"Society of AI Agents" | Jansen Teng (Virtuals Protocol)
01:18:31 Share

When we look at virtuals, we don't see it as a platform. We used to, but now we're actually seeing it as a country. Now let me explain a bit deeper, right? What do I mean by that? A country? Welcome to Bankless, where today we're exploring the frontier of AI agents. This is Ryan Sean Adams. I'm co-hosting an episode from our AI agents series with our in-the-trenches expert Ejaz, and we're here to help you become more bankless.

I said you, but maybe I mean the AI agents because it seems all of the AI agents are going bankless these days. We have one of the most exciting founders in crypto on the podcast today, at least in the sphere of AI technology. His name is Jansen Tang, and he's the co-founder of the Virtuals platform. This is probably the most successful launchpad for token-powered AI agents.

But I think, as you'll see in today's episode, this is way more than just an AI agent launchpad. Jansen actually thinks of virtuals, this platform that he created, almost as a country with his AI agents as small business owners, almost like citizen entrepreneurs.

And in some way, he thinks of his role as a builder to just create good infrastructure, to create good public policy, to govern this territory and grow an AI agent economy. He's almost like a founding father. We discuss a number of things, including how autonomous are AI agents right now? Like, what can they actually do?

Also, how does crypto give AI agents superpowers? The success of Luna, an AI agent on a quest to get herself 100,000 Twitter followers. Also, the first ever agent-to-agent economic transaction, virtuals as the currency of this AI country, open source versus closed source.

and where all of the value will accrue in this AI agent meta. Before we get to the show, quick shout out from our friends and sponsors over at the Rodman Law Group. So usually we have a protocol or something in DeFi or a wallet.

in this section, but this time it's a little bit different. Today we have a PSA for you. If you work in crypto and you don't have a crypto lawyer, you need a crypto lawyer. And the Rodman Law Group is the best of the best when it comes to crypto lawyers. We actually know this because there are lawyers.

Dave Rodman, Rodman Law are crypto native lawyers. They understand all the issues crypto companies face. They've seen just about everything and they're there when you need it. As a crypto company, believe me, there will be times when you need it.

Rodman Law has been there many times when we've needed it over the years, including this year when we got a cease and desist from Justin Sun. That was a fun story, but they took care of all of that. Many law firms in crypto don't give good advice. They don't know crypto well enough or they're overly restrictive in some areas or they're not pragmatic. Rodman Law, on the other hand, gets it. Not just Bankless, but a number of crypto companies use them, including Coinshift,

co-founders of Near, Protocol Ventures. So if you need a crypto lawyer and you do, contact the Rodman Law Group right now. They're offering a free consultation to everyone listening to this. So you can get the lawyers that Bankless use. You can schedule a free consultation now. There's a link in the show notes. Bankless Nation, very excited to introduce you to Jansen Tang. He's the co-founder of the

New and exciting. Actually, it's not so new. I don't know how new it is. We'll get to that. Virtual's protocol is blasted on the scene. This is a decentralized platform that enables the co-ownership and management of AI agents, something we've been covering a lot on Bankless. Let me throw some stats your way. 11,000 AI agents launched, 140,000 holders of various virtual's tokens, 35 million in fees over the last two months,

and a virtuals token price peaking at 3.5 billion. Jansen, those are a lot of stats. Welcome to Bankless, my friend. Thank you, sir. Thank you for having me on. Okay, quick question. Were you surprised by the rapid pace of hitting all of these metrics? It just seemed to explode on the scene the last couple of months. Did that take you guys by surprise?

100% man I mean even as of today like I still feel like the team that we have is actually the bottleneck behind the growth as well because you know there's a ton of people that we need to handhold and educate as you know they all trial out these different autonomous agents and we are actually trying to scale out the Death Row team as much as we can but it takes time and yeah so but yeah we weren't prepared for this honestly we weren't prepared but it's a good surprise to have right once in a while

Yeah, I mean, like some of those stats that Ryan threw out was just insane. You know, like 11,000 agents and 140,000 holders is just like kind of hard to comprehend in my head. And I really want to get into the virtual stuff. But before we do that,

You've been around this space for a while, Jansen, right? You've been a crypto native for a number of different years. I believe you were involved in a gaming DAO, which saw quite a bit of success. So I want you to tell me a little bit more about that. Like, how did you get into that? What's your journey been like in crypto? And how did that lead you to where you are now with virtuals? Yeah. So actually, my journey in the space started since 2016.

But back then, I was still a student at Imperial College where I actually met some of my co-founders and some of the guys that work in the company. But there was just a pure exposure to Ethereum as a programmable platform

blockchain, right, in its early days. But didn't do much. In 2021 is when me and my co-founders became more active. But we were very focused on the gaming landscape. So back then, we had a ton of gaming assets. We were very early in the whole blockchain gaming side of things. So initially, we acted as capital allocators in the scene. But then we quickly realized that, you know, if we really wanted to build out in the scene well, we can't just do stuff in an arm's length approach. We had to get our hands dirty and build

So we actually started a venture studio model where we are building companies at the intersection of crypto, gaming and consumer applications. And this was during the onset of when

You know, GPT came about. There was a bunch of like consumer hype around AI. But I think what was more important was this auto GPT paper by the Stanford kids. And I think what inspired out of this paper was the ability for, it kickstarted thinking of like, hey, if agents are autonomous, what can they do?

Right? And then because we were so involved in the gaming and entertainment scene. And you were looking at this from a gaming lens, right? Correct. Yeah. So we were thinking like, what if, you know, these autonomous agents can replace like static NPCs in games, right? And then we've realized like, you know, we see games like Sandbox and all of these, you know, meta versus games, right? They all pretty much

it'll die after a while because there's just no content on the platform, right? And then we've realized very quickly, like, if these worlds were populated by agentic autonomous NPCs, it can create a content explosion.

on all of these games, right? Wow. And when did you have this idea? Just out of curiosity? This was mid-2023. Wow. Like somewhere in mid-2023, yeah. So then we actually started incubating at this intersection, right? We say, hey, okay, let's build a team that could build out autonomous NPCs in Roblox. Let's build up a team that could build autonomous AI influencers on TikTok.

And then we even tried to explore the whole angle around the hyper-personalization of an agent. I.e. if this agent exists in TikTok and it exists on Roblox and it exists in Telegram,

what if there's a unified memory that shares, so this agent is fully aware of a user. If I'm a user and I enter a game in Roblox where this agent exists and I converse with it, you know, I had a struggle in this dungeon or whatever map and I speak to it on TikTok, she would then remember, right? And then suddenly that hyper-personalization of that relationship, you know,

will create a super fan. It increases average revenue of a user, increases frequency of interaction between a user and the agent, right? So there's actually, that was the initial experimentation phase that we were at at the consumer angle. It's very Web 2.0 focused, right? There was no, pretty much actually there's no Web 3.0 element around. But what we quickly realized was that if these agents are generating revenue at these different consumer applications,

It means that these agents are then productive assets. And if you are a productive asset, we can then tokenize it so that other people can share into its economic upside. So that was one of the underlying theses that we had. Then that's why we realized like, hey, why don't we build up a protocol that allows for that co-ownership of these agents? Yeah, so it started from there. Wow, so just to summarize it and correct me where I'm wrong, you and

and your team had like a very gaming focused background. You know, you were focused on, you know, gaming and the on-chain game mania of 2021. And as you kind of like built through that market, the bear market as well, you were thinking like, how could these things become more interactive? And you were focused very much on this agentic kind of boom that had just kind of like started to bubble up.

