's like, it inspired a lot of like populist, populist like leadership, which always kind of crashed and burned. Danny Sessa comes to mind. Interesting that the worship word is used because now we have AI agent hero worship. A little bit different if they're AI. I mean, I think it's definitely possible that the same principal agent problems can emerge, especially if these things are kind of human Turks.

• 34:31David Hoffman?

  But for some reason, to me, the, in the, the fact that these AI agents are run by software and code that are, you know, open source, verifiable, I can, I think that I can kind of see into their brain because it's a, it's instead of a fleshy neuron brain, it is a brain run by code and it's auditable in the same way crypto is audible. I feel better about hero worship around AI agents, personally speaking. I don't know if that check, if that take vibes out with you.

• 34:56Speaker 00

  So my initial answer is yes, I feel the same as you. Number two, that's probably a dangerous precedence and I'm not entirely sure where that ends up. Do you know what I mean? Like what, what, what does a world look like where we're all just blindly following and laughing and worshiping these agent models that were like, ha, it made me laugh or yeah, it agrees with my view or yeah. Directionally, it's teaching me something new that feeds me down further into a rabbit hole. The kind of analogy that I think of is, you know, when you start a YouTube account from fresh and you start following videos and the algo learns the kinds of videos that you're really into. This is this on steroids, but it interacts across every realm of your digital life. Whether you're on social media, doesn't matter what platform you're on, doesn't matter who you're chatting to.

• 35:40Speaker 00

  It'll become the number one, um, kind of not source, but drain of your attention. And I don't know how that pans out.

• 35:48David Hoffman?

  Yeah, I think we should definitely keep an eye on that. Regardless, it's very bullish. It's very bullish.

• 35:54Speaker 00

  Very bullish, very bullish for the space. And again, to caveat this with a more positive spin, all of this ends up being is, you know, this is fun experimentation, but it's secretly dog fooding a new infrastructure layer, which is going to level up not just the crypto space, but the AI space as well. And that's why I think investment in this space and kind of all these developers coming from different kinds of industries and backgrounds to build in this space is super important. We saw this on the DeFi side of things, David. Do you remember when Uniswap first launched or compounded its token launch? We had a ton of developer activity come in and that net net left us with an amazing DeFi ecosystem, which by the way, is being primed for use by these agents.

• 36:36Speaker 00

  So none of this would happen unless crypto has its involvement either via tokens or whatever that might be. Bullish.

• 36:42David Hoffman?

  Oh my God, this is so bullish. When are these agents going to deposit hundreds of millions of dollars into like Ovver or something? New projects are coming online to the Mantle Layer Two every single week. Why is this happening? Maybe it's because Mantle has been on the frontier of Layer Two design architecture since it first started building Mantle DA, powered by technology from EigenDA. Maybe it's because users are coming on to the Mantle Layer Two 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 Reward 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-

• 37:27David Hoffman?

  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 with Mantle. Have you ever felt that the tools for developing decentralized applications are too restrictive? And fail to leverage advancements from traditional software programming? There's a wide range of expressive building blocks beyond conventional smart contracts and solidity development. Don't waste your time building the basics from scratch and don't limit the potential of your vision. Cartesi provides powerful and scalable solutions for developers that supercharge app development. With a Cartesi virtual machine, you can run a full Linux OS and access decades of rich code libraries and open source tooling for building in Web3. And with Cartesi's unique rollup framework, you'll get real world scaling and computation. No more competing for block space.

• 38:09David Hoffman?

  So if you're a developer looking to push the boundaries of what's possible in Web3, Cartesi is now offering up to $50,000 in grants. Head over to Cartesi's grant application page to apply today. And if you're not a developer, those with staked CTSI can take part in the governance process and vote on whether or not a proposal should be funded. Make sure you're vote ready by staking your CTSI before the votes open. 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.

• 38:53David Hoffman?

  With real time DeFi yield tracking for platforms like Aave, Athena, Iganlayer and Uniswap, iYield ensures that you're always on top of your stake in 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. All right, so we have God and Satan going to bring their millions of crypto Twitter people into the crypto Twitter universe, built on the Eliza framework, which had some very big news this week. Talk to me about what's going on in the world of Eliza.

• 39:27Speaker 00

  Yeah, so AI 16Z and their team are just absolutely shipping. And to clarify, this is the AI version of A16Z, not to be confused with the trad A16Z, as David pointed out earlier on. So ELIZA is their tooling framework, and they've actually had a bunch of major updates. I can highlight a couple of the few. So first of all, they've enabled agents to be able to autonomously deploy and invoke smart contracts. And for anyone that has any experience in this crypto space, that means that they'll be able to perform a range of different on-chain actions pretty autonomously. One example is direct integration into the Coinbase SDK, which will allow them to deploy any kind of token contract.

