00:19:40 · Joe O'Donnell
I think, say, something that I think most people do these days would be something like, you know, summarize an earnings call for me or something like If if one were to compare the actual earnings call to the AI summary of that earnings call or if it... Or or an AB test of take my best analyst, have them look at the earnings call, summarize it, and then compare it with any model or, you know, system of experts or whatever, they're not equivalent off the shelf. And I question what's actually achieved by doing something like that if it's lossy. You know, I think... And that's just one example. There are many other ones. But, you know, at least in my own investment career or any of the investors that I know, no one ever looked back on a year that didn't go as planned and said, you know, if only we had read more earnings calls or something like that, or if only we had looked at more ideas. It it actually... Almost always, it was much more simple than that. The year didn't go well because a few times a year, we had a big decision to make. We made the wrong decision. And probably that... One... Another reason why we steer away from productivity based things because those things, I think, including today's models, all the best ones, they're lossy. And what would I tell one of my analysts from an efficiency perspective about... Let's take the example of earnings calls again. Read the earnings call. Read the transcript. Listen to the earnings call. That is the least lossy thing. And AI sounds super smart, and it is increasingly smart. But it's still lossy. And in many things that people use it for, it sounds like it has good judgment and doesn't, or it's just has different judgment. And those things can be dangerous. It helps people... It helps investors build conviction when they actually should have lower conviction if they're using that for their investment research.
00:21:46 · Khe Hy
[00:21:46] S2: Is it fair to say that there's kind of like a hierarchy of judgment? So for example, one judgment could be like what data sets to use and another set like layer of judgment could be, should we even be in the sector? And like, and then the lossiness angle is kind of a new one that I hadn't really kind of put into like the layer cake of judgment, but maybe pulling on that thread of, you know, earnings calls and the lossiness, like where is the judgment strongest? You you gave an example where it might be too lossy, but where where does it work well? The judgment layer.
00:22:25 · Joe O'Donnell
[00:22:25] S0: Yeah, I think specifically for AI, I would say that there's tons of utility in generalized investment research for AI. So I don't want to say that it's very clearly very useful. But I would say that the overwhelmingly best use case for off the shelf AI tools is information gathering purposes. Sort of a couple orders of magnitude better version of vertical search. And there's tons of value that can be generated from that. You know, just finding things you want to have otherwise found and then doing deeper research as a talented human or whatever from there. That kind of stuff I think is very, very useful, especially across enormous and increasingly large data sets that exist in the world for investors. For judgment purposes, what to work on? Yeah, I think for sure, we've built, to pat ourselves on the back, we've built some very useful idea generation tools for... That that use AI to help people discover new investment themes or things that align with their thinking on on specific topics that incorporate our proprietary datasets and have more generalized search and judgment tools to help people figure out what to work on. And then we're, you know, as I mentioned just now, we're building basically the stronger form version of that right now. Well, no one... You know, for... Take take idea generation. Nobody really wants a list of, you know, 14 companies that meet their investment criteria. What they really want is of the 14, which ones are good investments? That's the ultimate goal. And so that's the purpose that I I would say our idea generation agents are effectively the final boss version of investment idea screening. Yeah. Instead of having, you know, 14 companies on a list, you get somewhere between whatever, call it, zero and three finished investment reports done at the capability level that one would expect from a good analyst. That's the right way to screen. And and so, yeah, I think I think it it really depends on the use case. AI for judgment related purposes, would say to a large extent is still an unsolved problem.
00:24:43 · Brett Caughran
[00:24:43] S1: Joe, I'm with you on the the the chatbots being not that helpful across the institutional stack and primarily good for information gathering. But I think there is this belief now that with Claude Code, that funds can build a whole new stack of internal tooling that they can deploy Cloud Code to build a live earnings preview or an idea generation agent based on their own historical trades and ideas and thesis work, etc. I still wonder whether that's a prototype or whether that's production grade to actually deploy institutional capital. Could you talk a little bit about that movement, where you think that could work, and where you think that approach may have issues?
00:25:33 · Joe O'Donnell
[00:25:33] S0: Yeah. I mean, I think, you know, there's something against the laws of physics for an investment organization to build, you know, canary inside of their own organization. I'm sure, you know, most investment organizations have a resource set and a team of extremely talented people that could go reconstruct a large percentage or all of what we do. The level of effort required to do that would be extremely high, and I suspect that some people will decide to do that and others will not. I think what we found at least is that in order to, you know, say whatever, our firm's investment process is X, In order to codify that into software, it's nontrivial. I can... Everyone kinda knows what it is when they when they when they think about it or when they look at an individual investment. But but but then when you actually try to turn it into some sort of deterministic logic that sits in a software program, it's not so easy. You know? People plug in a set of things that they think they like, and the output from whatever the... Take your idea generation example is nothing like what they actually think they... What they actually do like. And then there's this iteration process. You know, sort of like automated driving where you can you can, you know, whatever. Maybe this is harder than understanding it, but you get sort of 90% automated in a really easy way. And then the the next 10% end up taking five times as long as the first bet. And, you know, some people will make that investment and maybe they'll build great products and and that... That's great. I think in order to get to that level, it still requires all the infrastructure that I mentioned too. So do you have the right data, which requires building integrations with who knows how many data providers, but that by itself is not sufficient. Well, then what is... What do each of those individual datasets actually mean? And every single one of those needs to be fine tuned. Every single one, at least as far as I can tell. There's nothing that's really just obvious, you know, plug in dataset x, and then a foundation model just intuitively right now understands it at the level of a really good investor. And in order to do that next layer, well, you got to get your really good investors that should be working on investments to go train all those datasets. And that's an ongoing process.
00:27:56 · Joe O'Donnell, Brett Caughran
And then you can build the emergent sort of, you know, what I... Maybe I would call microservices on top of all that stuff. And then you can build that emergent layer of investment analysts. It can be done for sure, but it requires... Well, I can speak for ourselves. It requires years of investment and many thousands of hours of fine tuning and data collection and cleansing and aggregation and stuff like that. And I think every firm has to decide basically what the demarcation line is between what they wanna do themselves and what they wanna outsource to someone else. And there's... Yeah. Every... You all have different answers. [00:28:34] S1: And how's that work today? Because where you may get pushed back is the argument that while correct that most public investors don't have alpha, no public investor believes they don't have alpha. So I think a lot of the mindset with agents is I want to clone my process. I a want to digital analyst that acts and behaves like a senior analyst or a junior analyst on my team. And to that degree, agents have been interesting as a flexible substrate to be able to actually train that analyst. How do you work with funds? Like if I gave you an analyst training manual and said, Joe, make this agent behave like a junior digital analyst. How far away from that reality are we today?