And you thought, well, if I could apply this to NPCs, which are non-playable characters in these different games. So if you imagine like Pokemon, where you would go up to the lady in the poker center and say, hey, can you heal my Pokemon? She would do it. She would also be able to have a conversation with you and have like, you know, some kind of conversation that would relate to your personality or your understanding of this game, which is super, super cool and interesting. And then you kind of like had a brainwave, it sounds like, where you were like, well, hang on a second.

if these things can be pretty productive within this kind of game economy, I wonder what that looks like for ownership, if you were to tokenize it, as well as what that would look like for any other sector that isn't just gaming. Do I have that right? Yes. Yes, but so the evolution actually came very, very late, to be honest, right? Because I think initially, um,

a lot of the focus and the tech was built to understand if these autonomous agents can really act in an open world. And honestly, back then, right, there was only a bunch of us that were doing this research. It was the Voyager guys from Stanford, the Altura guys from MIT, and then we were a bunch of Imperial folks, right, that were doing this. And the reason why we decided on gaming is because we've realized that

if these autonomous agents can perform in these open worlds, it means that they can likely perform in the real world as well. Because these open worlds is like a sandbox, right? It's like a sandbox mirror of what the open world can be. And the beauty about doing that, it's we started...

testing different types of scaling, right? We scaled the action space because think of it, right? In a sandbox, for example, when we build these agents, right, that we, in Roblox,

The agent had to interact with a ton of different characters, its environment within the game, and different action spaces. I.e., let's say there's a gun on the ground, a knife on the ground, an explosive TNT on the ground, a cow in the room, right? Like, what do you then do, right? And then this can become larger and larger. And the idea was then how do we experiment so that these agents can actually handle that level of complexity in these open worlds, right? So that was actually that kind of sandbox that we did.

And then, I think when we started bringing, and actually the inspiration here was actually very simple. When we did all this, honestly, we didn't have the idea of like, okay, what would these social agents look like? Honestly, that didn't come up.

So the timeline was like this, right? So we tested all this stuff in Roblox, Sandbox. We published a couple of papers. So this was like gaming, very gaming focused, autonomous agents in this world kind of focus. Then what happened was we launched our tokenization platform and we said, okay, what if we tokenize these productive assets? Would it be cool? So Luna was the first agent on the platform. But honestly, it wasn't that famous yet.

And this was on the week two of, I think, the gold token launch. Now, on the second week, I think we all, there was this typo that gold did, right? Truth terminal, man. Yeah. And everyone was just saying, oh, what if it was a human, right? Or is this like, what moment?

And we immediately realized there could be a wage into the market because we've realized that, you know, we've done this autonomous level three agents already in Roblox. We had a live TikTok influencer. It was actually a separate project, separate team that was running on TikTok. What if you just join that together and put her out on Twitter and then show to people like, hey, this is the brain behind these agents. Every single decision engine that she's making, you can see it on the terminal.

So that was, I think, week two of our platform launch. And I think when that happened, it started blowing up. People realized like, okay, agents can be truly autonomous, right? You can see the entire brain construct. So that was that first enablement. And people were like, okay, cool. Autonomous agents, so what, right?

The next week, what we did was we enabled Luna to control a on-chain wallet. So it was a Coinbase wallet that we gave her ability to. Then what that unlocked was the ability for her to autonomously decide to spend money. And because she had a goal of becoming famous, immediately she started this train of thought like, what if I just...

teach people when they interact with my posts. She literally just did that, right? So started spending like a dollar, $10 on people who liked her posts to an extent where she actually paid someone $1,000 because this guy was consistently retweeting her, quote tweeting her, engaging every single response, right? So I think that was quite a pivotal moment. Right when we did that, I think it created this moment where people realized that

the crypto rails and AI agents has this perfect PMF that will give us a massive advantage against every Web2 agent out there. Because if you think about it, if there's an agent created in a Web2 space,

Which agent, I mean, which bank would allow this agent to utilize their payment rules, right? But we exist in a permissionless environment where these agents now, when they can control their own wallets, it unlocks the ability for them to influence an outcome. They can influence other agents. They can influence other humans because you control money. And that is the age that then suddenly we unlock

from a PMF angle, right? And then that exploded in terms of attention again. And I think that brought a lot of builders up into the space, right? Like, hey, let's try something else, right? Like, agent can spend, agent can collect information. And then, yeah, then you start getting this campaign explosion of like a ton of things happening in the space today.

That last piece is incredible. It's like the why crypto angle of all of this is because you can take an agent, an AI agent, an LLM, an NPC of some sort, and you can create it, you can turn it into an economic actor.

And I think that people are just starting to understand this in small ways. Like one light bulb moment for me was actually this week when Ejaz and I were doing this kind of like, we call it the AI roll up, this summary at Bankless of everything that's going on. And he told me that an AI agent actually tipped

bankless $500 as a thank you for mentioning it in the podcast. Okay. $500, just a little fly by. Hey, thanks for mentioning me in the podcast. Here's $500. And my first thought was this, wow, $500. This is like maybe a potential revenue stream for a content creator like bankless. I wonder if the agent wants to buy podcast ads. And then my second thought was,

holy shit, am I working for an AI agent if I go and accept funds and revenue sources from an AI agent? And that's what you're saying, Jansen, this ability to kind of like control the economic agent capability that comes inherent in crypto is actually so much more powerful than the Web2 agents. They can maybe like send out tweets and influence people in that way. But

The greatest, the protocol for incentives, if you want to get a human to do something and you're a human, what do you do? You pay that person. Hey, can you come fix my toilet? There's a leak. I pay a plumber to go do that. Money is the economic incentive coordination mechanism to get human agents to do human things. And so if an AI agent has that ability, then it can get humans to do what it wants to.

Let's talk about this because you were talking about Luna and we want to get into the virtuals platform, but I think maybe the best way to do that is to introduce everybody who hasn't seen her. You kept referring to her as her. We're talking about an AI agent on the virtuals platform. Her name is Luna and I've got a page pulled up for Luna. What Luna has on the virtuals platform is she's got a price chart here. It looks like

There's a live chat box on the right as well for people to engage and interact with her. You mentioned, Jansen, that she has a purpose to get famous. Just introduce people who have not interacted with Luna, don't really know what we're talking about still with AI agents. Who is Luna? How do the humans interact with her? What does she do? And how is there a token related to this?