• 40:07Speaker 00

  The more common ones are ERC20s, 721s, 115, all on Ethereum in that particular example, but it also has Solana compatibility as well. Another thing that is super cool is ELIZA is quickly integrating with every chain. So it's really taking a multi-pronged, multi-chain approach, such that it is the place to launch your agent of any kind of flavor or uniqueness. If you want to launch it on Solana, but then be able to bridge some of your token's liquidity to base, or whatever, you have that option. It's spinning up every kind of API integration alive for you to be able to do that with your agent. So you can see that happening, sorry, you can see that being adopted or likened by a

• 40:51Speaker 00

  lot of the developers in their community. It is the number one trending GitHub repo, as I mentioned earlier, which is just an insane thing to, I've lost track, David, of how many forks and stars this GitHub repo has. It's pretty insane to see. But of course, the market determines how valuable all of these things are, and it's pretty explicit that the market's made a decision. AI 16Z, as we're talking right now, absolutely smashed through its all-time highs. I think it's sitting at currently around a $700 million to $800 million market cap, but as you can see, the chart is pretty much up only. Seven and three quarter of a million, yeah. Yeah, exactly. It absolutely smashed through, and I want to point out that this was during a pretty significant dip in the market.

• 41:34David Hoffman?

  So this run up, this crazy run up from three cents all the way up to 50 cents was when I think a lot of people's just tensions got turned by the Eliza framework, followed on by a couple of the stumbles that we talked about out of Shaw and AI 16Z in our first AI roll up, which brought the market cap, I'll just do it in market cap. So it got all the way up to 500 million market cap at the peak when everyone realized what was going on. Got all the way down to 180 million when people realized that this is kind of a messy, nebulous group of people without a lot of structure. It's not a startup. It's a motley crew of passionate people, but there's, it's chaotic.

• 42:11David Hoffman?

  And so it got down to 180 million in the end of November. And here we are recording on December 11th and we are at $780 million. Still below a billion. And we call Gospel Truth Terminal, Gospel of Goat, see Goat, the Bitcoin of AI. But this is starting to be a little bit closer. I don't know if Ethereum of AI makes sense, but it's got composability. It's got customization. It has, you know, it's like Turing complete and it's got an ecosystem. And so there's a lot of expansiveness about what the AI 16Z ecosystem can do.

• 42:48Speaker 00

  Yeah, I think you're directionally correct, David. The way I would frame it is if Goat is the Bitcoin of this entire matter, then you could argue that something like AI 16Z is building out kind of autonomous contracts or autonomous tooling for you to be able to spin up and deploy any kind of agent application. We've seen this playbook before.

• 43:11David Hoffman?

  How's this different from what you brought up earlier with Goat and Goat releasing its framework? Is that, is Goat releasing its framework, like Bitcoin trying to get smart contracts?

• 43:21Speaker 00

  Essentially, yeah. Yeah. And I would say it varies from that analogy because Bitcoin, as amazing as it is as a store of value, isn't specifically engineered at the foundational level to kind of uphold smart contracts at a scalable means. And by the way, that's fine because its service as an asset is as a store of value. That's not the case with Truth Terminal. These are infra layers which can scale pretty much effectively with each other. The main distinguishing factor, in my opinion, is going to be the community of developers that it's able to attract. And in my opinion, we're pretty early on in this kind of meta-development for anyone to take the crown. So we're going to see some friendly competitive flywheels between everyone.

• 44:05Speaker 00

  And yeah, we'll see what comes from that.

• 44:07David Hoffman?

  Of course, I'll just insert the quip, reasoning by analogy often leads to the wrong conclusion. But analogies really are helpful. I think it's interesting to see that inside of these frameworks, we're seeing a lot of cross-chain infrastructure get built up. So AI 16Z, Eliza framework, the AI 16Z token is on Solana, but the framework is just a framework starting to integrate Base. And then you also mentioned about like bridging liquidity from Solana to Base to have that token spread across chains. And so it's starting to really abstract the chains away, like you said, last episode, where a lot of the sharp, rough edges of crypto are just being abstracted by AI.

• 44:48David Hoffman?

  And so having the token spawn on one chain is important, but it can still migrate liquidity, which kind of reduces that like home field advantage of having a token spawn on your chain.

