George Kailas launches Aethon Fund, an AI-powered hedge fund launching in August

want I want some exposure.

Yeah. Get in on it while you can. Well, uh, our next guest is from Athon Fund. George is in the waiting room. Let's bring in the founder and CEO. He's launching an AI powered hedge fund. George, how you doing?

I'm great. Thanks for having me on.

Thanks for hopping on. Uh, please uh, introduce yourself for everyone who's watching.

So, I've been in investing for a long time. I started in value investing when I was 17. And I actually taught myself accounting to get that that job. Uh I lost a little faith in you know the fundamentals you know things like price to earnings ratio as I kind of saw some holes appear got a lot more interested in AI and actually was able to parlay a novel mortgage algorithm that I put together actually in Excel at that time uh into an AI company. Uh that was interesting for a while but in 2019 I started a company called Prospero AI uh because I thought that data was going to be a lot more valuable uh than models uh in the long term and just you know technology overall and you know we kind of had a couple approaches to that. One is we focused on compression. So basically complexity in the back end, simplicity uh in the front in the user interface and then we were pressure testing those you know signals looking at edge cases with the public which I think is one of the best ways to do that uh especially because not a lot of hedge funds uh get to have that kind of testing. Um as we evolved that we also saw some very interesting um data and this is part of our thesis where the more trust we have in the p from the public the more you know follow on and and kind of a self-fulfilling prophecy gets created and a lot of that was the thought what behind what we started with Athon recently and we're basically taking these signal libraries from Prospero we're expanding into like my R&D backlog that you know a startup wasn't able to dip into as much uh and then we essentially feed it into you some of the world's best LLMs. We create thousands of strategies with that and then from those strategies we figure out intelligent ways to you know combine them to work in many different markets to opportunistically allocate amongst them. Uh that's what Athon is. We have a lot of excitement behind like the risk and reward combination that that creates and obviously have a big launch uh for a fund. Um

so so fund is already closed. You're deploying it already. Like where where are you at in in the process?

So we have an agreement with uh you know the fund of funds uh that we're dealing with. Um we have you know some traditional LP capital as well. Um we expect to actually launch in August September at the latest. what uh what data sources are being unlocked this year that were previously either too expensive, too manual. Uh you know, we've heard everyone tells the story of like the hedge fund that's looking at Walmart parking lots with satellites to determine if they're going to beat earnings, but what are the interesting sources of data that you're seeing pop up around the industry these days?

Oh, I mean, I think a lot of that's exhausted. Like I've been talking about real time credit card swipes, camera data, satellite images, you know, for a long time. Um,

I mean, I think some of the interesting ones that I'm starting to hear are coming from like the the medical field. Um, I think a lot more people are starting to use interesting uh data sets there. Um, but I I I think most of the interesting things going on are in this, you know, signal compression. like one of the smartest guys I know on Wall Street um is starting a firm that is focused on like data containerization. Um and I think that's a really good way to look at basically all of these alternative data sets mostly established giving people kind of intelligent access to them that's kind of minimizing their training costs while also getting a broad exploration of the data.

What what does data containerization mean exactly? Uh, I believe it's, you know, it's like a new term that, you know, his carbon arc is the firm that, you know, I've heard from them, but I I happen to think they're very sharp. I've been on a panel with Kirk who runs it. Um, you know, essentially taking, you know, the same concept as, you know, dividing up compute, but basically packaging data sets in these smaller components, not like accessing a huge credit database, accessing it in only the ways that are relevant to the, you know, that current training. iteration for example.

Cool. Uh where do you want the fund to go? Where do you want the firm to go? Do you uh imagine adopting multiple strategies going into other asset classes? How do you see this growing over the next decade?

Uh certainly I think we want to go into a lot more asset classes as we grow. You know the way we kind of look at this is you know we will test strategies within our existing signal libraries. Then we expand our signal libraries and some of that will be chasing down, you know, different asset classes, different arbitrage opportunities, you know, even similar asset classes in different countries. Then we kind of feed that into, you know, the broader system um and do kind of the data compression thing that we're really good at and then say, hey, we might start mixing and matching a lot of these different strategies in kind of these larger waterfalls. and then you know amongst these different waterfalls combinations of you know strategy opportunistic strategy allocations um we might then say hey divide the capital we have between 10 of them 20 of them and just become more and more distributed.

That makes a ton of sense. Well thank you so much for coming on the show and breaking it down the rest of your week and we'll talk to you soon.

Good luck for having me on. Cheers.

Let me tell you about Codeex. Codex is a powerful workspace for getting work done with AI agents. Whether you're writing code, analyzing data, creating content,