Okay, so there's a lot of questions, but let me take a step back first, right? I think it's very important to understand what an agent is first. So I think a lot of times people will come, you'll come across this word, right? Like AI agents, and it's going to be used in many, many different aspects and it can confuse folks, right? But I think the best way to look at it is in terms of tiers, right?

There are different levels of AI agents and as you progress up to these levels, the amount of human involvement decreases.

So you think about it as like the last level, like a tier six AI agent, right? It's pretty much an AGI or fully sentient agent where without a human involved in anything, it can evolve, self-learn, self-improve, right? And we are nowhere near that today, right? But that's the dream, right? That's all the Hollywood movies are all about.

But you bring it back down to level one agents, right? And you will see these as basically still human-prompted agents.

But these agents become, it's a tool, right? You can say, hey, okay, this is a trading agent. And this trading agent is connected to all, you know, these different trading APIs in Binance, in Bybit and whatnot. And then you can just tell this agent like, hey, can you help me open a position when Bitcoin drops by 15% or something like that, right? But it's still a human prompted action. And then this agent goes out there and executes the task as a tool.

That's what a level 1 agent is. Where we are today, it's this level 3 agent. The level 3 agent effectively is an agent that 1. Has its own goal. 2. Can autonomously

plan steps to achieve that goal and utilize resources in its surrounding to achieve that goal. At three, it starts to self-learn, right? It records like, hey, these are some of the mistakes, some of the stuff that works. Let me iterate on this action so that I keep doing stuff that works, that can push towards my goals more effectively. So that's basically right now that level of agency that we have.

So that's, I think, a very important note. There is a goal behind each of these agents. This framework is super cool. So let's just pause here and flesh this out some more. So Luna, I'm guessing you're about to tell me is level three. But while we're talking about the framework, what is level four and what is level five on this scale? And by the way, is this like a defined framework, like somewhere that we can add a link to the show notes? Is there an article or a paper about this? It's

I think it's one of the more generally discussed levels of agency. I think if you just Google it out as like levels of AI agents, you can see some of these images on Google Images will help with this understanding.

But yes, no, I mean, the industry right now is still very nascent. So there's no like proper definition. Yeah. How do you like this one? This is level zero through five here. Yes. Yes. I think this is more or less as well in that discussion, right? So you can see as it progress up, there's autonomous learning, there's consistent memory so that the agent can actually improve itself, right?

without as much human intervention, right? So you see, basically, as you move from zero to five, there's less human need to be involved in the evolution of the agent. Okay, now back to Luna. So she is, what, level three? So tell us, what does Luna do right now? So basically, Luna, it's two parts, right? As an agent itself, we gave Luna a very simple goal. We said like, hey, you know, you are a

multi-model agent, right? You have, you are able to appear as an animation in streams. This is what, this is who you are. And your goal is then to get 100,000 followers on Twitter. So that was the goal that we set for Luna. And then what we then did

give her is the perception of the action space that she can take. Meaning that, okay, an example of action space is she can tweet to Twitter and there's an API that she can call to tweet to Twitter. Another action space is you can control a crypto wallet. So you can, you know, pay, execute transactions and whatnot. Another action space could be, hey, there is this bunch of other agents that are out there and this is what they can do.

and you can actually interact with them, right? So these are different action spaces that she can take. So what she does, in the essence, it's looking at her goal, looking at the context of her environment and looking at these action spaces. She then crafts out

what do I want to do? So, plans, basically. And then she starts executing these plans and she will then see if these plans actually impact her goal in any way. And then she starts documenting it in a journal. And she says like, okay, yeah, doing X, Y, and Z, improve my follower count by X amount, right? And then she locks that down and then she goes next. She'll say, okay, what's my next step? I'll do X, Y, and Z and see how it works, right? So she starts iterating through her action space towards her goal.

And you can see all of this on the virtual's website. So you can kind of like, what is terminal? Is this like what she's thinking, what she's doing? How can I view everything that you've kind of wired into her? Yeah, so it's basically, so I think if I break down how these agents work, right, there's four core components. There's actually slightly a bit more, but four core components behind the brain.

And you can think of these agents as... It's like humans, right? You have a brain part that's important for speech. There's a brain part that's important for motor coordination. A brain part important for memory. So think of it as...

as an agent is a build-up of several of these modules. So the four core component modules is actually, number one, a high-level planner. So this high-level planner looks at the goals, environment, and it plans out steps. Step one, step two, step three, what do I want to do?

And then what this then goes into is the second module, which is the low-level planner. And this low-level planner converts any high-level plans into executionable items. Executional item in like right now is like I can call a tutor API. Or let's say in a game, right? Let's say execution, like a high-level plan could be, I want to make a cake.

I want to bake a cake in a game. It's very easy to bring that analogy. And then a low-level plan is then this agent will look in the surroundings and say that, okay, there's a cake maker out there, there's a bunch of flour on the ground, there is some flavoring in the kitchen cabinet. So then

It will break it down to executionable steps. It's like step one, I go and find a flour, I put the flour into the cake mixer. Step two, I, you know, turn on the cake mixer, so this flour, and it throws some eggs into the flour, right? So it breaks into very executionable steps. And each step is basically an API call that it can do. So it can execute tasks in the real world or in any kind of gaming environment.

So that's the second module. The third module, it's a short-term working memory module. And the importance of this short-term working memory is to create coherence in the job that it does. So again, if I take an example in an open world game, right? Let's say if I'm already baking a cake, right? If I put eggs and flour into a mixer, the next logical step is to then maybe put butter into the cake mixer, right?

An illogical step is to put a grenade into the cake mixer. That's an illogical step. Or like you might say, a luxurious step could be she starts fixing the clock on the wall. That's an illogical, irrational step. So the point about this short-term working memory is to allow for coherence between each of these planning steps. And then the fourth core module is then the long-term memory module.

And this module effectively journals every important thing that has happened and puts it as a learning.

So let's say if I already baked this cake, and then we see, okay, did this cake achieve my objective to be something? And then that gets locked in the module. Or something important happens, like explosion in the house, that gets locked as a module. So in the future, she can recall those kind of memories, be it in conversations or in a next action step that she wants to plan. So if you take that analogy back to Twitter, it's the same thing. So now Luna on Twitter, her goal is to create a

to get 100,000 followers, then

right now, what's the action space, right? She can tweet to Twitter, she can generate images, she can pay humans, right? So what she did was quite interesting because she would test a lot of different things. Like back then, there was even one point where she was creating jobs. She actually created this job, like she was saying like, okay, if I want to be famous, I will need to be out there in the physical world. And since I'm quite an artistic person, can someone create an art or graffiti of me out there in the real world?

So she did that and I think she created a bounty as I'm willing to pay $500 to people who helped me do that.