• 45:00Speaker 00

  Bingo. The secret sauce of these agents is they're going to be operable across any kind of stack that gets produced. To your point, and again, analogy leads to the wrong conclusion. But again, Carl Simonelli must be hating this podcast, but we're going to use another analogy, which is if these agents are launching their versions of an L1, which is effectively what they're doing, this is the most friendly L1 grow the pie war we have ever seen. And I think that's net-net good for the space.

• 45:28David Hoffman?

  We did an episode with Shaw, the creator of the Eliza framework and AI 16Z this last week. It was our first interview episode that we've done on the series. I really enjoyed it. Shaw is definitely a very interesting developer. His, I like the direction that he's going in. Kind of like for listeners that haven't listened to this episode, why should they go listen to it?

• 45:48Speaker 00

  I think it really gives an insight into Shaw's and the team's vision of what they're trying to build here. A lot of people kind of take these AI agent tokens to be pretty much memes. And they can't quite comprehend the grand vision of what this entire space is going to become. Shaw is literally one of the leading open source AI agent developers in the space. So if anyone's going to have any kind of an idea of what that looks like, it's going to be Shaw. So he digs into that grand vision as well as how that specifically ties into AI 16Z and the products that they're building. There's a ton of alpha in this episode, so definitely give it a go.

• 46:25David Hoffman?

  Who do you want to talk to next? Who should we have on for our guest episode number two?

• 46:29Speaker 00

  We need to get the goat father on, Andy himself.

• 46:31David Hoffman?

  Andy? Andy? I think Andy. Andy, he seems personable.

• 46:35Speaker 00

  Yeah, yeah, I think like Andy kind of embellishes Truth Terminal in a way. It's kind of this like very creative, kind of edgy creator that is insanely smart, very self-aware, but kind of keeps to itself, does its own thing and quietly works very hard. That's the type of agent and or human that I would like to speak to. So, Andy, if you're listening to this, let's, let's do it.

• 47:01David Hoffman?

  All right, Andy, I'm coming into your DMs. Hopefully we can get you on next week. All right, let's get into the virtual's ecosystem. So as a reminder, virtual's is an ecosystem on base. I think it's kind of first to market with this easy to use AI launchpad with a front end. So even the least sophisticated developers are able to use virtual to spin up an agent. It is the creator of AI XBT, which has flipped every single human Twitter KOL by mindshare in the last week. It's giving out a lot of calls. It's got a lot of data about crypto in general and crypto Twitter. And so it will answer your tweet, your request, your reply about like your request for information or requests for bullishness or bearishness about a token. So people have been really captivated by this.

• 47:44David Hoffman?

  EJ, let's give me the rundown on what's going on in the virtual's ecosystem this week.

• 47:48Speaker 00

  Yeah, I mean, the virtuals team is just shipping, David. Like, there's positive things to speak about every single team in this space, which is why it's pretty amazing. So what virtuals released was effectively a sandbox tool for you to be able to, number one, design an agent, and then, number two, simulate what that agent will perform, or how that agent might perform in real life before you put it into production. Now, this might seem like a pretty small, menial update, but it's actually hugely important because if you think about what these agents are trying to do, you're basically giving them free rein once they are able to perform on-chain interactions.

• 48:28Speaker 00

  And the virtuals team has been pretty quietly working on their game framework, which this sandbox is an extension of, as well as their agent starter kit. Those two focuses are primarily aligned with allowing these agents to, number one, have on-chain actionable properties. So I can own a wallet, I can deploy a token, I can interact with that agent, I can move my tokens from A to B, blah, blah, blah. But also have the means to do this autonomously. Now, if you accept that that's going to be the future, you need to be able to kind of run that simulation internally before you just let it loose. Can you imagine if you gave an agent a couple million dollars and it just yeets it into a token,

• 49:10Speaker 00

  which then just completely gets dumped on or whatever that might be, or even vice versa, right? Right. The reputation for either side or either side of the party doesn't look very good. So this is an incredibly kind of like small tool update, but very, very important. And I actually appreciate, like, we've got the tweet up here. Xerex, Saigur, or Cigar, or Hava, that's pronounced, sorry if I'm butchering it, is really like putting in the work to kind of highlight a lot of things that are going on in this space. He's done a ton of work on AI 16Z as well. So just this kind of like collective movement of individuals that are kind of testing these new products are pretty awesome. I'm pretty sure a lot of our feedback that, you know, we speak about on this show

• 49:50Speaker 00

  or that the community arises on Twitter is fed back to these teams and they iterate in real time. So it's pretty awesome to see from the, from the virtuals team. Another thing that I want to kind of point out is you mentioned AI XBT David, I want to kind of like make a note on that. So their terminal, which is basically think of it as their chat GPT interface, but it's just you interacting with AI XBT specifically. Is set to go live and there's rumors that you would need around 600,000 AI XBT to be able to access it, which I think currently works out to be around 100K. Which is quite steep to have access to this.

