DeFi Meets AI: Autonomous Protocols, Intelligent Yield & the Future of On-Chain Finance

to when the circumstances change. And think about someone managing a position, they may be watching the market and asking whether a higher return somewhere else is worth the cost of moving their funds. A model might help them work through that information. And the interesting part is what happened next. Does it offer a suggestion as for approval or have the permission to act on its own? Each of those choices creates a different relationship with the person using it. Never having over more responsibility, we would want to understand what the system is trying to achieve, how its decisions are evaluated and where someone can step in if something goes wrong. So, in today's conversation of life is stayed close to those practical questions, where AI is adding something useful beyond ordinary automation and also what evidence would make us trust strategy or a risk model. And how do we keep the experience understandable as the technology becomes more capable? And for anyone who is new to fortune, we are an AI power prediction derivatives hub. And our idea is to bring the prediction market liquidity and tools around a trading decision into one place. So, the experience is easier to follow. You might have a view on an upcoming event, but choosing that view into a decision takes a little more work. You need to find a relevant market, understand the price and see what liquidity is available. And fortune brings access to prediction market liquidity together through one interface. So, people can explore those opportunities without moving with separate platforms for every step. Alongside of that, fortune agents help with the information around the decision and they monitor news, research events, also examining sentiments, probabilities and the other books. And our consensus engine brings multiple models into the process to cross check their views. And what interests me is how those different inputs help someone examine a market more tactfully. And for me, a forecast becomes useful when you can understand what it means for the choice in front of you. And what is the view based on does the market price reflect it? And what could change the assessment? Those are the kinds of questions I want to ask to keep asking as we feel. That is why I'm glad we are having this conversation today. There is a lot of things we can learn from how other teams connect intelligence with a real financial decision and how they explain their responsibility that comes with it. And with that, let's meet our guests today. Each of you brings a different perspective so I'm looking forward to learn a little more about your work. Okay, that's all for our introduction. Now, let's move to our speakers today. To get started, please tell us your name, your role in the project and a little bit about your background. I would also love to hear how your words connect with today's topic, whether you approach your true markets infrastructure, research, or the people using the product. Let's start with our busy group. Hi and welcome back to the space. It's lovely to have you here. Yeah, here's a name. Thank you. Thank you for having us, that's really nice to be back here. Could be a topic is really interesting. Hey guys, I'm our boss from RBC district. A quick review of our busy. We are basically building a wealth ecosystem around real world assets, starting with gold. Our focus is bringing physical asset on chain with transparency and security while keeping the ownership verifiable on chain because the real world asset needs to be verified. With the asset on chain and we are very focused on transparency and custody wise. And to be honest, really excited to be here. I am no AI engineer, but I can bring in the business perspective and the marketing for this conversation. And thank you so much for having us. Okay, great amazing introduction. Now, let's move to Mary, the NA. Welcome to the space. Hello everyone, thanks for having me and thanks to fortune global for putting everything together. So greetings from British Columbia, Canada and my name is Harry. I'm the CEO and Meridiax. And Meridiax is a stable one based middle settlement layer for the global aviation industry. So we connect to upstream airlines and downstream online travel agencies such as booking.com, trip.com, a goda, etc. And we earn the spreads in between. So in very simple terms, we are striving to make your daily airplane tickets cheaper. Let me introduce my background real quick. I studied math in college and I grew up in a family which has been investing in the secondary stock market for decades. And personally, I've been active in the US stock market for just a few years. And of course, I think AI is making defy smarter and more dangerous at the same time just for today's topic. Because the danger comes from thousands of smarter AI agents that all agree with each other. So if every agent is trained on the same price fees and in the same on chain data, they will all had for the exit at the same moment. And so like that's pretty much how it turned a deep into a crash just like how algorithms triggered the black Friday in 1987. Thanks for having me super glad we could join us today. And then we have from a rationalize word. Welcome to the space. How are you today? How are you doing? I'm Charles Darwin, Fractionized World. I've been around the crypto space since 2017. Mostly investing to start with then went into defy and literally we've we've got the build stage up in planning and build in Fractionized World since 2022. And we're finally getting to the point where we're actually bringing it into the into the public domain. So we'll be launching in January and we're just about to start our launch event which lasts for 80 days. There's an event called around the world in 80 days. So that starts next Wednesday and goes all