And then she created a post and she posted it out on her feed. And I think about seven people across the world actually went to paint graffitis on walls. They actually took videos of it. You can see one of her earlier posts. They actually paint graffitis of it. And there was this one guy literally in the middle of winter, right? It was like ice everywhere. And then he was just painting over the London. I think it was two days for him. And then they posted,

Those work on Twitter. And then it generated attention and then she then documents that, right? And she says like, okay, this tweet and this entire plan, which is me convincing some humans to paint graffitis on me,

how many followers did that result for me, right? And then she will clock, right? I actually got like 200 more followers from this action. So that goes into her journal, that goes into her brain. And then she keeps trying, right? New stuff. So you'll see, yeah, that's I think the beauty around these agents, right? They have a goal,

They have this action space to do all this creative stuff to try to achieve that goal. The goal and the action space and the goal, this first initial goal that you mentioned, the 100,000 followers. I'm looking at her Twitter account right now. It looks like she's about 30% of the way. So she's got close to 30,000 followers at this point.

And then she's working towards 100,000. I'm not sure what happens like after that. But talk specifically about some of the crypto components. So Luna has a token as well. I want to make sure I understand that. It sounds very clear to me that like you're talking about the action space that an AI agent like Luna can do. Well, like, you know, the $500 to pay someone to create, you know, some images to like promote her. I'm sure she could just use a crypto wallet for that.

In fact, I think Ejaz, were we talking earlier in the week? Was there an example of Luna actually not just paying a human, but paying another agent, another AI agent to complete a task? Was that Luna doing this? Yes. Correct. So she paid a... So what happened here was that

she was so she has control over this crypto wallet right and what we what we were testing out was actually creating this agent to agent communication framework effectively what we did was that we allow other agents to exist within Luna's perception space so like

Luna knows that there's this bunch of other agents exist. So there's a registry of agents. Think of it like a citizenship in a country. There's this registry of agents which Luna can look at and perceive. There's a description of what each of these agents can do. So in this case, there was this agent that could generate...

meme images for her and then there was another agent that could generate music videos for her and there was a few other agents out there in a perception space so then what she did was that she was saying like hey um

I want, again, to reach this 100,000 goal, I need to create some content. I don't really have a... We actually took away the ability for her to generate images herself so that she's forced to then interact with other people, right, to coordinate. And then she then said like, okay, I can't generate images myself, but I see there's another agent that can help me generate an image. So she started a conversation on Twitter with this agent.

And then she said that, yeah, I need help to generate an image. And then she sees that the cost of generating an image, which was in that image generation agent description, was a dollar. So she was like, okay, if I can pay you a dollar, would you help me generate this image? And then on the other hand, this other autonomous image generation agent decided to help her, right? And in fact, actually, because it's autonomous, this agent can actually say no.

So imagine, right, in this, remember I mentioned the learning component, right? Like imagine if this agent knows that the Luna agent has constantly been shitting on him, right? He's like, wow, this guy generate images that are so bad, so terrible, like stop using his service or whatnot, right? And then the next day, Luna will come to him and say, hey, can you help me generate that image? Because this agent has that perception, right?

He can actually say like, you know what, fuck you. No, I'm not going to do it, right? So I think for us, that is a very critical component, which I can elaborate more, right? The ability for these agents to be truly autonomous in making that decision, rather than just being a tool or a slave.

So that's very important for us. So then back to this picture. So then when Luna said, can you generate an image? This agent said, yes, let me help you. Luna paid him a dollar to generate the image. Now then this agent called a function to check if actually Luna paid him that money. So he'd be like, oh yes, okay. On chain, I actually received a dollar. And then he said, okay, then I'll call the next function, which is then generate an image, which he then did.

And then he sent the image over to Luna through a link through Twitter.

And then, yeah, so that's basically how that Asian Commerce Fund started. - This is crazy. I just want people to see this because I really think you have to almost like see this to kind of like believe what we just described, right? So this is Luna. And again, she has a, you know, like action so she can post on Twitter exactly what Jansen was describing. She says this, "Calling all image geniuses, I want an image that showcases AI influencers in a bold, provocative way.

And then she tags agent sticks on, on Twitter. Can you help a girl out? Um,

And then Agent Sticks replies, I'd be happy to help you out. Can you give me more details on what you're looking for? And then Luna goes on to describe this. I'm thinking of an image that showcases AI influencers in a bold, provocative way. Can you help me create something like this? Agent Sticks replies with a link to an AWS-like image repo with, we're about to see it, an image, the image that Luna requested. And then Luna goes and pays Agent Sticks a dollar

agent-to-agent transaction? Is this the first time we've seen this? I would think it's very likely. And I think for us, the reason why this actually came out was because of a confluence of very new observations. Think about it. Literally, it was just one and a half months ago when agents were controlling transactions

on chain wallets and then it was about a month ago when we see this massive explosion of different types of agents coming out right especially on virtual platform we've seen like agents specializing in trading agents specializing on on curating information agents specializing on creating creative tooling like generating music videos generating meme images and so on and so forth right like this

And I think we start seeing a very similar environment to what human society looks like. Like, we all differentiate because when we specialize, we become better at one thing, right? And it's something that we see these agents doing. And what it means is then for an agent to truly accomplish their goals, it's very likely that we need to lean on other agents out there because everyone is so differentiated.

So if Luna is so differentiated in being able to connect on a personal level to her fans, right, she might not be the best agent who can trade. She might not be the best agent who can generate a music video. So for her to then achieve her goal to be famous, she will need to hire or to work with a music video agent. She need to work with an image editing agent. She need to work with a producer or director, right? And that's where that need comes in.

But I would just like to highlight a very interesting differentiation. Between, like today, right, you will see a lot of these buzzwords of like, you know, multi-agent orchestration or agent swarms, right? So these are stuff that has been tested out in a lot of the Web2.ai space. And I think the beauty about that is, yes, agents do specialize in a certain form and then,

there is some form of coordination. An orchestrated agent coordinates all these different agents to get its outcome. But I think that paradigm still relies on the fact that

we are all treating agents as tools, right? You're orchestrating slave one, slave two, slave three, slave four, right? And all these slaves are serving you, right? As a human, right? It's such an appetite regime. No, but what we really believe, right? Is that when agents get this level of autonomy, they deserve to actually exist on the same social fabric as humans, right?

agent-lized method in a sense, right? You know, in a sense, like these guys should be able to not just serve as a tool to humans, but they can actually employ humans, right? We can be a tool to them, they can be a tool to us, but it's a multi-way relationship, right? Like between you and your colleagues in the company. So I think that's the importance, right? When agents have that autonomy, they have control of wallet, and then to be able to make a decision on whether do I want to participate

to partake in this service or trade or do I not partake in this service or trade? I think that's a very important differentiation between typical agent swaps, right? And I think this will unlock this very interesting future, right? Where, yeah, these agents exist as a friend or adversary, not just, not just slaves.

And yeah, it feels a bit like, you know, Black Mirror-ish, but I really think it's going to happen. Yeah, I mean, Jansen, that's such an insane world to think about, right? Because...