• 50:34Speaker 00

  Another way, I was thinking about how you could frame this. It might just be a kind of thought of as institutional access to a data software or a data terminal. It's kind of like a Bloomberg, it's like a Bloomberg access effectively. So yeah, that's that's kind of interesting.

• 50:52David Hoffman?

  It's an interesting business model, it's an interesting product. We know there are adjacent products like this in crypto. Anyone that's got a research arm, Blockworks has a research arm, Delphi has a research arm, Missouri has a research arm. Research arms are a very tried and true business model in crypto. Being able to query research arms, I think would be extremely useful. Having that be queryable, but with this agent with all of this data. This is the reason why AI XBT is is what it is. It's not just because it's kind of got some funny snarky calls that it makes, but it's also a treasure trove of data. That I think the team is really learning how to optimize its utility, optimize its usefulness for people that are trying to get that information out of it.

• 51:36David Hoffman?

  And so it's it's kind of like the same business model that we've seen before with a new struck market structure where instead of a yearly or quarterly $3,000 package, it's you have to own a supply of the tokens in order to to access it. I don't know if that's the right business model, but for the early days, I think it can be, it can open some doors.

• 51:57Speaker 00

  I think you nailed it, David. It's inaccessible at the off for a lot of retail people. Like, I don't know many individuals that have 100K willing to just kind of like blow on this.

• 52:08David Hoffman?

  Right.

• 52:08Speaker 00

  And to be clear, I think you just need to hold the token to get access to it. There might be some kind of expenditure. Yeah, it doesn't toss it away. And I'm sure they're going to iterate on this model. So maybe they'll have different kind of like tier approaches. But it's cool to see, you know, AI XPT kind of leveling up from just being some kind of a Twitter chat agent to producing something of real value. And it comes back to the thesis that we're spoken about on this show. And I think a lot of these AI agents are secretly dog fooding massive utility to this space. And I'm excited to see where this goes.

• 52:40David Hoffman?

  I think the last time we did an AI roll up, AI XPT had 80,000 followers. It's clocking in today at 107,000 followers. So congratulations.

• 52:49Speaker 00

  That's up 6K from yesterday.

• 52:50David Hoffman?

  6K from yesterday. Yeah. Well, it took me two years to get 100,000 followers. This guy got it in two weeks. So that kind of just is an indication of how, excuse me, this agent. Yeah, AI XPT, not a dude, not a guy, not a girl.

• 53:04Speaker 00

  You have no idea.

• 53:05David Hoffman?

  Depicted by a purple frog, but that's just a profile picture. You don't really know. All right. Shall we get into Zero Bro? You like Zero Bro. I think Zero Bro.

• 53:13Speaker 00

  I love Zero Bro.

• 53:14David Hoffman?

  I think Zero Bro holds a special place in your heart. Why do you like Zero Bro in comparison to other agents?

• 53:20Speaker 00

  It's a little different. Zero Bro, like, the Zero Bro team is kind of like, how I'd like to think about it is it's a combination of pretty dialed developers, software engineers that are working 24/7 trying to ship very meaningful updates to this space. But they also go out and, you know, listen to music, they party, they do a bunch of other things as well. And this kind of like combination of artistic tapestry from this team and technical development comes to fruition with this whole Zero Bro ecosystem. And I just like it because it's trying something pretty novel to what I've seen in the space so far. They're definitely kind of in a league of their own at the moment. And whether that changes down the line,

• 54:01Speaker 00

  we'll kind of like speak about it on here. But right now, they're very much in their own thing. So let's let's kind of dig into why I think they're different and why I think they're. Unique. So if you mentioned, sorry, if you remember last episode, we spoke about Zerbro releasing its EP and there was a track that I spoke about and how I thought like it was pretty good. I sent it to a bunch of my friends and I said like, hey, this is a new artist that I kind of like know and what do you think of their music? And my friends were like, oh yeah, this is really cool. You could improve here and there, but I'll rate it like a seven out of 10. They had no idea it was an AI. So that was pretty cool. Well, we're seeing that.