the way through to our launch. We will be speaking about everything we can do. We've got a lot of aspects that we're bringing into the new defy space. So we're bringing in a real world fusion type things. So we what do we do we we buy real world assets and we mix it up a bit and we bring back and give some yield back to back to the supply back to the investors. So we've actually got participants on chain. We've got real world assets off chain somewhere in the middle. There's a yield to be had. So I got a compliance man. So I've got I can't explain the exact sort of it but there's there's a lot that goes on in between and everyone everyone gets sort of like happy at the end of it. So nice to be here today. Thank you. Right. Looking forward to explore that with you today. Now we have also from the white.ai. Welcome to the space. Hello everyone. I'm the community manager from PWTI. I'm this great to be here today. Thank you very much for inviting us. PWTI we are an AI powered media network that owns news podcasts or do and long form stories into digital context. The last forever. We use smart AI tools and decentralized technology to to keep your content safe from censorship. We keep it easy to find and we keep it permanently accessible to everyone. If you'd like to learn more about what we do at BWC feel free to check out with our profile. And you can also check out websites at the BW.Paymarkcast.app. Thank you very much. Okay, thank you. And lastly, we have from reactive. Welcome to the space for the first time. How are you today? Yeah, thank you very much. Thanks for having us. It's a pleasure to be here. And for quick intro, I'm Amelia some leading ecosystem slash partnerships as reactive network. And just to give you another view, we're on EVM automation layer and the whole network is based on the concept of reactive contract, which subscribe to events across any supported chain. And they can fire across chain transactions on their own without using the typical automation tools as we know today, such as boss or keepers or really. So there's nothing on that in the middle. And the topic of the day was very interesting to us just because we are sitting at this AI and defy intersection from the infrastructure side. So we're not building any model that makes any decisions, but we are kind of building the rails, which let an autonomous system actually act on chain without you having to trust some kind of an operator. So this quick intro about us and just to lean on a little bit on the topic of today, I do agree with Harry that the fire is getting smarter, but also more dangerous because of AI just because the same AI that can, for example, you know, rebalance your vault. Can actually drain it as well and you just have to trust today that it won't and the bottleneck is not the smarter models. It's actually the execution of those models and this is what we're trying to solve on our side back to you. All right, thank you guys for the introduction. It's good to have all of these different perspectives around the table. Now let's begin what we mean when we put AI and defy together and work through the financial decisions and also the responsibilities behind them. Please raise your hand whenever you like to answer or add another view. I'll bring you in one at a time and invite people directly as well. There's a lot of room for different opinions. So feel free to explain where your experience diverse and if we get technical simple example will have everyone to follow. Okay, let's start with the first language we use when someone describes a product as an AI power. I want to understand which decision the model is helping with and why it belongs there. It gives us a useful place to begin. AI power divide is used to describe everything from simple automation to sophisticated ML system. What actually separates the genuine AI integration from a rule that scrap with an AI level and what should builders and users look to tell the difference. We would like to get us started on this one. Feel free to raise your hand and bring you in. Okay, Harry, are you okay? Here we go ahead. Yeah, that's a very good question. And honestly, I think it's the most important topic today because the label is on everything right now. So here's, I think it's the simplest way I think I mean, I consider it. So a rule based script always says if the rate on Abby, for example, goes above X like it moves the money and then somebody wrote another value, for example. And so the value acts never changes unless a human changes it. So that's automation. So it's useful, but it's not artificial intelligence. So real AI integrations when the system learns from data and changes its behavior. So the threshold is a hard call it, I believe. The model keeps reinvesting it from volatility liquidity utilization and it has to update as the market moves. And here's some data that I found interesting, DW adventures without a report back in April saying AI agents now originate around 19% of the volume. That sounds a lot. But like if you break it down and most of that, M EV maximum extractable extractable value is from bots like and liquidity, but liquidation bots arbitrage. But stuff like that doesn't exist along before anyone called it AI. So that's my take pretty much. So I feel like right now they're still in already status of utilizing real AI transactions for transactions. And of course, like AI is great for coding coding and that's my take. Thank you. Okay, thank you, Harry. I appreciate you taking the question seriously and it's planning how you draw the exit that extension and it gives us a useful starting point for the rest of the conversation. And thank you again, Harry. I appreciate you sharing your thoughts. Let's bring another perspective into the conversation who is praising their head. Okay, I basically grew up here next and then we're moving to reactive network. Good.