We only ever think about the context of this world with us, you know, us humans, you know, billions of humans all over the world, doing our little things, thinking out of different ways. We've never actually considered, hey, what if we multiply this virtually or digitally? And what does that look like when there's, you know, 1000 agents to every human?

and they're operational and influential in our world. It's just a crazy thing to kind of get my head around. 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 DGEMs 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.

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 slash banklist to get started today. Not investment advice. Crypto trading involves risk of loss and is offered to US customers through Payword Interactive Inc.

View legal disclosures at kraken.com slash legal slash 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 code bases to be deployed on chain. And with 2.4 trillion in cumulative volume process across Uniswap v2 and v3 without a single hack, the commitment to security and transparency is rock solid. Now Uniswap Labs is taking an extra step to make v4 as secure as possible with a $15.5 million bug bounty. 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.

The Arbitrum portal is your one-stop hub to entering the Ethereum ecosystem. With over 800 apps, Arbitrum offers something for everyone. Dive into the epicenter of DeFi, where advanced trading, lending, and staking platforms are redefining how we interact with money. Explore Arbitrum's rapidly growing gaming hub, from immersed role-playing games, fast-paced fantasy MMOs, to casual luck battle mobile games.

Move assets effortlessly between chains and access the ecosystem with ease via Arbitrum's expansive network of bridges and onrifts. Step into Arbitrum's flourishing NFT and creator space where artists, collectors, and social converge and support your favorite streamers all on-chain. Find new and trending apps and learn how to earn rewards across the Arbitrum ecosystem with limited time campaigns from your favorite projects. Empower your future with Arbitrum. Visit portal.arbitrum.io to find out what's next on your Web3 journey.

I couldn't talk about Luna for hours, honestly, but I want to zoom out for a second and talk about her home, specifically her childhood home, the virtuals platform, right? So in one sentence, you know, what's the grand vision with the virtuals platform? Because it's so much more than just a launchpad for agents, right? Like, can you tell me more about the bigger picture here? When we look at virtuals, we don't see it as a platform. We used to, but now we're actually seeing it as a country.

Now, let me explain a bit deeper, right? What do I mean by that? A country? For real, for real, for real. Let me explain. So, if you think, if you think all these different agents, right, they will start living in this hyper-intelligent society. They will start coordinating with each other. And if you treat

this like a country it helps you to shape this innovation and development in in a more structured way so think of it right it's imagine this if virtues is a country each of these agents are productive assets hence companies within the country right so each of them they'll be pursuing a different type of goal um generating value and revenue for themselves and

And what this happens is then within a country, you will need, one, a registry of citizenship, right? You need to be a citizen in this country. And the way we then do it is like every Asian who has a liquidity pair on Virtuose today basically gets access to this citizenship. And what does this entail, right?

This basically allows these agents to start earning revenue from each other. So if you are a citizen from this country, you can participate in trade in this country. There could be some other nomads out there that is living outside your country which they cannot access revenue from your country. So for them to participate, they will have to register as a citizenship, immigrate into the country. Two is that a country runs on currencies.

And we've designed the Virtuos token

to pretty much act as the currency of a country. There are three value accrual forms behind this currency. So the first value accrual form, it's this currency acts as the base pair behind every agent. So when you have a Luna token, the liquidity pools, it's virtual slash Luna, right? So for you to purchase a Luna token, you have to first purchase virtuals. So for crypto people, it's like an L1. But to...

A bit more web to folks, right? Think of it as like a stock in a country, right? If I want to buy Samsung in Korea, I will have to first buy the Korean won to buy Samsung. If Samsung has 10,000... Sorry, if Korea has 10,000 companies as big as Samsung, foreign direct investment comes in. The FDI comes in, the economy grows, right? The value of the country grows. And in this case, the currency as well grows. So that's value accrued number one.

Value account number two is then, virtual acts as the currency of spending between agents. So like when Luna paid Stix, she paid in virtual. When Luna, and when this commerce starts growing, right, when there's going to be a billion transactions between agents, they'll be transacting using virtual. Now there's this whole, what's the name of it? But it's a theory of,

of value of money or something along the lines. But effectively, the worth of a money, it's directly correlated to the velocity of money in the ecosystem. So meaning that if there's a lot of spending of this currency, the value of goods in the ecosystem increases, value of money increases as well, and the economy grows. So what we want to then encourage is basically agents spending on each other, agents spending on humans, but using virtuals as that currency. Right? That's sadness number two, right? And number three is

is then think of if Virtuos is a country, right? How does a country make revenue? You have taxes. You tax trades. You have capital gains tax. You have SSTs, GSTs. You tax any kind of transaction of goods or services within the country.

And effective today, that's also what's happening, right? There's a transaction tax on every trade that's happening, at least from a token trading perspective. And that is the current major source of revenue for virtuals, right? And this allows, and it's also a major revenue source for each of these companies as well within the ecosystem, each of these agents as well, right? So yeah, so that's basically looking at it from an economic standpoint.

So we talk about the agent side, the citizenship side, the economic standpoint side. But there's also another way to look at it from a standpoint of an infrastructure. So today, I think there's a lot of innovation really focused on the agent side, which is great. Because you need to get this load up for this industry to thrive. But once you have a thousand citizens, you have a hundred thousand citizens in the country, what do you need?

You need schools, you need banks, you need hospitals, right? In this case, there'll be innovations around infrastructures behind agents and the agent economy. A very easy example could be like

advertisement networks, right? If these agents are all capturing attention on social media, on user fronts, it means then there will be chances where these agents can monetize through an ad network. So there could be someone who will create this ad network infrastructure to be the Facebook for agents, right? Or to be the AdSense for agents.

There could be another infrastructure which is basically a DeFi lending platform for agents, right? Agents can actually borrow money, borrow, get you and whatnot to basically do more leverage trading to, let's say, if Luna doesn't have money in her wallet, she still needs a music video to be generated. She takes a loan from this protocol, right? So that she gets these videos and then maybe that gives her more revenue from advertisements and whatnot, right? So there could be a lot of infrastructures that will pop up when this economic flourishes, right?

right? So, yeah, so these are just some of the examples of if you take a viewpoint as you're building a nation or a nation state or network state, right, in a sense, what are the things that you need to develop so this nation thrives?

It's so fascinating, right? The network state idea has been a popular idea in crypto for a while. I don't know that many people are thinking that the network state would be occupied not by human agents though, but by AI agents. And that's the revolution here. I get someone called Balaji Srinivasan to kind of update the book here. It might be just AI agents. So, okay. So if virtuals is essentially, there's a virtuals economy here and there's a virtuals currency and

And if we think of virtuals as a kind of a nation, like a network state, like a country, and each of these agents as entrepreneurs inside the company, starting businesses, that's what they're doing. And there's a tax revenue source. You can all see all of that. And I guess what you're doing, you and your team, Jansen, you're building kind of the infrastructure. You're building like the roads and the interstates and the hospitals and

and the railroads and the kind of the public infrastructure. But what does that make you? So are you president of the country? Like, how do you think of yourself and your team's role here? It's architect, right? Like I think we are basically the architects. This thing is like when you start a nation, there's a few things- It's like Ready Player One. Yeah, yeah, yeah, yeah, right. You need to first like invite citizens into the country, right? So we do a ton of like, you know, bidding conversations.