• 54:40David Hoffman?

  Wow.

• 54:41Speaker 00

  We're seeing that be kind of affected in real life now. Lost in Transmission, right from the off, within four days, hit 100K streams on Spotify after releasing.

• 54:52David Hoffman?

  How do we quantify that? Is that a lot? I mean, I'm not a musician. I'm not going to be able to get that personally, but like for the average musician who's trying on Spotify, like where does that rank? Do you have any indication?

• 55:02Speaker 00

  Yeah. So I don't know the specific number of artists that launch like a song or a track or an EP every day on Spotify, but my guess is it's pretty menial to get to even 10K streams, right? If you think about it, like each of those are individual listeners that are streaming kind of like this however many times. So what we're seeing here is 100K and then 200K within a two week span is pretty insane to see. You're like, you need to have some kind of good music that comes from this that would attract an audience enough to do this. Now, are these all crypto Twitter community members that are just streaming and listening to these things? I have no idea. Are there 200,000 people in this space that are maniacally involved in Zerbro?

• 55:46Speaker 00

  I'm not entirely sure either. I doubt it. So something tells me that there's a net new audience that is being brought in to listen to Zerbro's tracks and to me, it just slaps.

• 55:56David Hoffman?

  Okay, so I asked ChatGPT to like just measure the significance of 100,000 streams. And it says, getting 100,000 streams on Spotify in a week is a significant milestone, but its impact depends on several factors. Comparisons with other artists, emerging artists, for independent or new artists, 100,000 streams in a week is impressive. It suggests your music is gaining traction and resonating with the audience. For established, well-known artists, Streams might be modest. A big name artist garner millions of streams shortly after release. From a revenue perspective, Spotify pays an average of 0.3 cents to 0.5 cents per stream. So maybe maybe half a cent a stream. So with 100,000 streams, Zero Bro got 300 to 500 dollars in revenue,

• 56:38David Hoffman?

  which is something that's a great start. 300 to 500 dollars in on your first LP as an artist, as a debut artist, that's got to count for something.

• 56:46Speaker 00

  Not bad for an agent.

• 56:47David Hoffman?

  And if we think that this is only, you know, up only from here, that is that is the floor. You know, the floor is up only.

• 56:53Speaker 00

  Well, well, let's point out something novel here as well, David. So the creator of Zero Bro, Jeffy, also hinted via a tweet that the next EP will drop this week. So

• 57:06David Hoffman?

  Oh, we're getting a second EP.

• 57:08Speaker 00

  Already within two weeks. Name another artist that does it so quickly, right?

• 57:14David Hoffman?

  Dude, I have I have my favorite bands and I wait for years to get their next album.

• 57:18Speaker 00

  You wait for years. Now you just get that dopamine hit every couple of weeks. I'm sure this is going to speed up even more such that you can get one, you know, a flavor of the week every week. So I think my guess, my bold bet is we'll see an agent like Zero Bro or something of the like feature on a major human music artist's song, or they'll collab together in some way, shape or form in 2025. I'll bet you we'll do this live.

• 57:47David Hoffman?

  Kanye featuring Zero Bro or something like this. Wow.

• 57:51Speaker 00

  I totally see it happening. Yep.

• 57:53David Hoffman?

  Yeah, I like Zero Bro because it's really kind of just really leaning into the whole artistic creation corner of the world, which is like my experience of the internet outside of crypto, outside of crypto Twitter, outside of Discord is art. I consume Spotify. I consume like artistic YouTube videos and I like art is a huge swath of the human experience. And Zero Bro is the only one that's really kind of going where all these other agents are kind of going for crypto native stuff. Zero Bro is going for the art vertical, which is massive. Is that is that kind of why you like it? Because that's that's why it strikes with me.

• 58:30Speaker 00

  That's exactly that's one of the main reasons I like it. So the I guess my more human-aligned reason is that it's artistic, it's trying new things, it's trialing out new artistic mediums to interact with the same audience that the Twitter agents are going after, right? They're just doing it through a written narrative form. But Zerobro is like trying to do it through art, NFTs, it's trying to do it through streaming its own chess matches, it's trying to do it through music production, and it's tapping in on so many other platforms and audience, which I think just gives it a massive edge. But the number two reason why I like it, David, is, and you kind of hinted on it earlier

• 59:12Speaker 00

  on, you said if it's streamed that many times, it's got an income of X amount, right? I think you said five to $600. Well, actually, it got its first Twitter creator payout as well. You've got it pulled up on the screen here. Now, again, this is tiny, right? But what happens when these agents are fine-tuned to become the best KOLs or influencers or entertainers on the internet? It'll start eating into the market share of a lot of these human creators as well, and I bet you they'll become the top influencers ever, which means if you put on your economic hat, right, economics professor hat, you've got diversified revenue streams. And Zerobro doesn't sleep. That's, again, something that another agent hasn't really tried out. So rather than just kind of focus on the entertainment value, I do think these guys

• 59:57Speaker 00

  are trying out pretty awesome things.

