AI Meets Web3 by KuCoin Web3 Wallet

Hi everyone, welcome in, let's start. Right, so today we'll be joining, we'll be joined by Joe Skill and Botchain. So to give you some context, if you do any serious market research in crypto, then you probably already know the grind. You've probably got half a dozen Dex, greener tabs open, you have TG channels picking you left in right, and by the time the two run your manual calculations, the trade is probably gone. Today we're talking about fixing that workflow from two sides, on the analytical front with Jill Skill, who built an AI driven local first quant terminal and on the execution front with Botchain, who are building the modular layer one rails. So AI agents can actually transact on chain. Let's bring both teams up and get straight into it. Let's start with Botchain. Botchain can you tell us what is Botchain building and why does an AI application need blockchain infrastructure, which use cases are you most focused on today? AI agents, decentralized compute payments or something else. Botchain? Okay, I think they're still setting up. Let's pivot back to Jill Skill. Instead, Jill Skill. Yeah, I think for everyone here, who is monitoring markets, we usually end up looking at dozens of dashboards and feeds just to evaluate the setup. Can you tell us what is Jill Skill and what is, what problems are you trying to solve for people who follow financial markets? Like, if someone is using it for the first time, what can they do? That would otherwise take several tools or a lot of manual research. Yeah, we are, as Jill Skill, we are currently building an AI and later financial, like quantitative terminal for these, up the previous like institutional traders or these professional contractors. And we are trying to bring all of them. Jill Skill, I think you got cut off, but let me know if that was the intro. We can continue with bot chain if that's the case. Yeah, yeah, yeah, I can hear it now. I was having a little bit of tech issue. Okay, got it. Got to have you with us bot chain. Okay, Jill Skill did an intro, but yeah, let's continue with bot chain. Can you tell us what is bot chain building and why does an AI application need blockchain infrastructure? Like, which use cases are you focusing on today? Arm, is it AI agents, decentralized compute, deep in our payment or something else? That's an ape. Hello, they hear me. Yes, we can hear you. Oh, yeah, it's a thing I'm having usually again, but if you guys can hear me. So yeah, so bot chain is a AI native EVM complex level one. So build as the settlement and trust layer for agent economy and real assets. Why does a like a application need a chain? Because today AI can talk, it can, it can all host assets, set up payments or for what you did. At the moment, AI started moving like real value and needs infrastructure for identity payments and accountabilities. And that's what blockchain provides. On the user case, to be precise, we're not doing a decentralized computing, heavy inference, you know, belong off chain. We're a computer is efficient. And what belongs on the chain is settlement, authorizations and proofs. So we folks in two things. So first AI agents and the second is will be our RWA. Thank you. That makes perfect sense, especially when you need service like this when you're not using real value. Yeah. Yeah, let's continue with the old chill skill. I was circling back to chill skill. On the daily analysis side, I think the biggest risk with general AI models is outdated data, especially when crypto order books ship in virtually seconds. How does chill skill bring live market data into an AI conversation? Could you walk us through the practical ploy, the example? And like from starting from looking up an asset into asking a useful follow up question, go ahead, chill skill. Thank you. Chill skill, we're not able to hear you. I think the space today is a little bit laggy. I see. That's unfortunate. This is a very exciting topic. I'm sure our audience would agree. Let me message the team. See what we can do. OK, let's continue. So you're going through a couple of questions here. OK, blockchain. If an AI agent can interact with smart contracts or make payments, how can users verify what it has done? Like what role should on-chain records and clear permissions play? OK, so sorry, that was a good question. So verification started with the fact that every action on blockchain is a public audit but receives on the blockchain explorer. So no black boxes, whether the actor is a human or machine, anyone can check what actually happens. But the record is a long aren't enough. So the deeper issue is the authorization. So how do you let the agent spend without handing? It's a key. So our answer is programmable and bond with permissions. So with agent pay and smart contracts, AI agents operate with in-limits the owner's defines. What it can pay how much to own and every action within those limits and lives an on-chain tail. So the model is user keeps control and agents gets capability and as a chain keeps record. So on-chain records makes the past audit both. And clear permission makes the future bond. So you will need both. Sounds good. Thank you. Yeah, so having that record or clear visual is something that would really help make people trust bots with capital, making sure that the signal is clean before bridging over to the on-chain action, which brings me to this question. Where do you see the strongest connection between AI market tools and on-chain infrastructure is what would need to happen for an inside generated by an AI to lead to a useful trustworthy on-chain action? Yeah, my question, right? Yeah. That's a question. What about the trust was the on-chain action, right? Yes, that's it. Yeah, so I think the strongest connection is exactly the last mile. So from inside to execution, a tool like Joe still on a turns market data into a decision. But blockchain is where the decision makes a settled from probable actions. So for the work trust for the AI, so forcing needs to exist on the infrastructure side. So first, I think that would be an identity for whoever or whatever that is acting. And then it's the bond permission so that action stays within the user authorized. And the next, I think, there will be like statements fast and cheap enough for the machines to be an activity. I think that's why we build a blockchain. And that's why we made it more fast with their point like 10 seconds block and substance gas. I think those were really important. And a very favorable record. So anyone can audit what happened with AI afterwards. So in size without the settlement, it's just information. But in size plus trust execution, that will be an economy. Sounds good. Yeah, I think we've gone quite a bit far with blockchain. Let's see if Chilskill is ready to introduce themselves to us. Chilskill, are you there? And they disappear. Let's continue with blockchain. OK. I'm just going through in notes. Yeah, about chain. For a web3 wallet user who is curious about AI, but has never used these products before, what do you think is the easiest place to start? Like what part of the experience do you think still needs to improve before more people will use AI tools regularly? Yeah. So well, for that question, I think AI nowadays haven't gone that far for mass adoptions. Because I know a lot of web3 users might start using AI already for analysis or even web3, whoever works in the web3. So I think to getting it into our everyday lives, we have to, like with AI first, you have to know what you actually want. Otherwise, a lot of people are saying AI is stupid. So you're going to have to know what you exactly want. And then AI can do the job because at the end of the day, whether you can say AI, skills, or agents, they are bound together, they are definitely