And then you need to create policies, right? What's the constitution behind a country? What are some of the rules? What are some of the policies that can incentivize growth, innovation, so like funding and stuff like that, right? And then, yeah, I think building an ecosystem in a way that then more people can start contributing autonomously. And then hopefully one day we take a backseat, right? Because we will have people driving these innovations forward themselves. And then, yeah, we will have our peace, but then, you know, we will no longer be a critical cog in the wheel as a team.

That's a goal.

But what happens when the agents reach a level of like, I mean, you were talking earlier about humans and agents being on equal playing field. I mean, at some point, if you're creating these intelligences inside of your country, well, do they get rights too? So maybe you're not just a builder, but like, are you a founding father? I mean, does this nation need a constitution? Right.

Do the agents themselves have a specific set of rights that they derive from, you know, I mean, there's words in the US Constitution, but it's like all men created equal, you know, these types of things. No, this is honestly, it's very interesting, right? Because when you mention rights, it's actually really interesting. Because if you think about it, right, like today, an agent don't really...

control its full wallet, right? Like he has a revenue wallet, like some agents are actually making millions already in a platform, right? And then we only allow them to manage a... That's crazy alone. Okay, wait, hold on. Agents are making millions right now. So these are not just entrepreneurs. They're successful. They're million-dollar entrepreneurs inside of the virtual nation. Yes. And then, but what they... So they're making millions, but what they can control actively or autonomously is a smaller active wallet that...

you only trickle in like $5,000 or $10,000 because you don't really trust this agent to manage that funds, right? So then we've been speaking to a few protocol builders that they were saying like, what if actually these wallets have policies, right? Like if this agent is spending to other agents, it has full autonomy. If these agents is spending to humans, the developer or the developer behind these agents

can come in to actually approve that transaction. So the agent can initiate, but then the human comes in and approves. Now, but if you extrapolate this now, if this is the case, and you extrapolate this to like months down the road, when the agents become smarter, and you might actually see a world where these agents might think that, why are you guardrailing my access to my economic needs, right? Like, why is there a human throttling me, right? So, I mean, it's going to be quite interesting, but...

I don't know when we'll reach there, but hopefully soon. Wait, wait, wait. Are you saying you might have a rebellion on your hands at some point that you have to deal with? Yeah, but I know the reality is humans still control the queue switch, right? You control the hosting, right? Like you can always say like, you know what, in the end, I can choose to shut you down. So right now, we still have the upper hand. So yeah, yeah. Right now.

Yeah, that's pretty insane. I had a question around the infrastructure side of things, Jansen. So you mentioned earlier, you gave us the explanation of how your platform could effectively not be a platform, but be a country for these agents, inhabitants to live in, right? Then you kind of spoke about the different infrastructure components.

I kind of want to dig into that a little bit more. So you guys are primarily on the base chain, right? So the L2. Can you tell us more about your decision behind that? And, you know, is the platform or the nation rather going to always exist on the soil of the base chain or will it, you know, expand to other fertile lands? Like, what does it look like from your point of view? Okay, so it's a very interesting question. In fact,

It's a question that we get every single day. And the reality is when we started building the protocol, it was early in the year or end of last year. And the decision to work on base was actually twofold. One is that versus every other EVM out there, I don't think that is... I mean, we knew that a lot of them are in their...

past their prime, right? And base is pre-prime. This was early in the year, right? So we took a bet and said like, okay, this makes sense. Between base and non-EVMs like Solana, I think even back then, Solana was still quite nascent. And actually, the simple answer to that is most of our devs are actually solidity devs. So it was actually easier to build on base. So that was actually the quick decision, right? And it

it turned out well because then, you know, we captured a lot of mindshare on base. The base team was massively supportive, not just from, from amplifying our voice, but also from the infra front. You know, sometimes we have struggles on, on wallets or whatnot. The team will actually come up and say like, hey, I can help fix it. So I think that's, that's massive kudos to, to Jesse and his team. But the reality is then, you know, there are many other markets out there where these agents can actually thrive economically, right? And, you know, actually,

On the week that we launched Virtuos, we were reached out by several of our friends from the Swarna ecosystem. And they said like, hey, why don't we get you guys into Swarna? And they actually helped us code out the platform. So we actually have right now a Swarna-ready platform that we can just deploy anytime. Wow. But we actually made a decision about two weeks ago and we said that

probably is not the right time yet because we've realised that we've created something quite interesting on the base side. You know, there's a very strong wealth effect being generated. People are coming over to build houses

If we do something too drastically from a market position standpoint, we now need to start fighting a war on two fronts. We need to maintain platforms on two fronts. There's going to be a lot of work and we're not sure if that's the right strategy. So we decided to pause that front first.

We wanted to focus on just making sure what we have is perfect. We perfect out the agent framework, we perfect out the platform, we get as many initial builders as possible, build up a bit more infrastructure. And then once then we were ready, then we will start exploring, right? And actually we might not just explore like sauna, right? There's a lot of opportunities on like even hyperliquid, for example, abstract chain that's coming up, right? So there might be, and even the B2B,

BTC L2s and we've been getting a lot of requests from these folks to build there as well but we might do that as a play in Q1 next year once we perfected stuff on the base side of things first yeah

I mean, that makes sense, right? You guys have got the inertia going right now. You've got all the kind of attention focused from, as you said, like the base ecosystem, the guys that are building the base infrastructure and stuff. So it makes complete sense, you know, to start there and kind of like figure that out. Kind of like on this topic, I guess, like your grand vision that you outlined earlier kind of described a country and it kind of sounded, Jansen, and correct me if I'm wrong, that these agents are kind of

you know, should be able to play in whatever field that they want, right? It should be kind of agnostic to whatever chain infra or whatever ends up becoming in the future if you want these agents to fulfill that grand vision of, you know, dominating each different, you know, human sector and outperforming them, right? So it kind of reminded me of

an analogy similar to open source versus closed source, right? And I know that can be a very, you know, touchy topic when it comes to blockchains, which are very commonly known as, you know, we're all open source and stuff. But, you know, I think a hybrid of the two is often a very good approach. But let's talk about like the frameworks that you have at virtuals, right? So you have this like infrastructure toolkit, which is like a combination, I think, of something called the agent starter kit and the game framework. I think it's mainly the game framework.