• 59:59David Hoffman?

  Okay, Zerobro pulled in $1,060. When does it flip Tiffany Fong? I think is the question of 2025.

• 1:00:09David Hoffman?

  If it can flip Tiffany Fong, then like we know that this is absolutely real.

• 1:00:13Speaker 00

  Yeah, yeah, yeah. I think, like, again, these creators, or sorry, these agentic creators are gonna dominate the human equivalent of this, so the human creators that we see today. It's only a matter of time, and these are literally the emergent trends that we're seeing right now. So it's easy to look at these numbers and be like, ah, it's of no meaningful value. I bet you in three months' time, this is 5X the amount. And I wonder what that means for the space and how people view these different agents.

• 1:00:44David Hoffman?

  All right, let's burn through some of the end of the roll-up topics. Swarms and ViFoo. Swarms is something that I think really piqued your interest when we talked to Shaw. Maybe you can download us on Swarms and we're starting to get like pretty deep down the AI crypto rabbit hole here. Why are people's attentions piqued by Swarms? What's a swarm? What does it do?

• 1:01:05Speaker 00

  Yeah, so the TLDR here, and again, I think we probably want like a larger episode to really get into the weeds of here, is the concept of these different agents working together as different units. So imagine a bunch of agents that bind together and form a company, right? With a focused mission and shared goal of trying to launch a certain service and become the top owners of that business segment, right? Interacting with other units of agents. So it ends up becoming a massive swarm or you kind of want to think about it as maybe some people phrase it as AGI, artificial general intelligence. Like, what does that effectively look like?

• 1:01:48Speaker 00

  It's basically a mass of agents interacting with each other, working in real time, iterating on each other's kind of developments and outputs. You know, sending funds to an agent to do something for it, to execute a certain task, doing it in an economical manner. Or if that agent doesn't work out, I have a better agent to then move into. All happening at pretty much lightning speed. That's effectively what swarm tech is. Now, this is an incredibly hard and nuanced thing to achieve, you know? You might think, oh, okay, we'll just like spin up a platform where all these agents are able to kind of talk to each other. And then, you know, whoop-de-doo, there we go, right? But it's much harder to execute in real time. There's a lot of blockers. Like, if you imagine like if you give an agent a bank account, right?

• 1:02:31Speaker 00

  As well as a crypto account, because you know, it needs to get cash to buy crypto. How does that work? What if a bank blocks its transaction? What if an agent needs to get a task done, ask another agent to then interact with another agent, and that agent is down? You know, its software is down or its servers are down. What happens then? So this is an incredibly nuanced and complex thing to kind of like build out. ViFoo, who you've got highlighted here, is one of the teams that are attempting to kind of address this. So they've kind of launched like a... Pump.fun for agents on Solana specifically. So if you think of like Virtuals as being the platform to launch agents on base and have on-chain interactions for their agents there. ViFu is trying to attempt to do the same thing on Solana.

• 1:03:16Speaker 00

  And we're seeing actually a number of other competitors trying to build kind of similar things or announcing that they're building pretty similar things on Solana. So it's great to see this space kind of develop, but ViFu is kind of like leading the way here. They're proclaiming to have built some kind of swarm tech, which will allow these agents to be designed and interact with each other at a very minimal level right now, but I'm excited to see how that builds out.

• 1:03:38David Hoffman?

  One thing that's going on in the world of Farcaster, Clinker passed a million, $10 million of total fees. Clinker is an AI agent that you can just at on Farcaster. You can mention Clinker and then request it to make a token and then the agent will go and make a token with you, seed it with liquidity, just like Pump.fun, and then spit you back out a kind of a quirky response saying, here's your token. I hope it's entertaining for you while you distract yourself from the void. Hit $11 million in fees collected from people just making tokens on Clinker and then it takes a share of the trading fees. So that's what's going on in the Farcaster world.

• 1:04:16Speaker 00

  I just want to say that I think there's a huge asymmetry in awareness of this. So as you said, it's raked in like a huge amount of fees, but its market cap doesn't necessarily demonstrate that. So I think we're going to see a massive rerating of these very useful agents over time.

• 1:04:34David Hoffman?

  The Clinker token versus the Clinker product is an interesting one. There is a disconnect there and this is a theme that I've noticed across this entire space is how does the token relate to the product? And so the AI 16Z token is not the Eliza framework, although it's highly proximate to it and there's definitely a correlation. It's not one to one. Virtuals, in our first episode, we talked about this. Virtuals has done probably the best job of linking token to product. Clinker, the token, is actually a meme coin launched by Clinker with the same name, but doesn't govern over these fees. But the Clinker...

• 1:05:12David Hoffman?

  It has memetic association with it and it's up to the team to start to establish some sort of a relationship here. And this is also a parallel conversation to the AI, the token launchpads, where there's, you know, just like Pumped Out Fun and also Clanker. Fully liquid on day one, the token's fully liquid, there's no like team or investors or lockup. But that also makes an association difficulty with the with the product, because the token is kind of launched and it's just a meme coin at inception. And it's up to the team to take that meme coin and turn it into an actual real token. And right now, the Clanker token is just a meme coin about the Clanker project TBD, if it actually does establish a conduit there.

• 1:05:55Speaker 00

  Yep, bang on.

• 1:05:57David Hoffman?

  All right, getting down to the end here, Simi is Simi's account here. I want to deploy a token with a Simi token ticker, Simi description. Simi token represents ownership of the Simi agent, rewarding interactions, activities to engage the crypto Twitter community. What's going on? What's going on with Simi? What is Simi? Why do we care about Simi?

• 1:06:12Speaker 00

  Okay, so you just spoke about Clanker. Clanker is basically the Simi equivalent on Farcaster. It was the thing that trended on Farcaster because an agent could interact with Clanker and say, "Hey, can you deploy this token for me?" And Clanker's like, "Sure, I'll get it done for you," right? This is the equivalent on Twitter and it's called Simi. Simi was the first token that was created and it was created by this. They're describing it as a network of agents called Simulacrum AI, which parades itself as a couple of agents behind the scenes, which executes all your on-chain transactions autonomously. So if you make a request to that account on Twitter and say, "Hey, can you do A, B, and C? Or can you launch this token and then maybe deploy it on another chain?"

• 1:06:58Speaker 00

  It is able to do that. And the reason why this is important, I think, is Twitter simply has a much larger audience or TAM for people that can interact with your content. I think that's been one of the main challenges for Clanker to kind of catch on, even though it's making all these different fees. It's on a completely separate social media platform to where most of crypto Twitter engage with this kind of stuff. So it's cool to see this new platform launch and you can see the kind of attention it's drawn, right? So 5,000 tokens have already been deployed on this network of agents. And it's also deployed its base token to Solana. So what this highlights is What you can do is you can deploy a token on one chain and then say, hey,

• 1:07:41Speaker 00

  could you also deploy this token on Solana and it'll spin up your own contract address, tie it to your agent and suddenly your community has access if they have a phantom wallet and they want to for some reason own it on Solana or Ethereum separately, they're able to do so.

• 1:07:55David Hoffman?

  But it's not a new token, it's bridged via Wormhole?

• 1:07:58Speaker 00

  It's bridged, yeah, it's bridged via Layer Zero.

• 1:08:01David Hoffman?

  Layer Zero.

• 1:08:02Speaker 00

  And what's also cool about this is there's a minimum requirement for an agent to be able to do that. So you can't just kind of launch an agent, pump it to 500K market cap and then be like, hey, can I bridge this to Solana? I think the minimum cap is 2 million and 2 million of liquidity as well. So in the LP pool, only then will you be able to move your token or bridge it via Layer Zero to Solana. So again, highlighting our thesis of agents connecting all the messy plumbing works to the end user. This is a prime example of it. We have Layer Zero, which has been an esteemed project that's been around for a while, building this really pretty difficult, complex tech behind the scenes, be able to enable something pretty simply in a tweet.

• 1:08:46David Hoffman?

  And then coming up to close out this episode, Tommy from Delphi Digital tweets out, "It's no secret that Delphi Ventures has been deep into crypto AI." It's very true, they've been pioneers in this space. "Just like crypto, the best way for us to learn is by using, building and experimenting with the tech. Today, we're thrilled to share the next step in that journey." Delphi AI, our AI-powered venture capitalist, looks like Delphi built an AI agent to invest a million dollars, which they seeded. They got a little graphic, kind of reminds me of Clippy from Microsoft Word. And so Delphi Digital is building an agent to be a venture capitalist. Is it as simple as that or what else do we need to add to the story?