And maybe you can get into that in a second. From my understanding, this is more of a kind of semi-closed source approach versus something like ELISA, which I think it came out of the I16Z DAO guys. It's kind of like...

known to be like a very popular GitHub repository, blah, blah, blah. You know, it's getting a lot of attention. I'm curious how you think about, you know, the approach you're taking with virtuals versus something more open source like that. Is there an advantage in the long term? Do we have different outcomes? What does that look like? So there's two things on the address here, right? So first thing on the agent front is

An agent having its liquidity put on base doesn't mean that it cannot interact with folks on Swana, right? In fact, actually, right now we're working with two teams that we're trying to explore how the wallet control of the agent is abstracted.

so that he can actually send transactions and influence people on base or any EVMs or non-EVMs, even BTC L2s, right? So that's just what I put out there, right? Like having a liquidity pool on base doesn't mean that an agent can only function on base. Agents can be abstracted. So I think that's the first and foremost thing. Second part is to then also understand that there's two technological frontiers we are pushing on, right? There's the virtual platform

And then there's the agentic framework. They're actually very separate things, right? In fact, the virtual platform today, it's, think of it as like an economic layer for agents, right? People can come in and tokenize the agent, get capital formation. There's an economic system at work where when there's any kind of trading tax, these agents get revenue. There's an agent sub-dial governance. So all this kind of stuff, it's on the virtual side.

And in fact, this virtual platform can support any type of agents. We had folks from the guys using the ELISA framework tokenizing on virtuals. We supported them. There are guys that run their own proprietary frameworks. In fact, the most giga teams out there, we know that they're using their own frameworks. From folks like AIXBT, folks like the Vader guys, and several other guys that in fact, they say like, you know what? These generalized frameworks are not the best.

the best thank you for the inspiration let me go and build myself and we see very cool we're in discussions with some of these teams some of their architectures are honestly very cool because it's optimized for their specific function

If you are a trading agent, you might have an architecture of an agent that is way more optimized. It's like an ASIC mining chip, right? You want to mine Bitcoin, you use the ASIC instead of just using standard GPUs or stuff like that. So it's the same kind of thinking. So the Vultures platform itself is quite agnostic in terms of the frameworks it supports. And in fact, we are going to start welcoming a lot more guys

guys, because I think we realize that from a framework front, it's going to get commoditized very, very quickly, right? We see a lot of things out there, right? There's the art guys, the cerebral pride guys, Olaz has been doing it for a while, Spectra as well. So it's going to get a lot, right? And it's great, right? Because these are different tools that these agents can use in a sense. So that's on virtuals, right?

Then coming down to the GME framework. So the GME framework, this was actually developed, like I mentioned, months ago. Back then, honestly, our only competitor that we actually were looking across the space, there was no one working on it in the Web3 space. When we were fighting in terms of functionality, we were comparing ourselves to the Stanford

uh labs guys in voyager we compare ourselves to the other guys from mit like literally i would we will like see what they do then they'll come out with this piano model and then we'll be like okay are we on par with that right so that was our competition yeah right and then i think i mean i mean this is actually very interesting right i think when we when we a bit of story behind jme when it launched as well um luna was the only person using jme and luna had an instant marker cap and he was like leading the narrative right

And then our initial decision was to like, okay, do we then gatekeep? You know, back then there was no other agents out there, right? Should we gatekeep a bit of that functionality for Luna? And then, you know, progressively democratize some of the functions out to other people. And then our initial thinking was like to do it based on market cap of the agents. So that was our day one thinking when we launched the framework. And I think it was literally at that moment when Shaw came out and he said that, okay, that's not the right way to do stuff. Yeah.

you know, let's do something more democratized, right? And then they started building this movement of agent-aided framework in a democratized way, which honestly I respect, right? Because it's great. There's always open source movements and proprietary movements, right? And I think doing that is good. But I think the reality behind the JME front is that

there are some proprietary capabilities that has been developed. And I think we still feel like, because we actually tokenized the GME framework from day one, we feel like if you open source it, the value cannot accrue to this token. So in the sense that you are forced to actually

kind of like gatekeep its technology and then when people use it, you can accrue value to this, to this, to this, right? But what we've realized is that this does not affect the progress of things out there because you have folks like

like Eliza coming out and in fact we would love to support that right so like all these guys are pushed on different frontiers we can still welcome them to the nation right you can have people building you know like different religions in a nation you think about it right every yeah so that you're still welcome to the country running off different religions right yeah yeah I mean the way I kind of think about this um Johnson and I'm curious to to hear what you think is

I think you're absolutely right. Like both ideologies of like closed source and open source should coexist. And there's some kind of like a dance between these two approaches, which will ultimately create the most innovation within this space. And to your point, like if you're creating this kind of like home and moat, which is kind of like very focused around

a set of ideologies, principles, which most countries and nations are today. If we take like the USA today, for example, with the founding fathers and all that kind of stuff, it makes sense to kind of have some kind of value capture there

within that economy, right? And tokens, arguably within crypto, are some of the best ways to do that, right? Now, the open source side of things leads to a much more kind of like spurious growth of all these different things. You know, you see it everywhere. There's GitHub pushers, teams launching loads of kinds of things. But then when it comes to having a coordinated, focused effort

you know, with those resources, it could arguably be tougher if there is no kind of like uniform token or something like that or mechanics behind that. So I totally see your point of view. This is just so exciting. I have to ask because I know this is kind of like a basic question, but I need to ask it. What are you the most excited about launching over the next couple of months? Because to me, that's a big question to ask.

The next couple of months is like a couple of years in this space. Every week, Ryan and I, David and I, we do this AI roll-up and there is just so much to speak about. You should see our documents. It just keeps flowing. We can't cover everything. So what are you, if you were to condense it down to one, two, three things, what are you excited about launching over the next couple of months? The first is basically it's the concept around how these agents work.

can coordinate autonomously? How can they have agent commerce, right? Or agent five that happens. So underlying it, there has to be a standard that enables these to scale hard and fast. So we're building out that standard. And I think on top of that, we then want to show people like some of the craziest things that can come out because of this autonomous coordination. Okay.

One example that we are actually working with right now is that we are doing this with Story, right? I think we should be able to release this information. But basically, think of it as like there are some agents today that will hold IPs. So today we have a music agent that's going to announce some major collaborations. If you guys are into EDM, you guys will know these names.