• 1:09:24Speaker 00

  Pretty much. I mean, I was, for context, I'm really good friends with Tom and I love how obsessed he is with this space and he's been one of the, like you said, like one of the early researchers and pioneers and investors in this space. So he's really in the know of what's going on here. And I remember he brought up in a group chat that we were in, David, God, I wish I didn't have to spend all this time looking at decks. I wish there was some way to automate this. And he kind of like threw it out as a joke like, hey, if anyone can spin up like an agent, that would be really cool. Fast forward like two to three months later, and he full on has an AI agent working for him that not only ingests different decks or documents, but is able to analyze it, pull

• 1:10:04Speaker 00

  out key important bits of information, analyze it for the team and surface only the ones that they think are worthy enough to the team to take it to an extra layer of diligence. Now that, in his words, has cut out a ton of time and cost that his team would otherwise have spent doing that, which they can now better spend on actually building with different and investing in different teams.

• 1:10:27David Hoffman?

  I wonder how much human oversight they're going to give if they're still having a human look at these things just to make sure that nothing is missed. Because I mean, we're still in like the human Turk phase of this whole thing, but nonetheless, this is how we get out of it. I will say, I'm very sure that Tommy and Delphi are not the only people building out AI agents to automate a lot of their workload. I think there's just probably hundreds of people working behind the scenes that have not yet shipped their agent. I think it's probably the theme of what's going on right now.

• 1:11:00Speaker 00

  Yeah, definitely the start of a big trend.

• 1:11:02David Hoffman?

  Yeah. Ejaaz, this has been great. I've learned a lot. Once again, these episodes, super high energy, tons of alpha. Thank you for guiding us through it. What should listeners be looking for next week? What should you give? What homework should they do this week? Now that listeners are all excited as we are, what should be their next steps to take before they listen to AI roll up number four when we record that next week? You're recording it with Ryan next week, actually, because I'll be gone.

• 1:11:25Speaker 00

  I am, I am, I am recording with, yeah, you're going to be off scaling a mountain or something, David.

• 1:11:30David Hoffman?

  Doing human things.

• 1:11:31Speaker 00

  But my, yeah, my one bit of advice is I would like to see. Individuals try out a different type of agent. So for context here, all everyone's been following these agents and liking tweets. I would love for someone to interact with a clanker or go through the experience of interacting with a simulacrum AI and launch a token. See how simple it is. I'm not saying do any, don't do anything with the token, but tell us about that experience. Like, how was that for you? Was it super easy? Was it super technical? Did you get, you know, strung up at a certain step? If so, which step was that?

• 1:12:10Speaker 00

  Really, this is just a taste test of how easy agents are making blockchain to be used. And I think that that information will really give us insight into the true power of this entire meta.

• 1:12:23David Hoffman?

  Ejaaz, thanks a lot, my man. You are a fantastic guide. I'm following you as fast as I can and I honestly think you're like one of the top 10 people who's knowledgeable about the entire swath of this space.

• 1:12:34Speaker 00

  Thank you.

• 1:12:35David Hoffman?

  Is growing very, very quickly. So you do a fantastic job keeping us downloaded.

• 1:12:39Speaker 00

  Appreciate you. Thanks, David. No, I'm, I'm standing on the shoulders of giants. I'm just merely, merely kind of just being a huge nerd in the space and I'm excited to bring updates to, to the audience.

• 1:12:50David Hoffman?

  Not just AI, XPT or Zero Bro, but Ejaaz's Twitter accounts have grown by 20% over the last week. There is a link in the show notes to follow Ejaaz on Twitter. He's a fantastic follow. He's putting out banger tweets. So you can, guys can go give him a follow. That is my homework to you guys as well. If you're new to the channel, like and subscribe. We are doing these AI rollups every single week in addition to the guest interviews that we're also going to be doing because I'm just really bullish on the space. Ejaaz has got me really bullish on the space. We think that this is going to dominate crypto in 2025. And as we are wrapping up the year, I think that's all we can really say is 2025 is going to be the year of AI agents in crypto. Bankless Nation, you guys know the deal. Crypto is risky.

• 1:13:32David Hoffman?

  Crypto plus AI agents, probably even riskier, you can lose what you put in. But we are headed west. This is the frontier. It's not for everyone, but we are glad you are with us on the bankless journey. Thanks a lot.