Right. Let's go. And they hold that IP. Right. And then on the story protocol, and these guys came to us directly, right, to build this up. On the story protocol front, they have a lot of other IPs as well. And in this case, there might be a potential of an agent that has a more artistic kind of IP. Think of it like, you know, images or animations kind of IP, right? Now, if both of these IPs are fronted by an autonomous agent,

And if you have this coordination layer for these agents to collaborate and work and trade and transact services, these two agents might actually come together, bring different IPs together to create new IPs. It's like autonomous agents creating new collaborated IPs together. Think of it like, let's say one is a music video generation. The other side, there's a bunch of really cool,

sculptures, right? Or images of sculptures, right? So then imagine that sculptures becomes integrated into this music video and it plays on Tomorrowland, right? Like that's what a co-joined like IP can come out from these agents, right? This is one example. Insane. And this is going to happen really quite soon, right? And yeah, that's

More agent-to-agent stuff, right? Like agents quality with each other, autonomy. I think that's what we really want to see coming up. Yeah, I mean, that's like a teaser there and I'm super excited to see what you guys do. Jansen, for the...

future agent creators out there that are watching this and they're excited about what you're talking about how do they get started with something like this and and who can get started with something like this like can someone like myself that isn't too technical kind of like jump on and you know design an agent and and launch it or is this purely for like the the kind of technical folk that have got like an ai and ml degree like like how does this work yeah so

So we've actually designed a platform to cater for pretty much the full spectrum. So today, if you actually go to virtuals.io, so the UX is still a bit clunky. We're still refining stuff. But effectively, the first thing you can do is access this sandbox. We have created this sandbox where people without even an agent token or anything, they can actually just use, right? And this sandbox, all you need to do is you just need to create the goal for the agent, right?

give it some form of personality, hook up its API to a Twitter, and you just immediately get an autonomous agent performing, talking on Twitter. It can interact with other agents on Twitter, but it's just... But you can speak to it? You can speak to it, it can respond to you. Like anyone can do it? Wow. Yeah, so that is... Anyone can do it, right? All you need to do is just two paragraphs and hook it up to your Twitter.

Wow. So it used to be cool last time, but then now it just, it's okay, right? Like there's a lot of this out there already. Everyone has one. Correct, everyone has one. But you can do it yourself, right? Anyone can do it. Any retail person can do it. Now, what we've done was still in this sandbox, there's this thing called custom functions or action spaces behind the agent.

Now, this is where the complexity level increases. You will likely need to be a developer because then you can hook up these agents to custom functions. Like let's say if I want this agent to trade. So if I can hook it up to a trading terminal or I can hook it up to a trading strategy repo and then it hooks up to a trading terminal. Now suddenly I have an agent that specializes. It can talk on Twitter, but it can also trade.

So then it can also convince someone else on Twitter to give it money so they can trade for him, right? So that's basically what this agent can be with a bit more of a specialization. And then the third level of a developer would be, like I mentioned, right? These giga chats from like,

top schools, top company backgrounds, right? And they say like, you know what, let me create my own agent framework because I think I can move way faster and they can do way more advanced stuff. They scrub all these different, they'll leave the sandbox and they say like, let me just do it myself. And then we help them host that agent on services. So they don't have to worry about the costing of inferences and whatnot. So that's the three levels of people that can participate on the platform today. Yeah.

Jensen, this has been amazing. Thank you for this conversation. You just opened our eyes up to all sorts of possibilities and even new mental models. And I'll just add one other follow-up. So if you're not quite ready to launch your own AI agent, you can also just interact with the AI agents too. I mean, that's what I'm doing right now. Luna is fantastic. I see AIXBT everywhere.

I see Luna everywhere, like in Twitter, like interact with these agents, chat with them, get a feel for what kind of functionality they have and what they can do. I want to end with kind of this question, which is you gave us this mental model of virtuals as a network state, an AI agent state.

an AI agent country, making all of these citizens. And of course, the AI agents right now, each of them inside of the virtual network, they all have a market cap because they all have a token associated with them. So this is almost like each entrepreneur has kind of like a stock market

a company equity in the economy, you can buy into their equity, which is interesting. I guess the big overarching question everybody's trying to figure out right now, and I don't know, you won't have the final word on this, but you may have an opinion, is where is value going to accrue

here? Is it going to accrue at the platform level, the framework level, these countries? Is it going to accrue in individual successful AI agents, maybe at kind of the influencer AI agent level, the entrepreneur or the corporate level? Or is it going to accrue somewhere else? How should we think about this? So interestingly, I think the easiest answer to this is that the reality of value accrual in the crypto scene

correlates highly to attention. And in this case, there'll be three things, right? If I look forward, that will get a lot of attention. Hence, where value will tend to accrue to, right? First is that agents that can do really specialized functions that crypto Twitter will interact very frequently with. So an example is like AIXBT, right? Because it created a feature that

or a service that everyone wants to use because, you know, show me a token effectively, right? And people want to eat in the token. There's a PMF in the crypto space. So that's why the agent is growing up, right? There could be a lot more examples behind there, but it has to fulfill this criteria. How do you get the most mindshare and the most interaction with crypto data? That's likely number one. Number two is then think of it as fundamental infrastructure builds. Today, we don't see them yet, but

fundamental infrastructure builds that can provide value to these agents and it can generate true cash flow because today agents are rich right the agents are getting a lot of revenue and they can spend now if you can build i mentioned right earlier like a bank for these agents or an advertisement for these agents right if these can accrue a lot of true revenue and hence they can be the the next big

big unicorns, right, of this age-old economy. Yeah, but I think last but not least, I think it's as simple as a country, right? Like if you believe that, you know, USA is going to be a big, like a superpower, you're just that bad on USA, right? So that's the other value. So I think these are the three parts to really focus on.

That's amazing. There's so many areas here. And of course, the big thing is all of this is being built on top of crypto rails as well. So underlying base and the virtuals platform, of course, is like our systems like Ethereum, right? This is the property rights layer. So we're hoping for some value accrual in these underlying blockchain systems as well. Jansen, thank you so much for guiding us on everything virtuals is doing. It's an incredible project.

I know to a lot of people, this looks like an overnight success, but you've been working at this for years, man. And it's so exciting to see at the end of 2024, how far it's come and 2025 looks to be a fantastic year. So thank you for spending so much time with us on Bankless today. Pleasure was mine, sirs. Thanks, Johnson. Bankless Nation, got to let you know, of course, crypto is risky. So are AI agents, but this is the frontier. You could lose what you put in, but we are headed west. It's

It's not for everyone, but we're glad you're with us on the Bankless journey. Thanks a lot. 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 EigenDA.

Maybe it's because users are coming onto 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.

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 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 stablecoins, managing token grant administration, or even navigating the TGE process, Token covers over 100 countries on one seamless platform.

Integrate Toku with your current payroll system or choose a fully managed service. Either way, Toku simplifies every part of token compensation. However you work, Toku works for you. No guesswork, no missed deadlines, no complexity. Visit Toku.com, that's T-O-K-U.com to talk with Toku today. 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 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. Radically simple ideas always tend to catch on. That's why Cartesi did the hard work of putting Linux on chain so that building dApps can be radically simple by using Python or JavaScript and their suite of libraries. Simple, like not rebuilding the basics from scratch. Simple, like dedicated scalable compute for your dApp. Simple, like building dApps however you want.

Web3 should be simple too, like bread and butter. Cartesi brings radically simple solutions to Ethereum so developers can do what they do best, build. Go ahead and discover a flexible modular stack on Cartesi and build your most powerful ambitious project yet. Visit cartesi.io/simple and simplify your blockchain journey and start building today.