01:06:10 · Brett Caughran, Alistair Smallwood
[01:06:10] S1: Yeah. Yeah, so folks, the application layer does not scale as fast in finance as it does in coding for this very fundamental, sort of very fundamental reason. The good news is once you're wired in, today if I am using Instinct, I can switch to Muse or whatever the next thing is, like the sort of mode around many of these businesses. The coding agent mode is almost zero, but it feels like the vertical layer here will be as large or larger, but really sort of a co developed collective system, which should make for some really high quality businesses that can serve asset managers to help them cross this transition. [01:06:55] S0: Yeah, totally agree, Brett.
01:06:57 · Brett Caughran
[01:06:57] S1: Well, I feel like we, I got to like a third of my list. I wanted to get to modeling and I wanted to get to so many other things, but we're over an hour now. So I think we're going to have to do a part two at some point in a few months. This was really fun to discuss and debate these things with you. And I would encourage people to follow Alastair on LinkedIn because he's putting just some interesting thought pieces. Really like just sort of the... If people ask like, What's the future investing? I don't know. Why are you asking me on Twitter? It's sort of an unknowable, right? But it is what we all as an industry make it. Best thing we can do is sort of study priors and pay close attention to the technology and try to get a sense of where frontier firms are at, frontier hedge funds and long onlys are at. So I think we agree most funds are less far along than most would think. The technology is just coming together. We didn't even get to the sort of the rewiring of the investment process. You have sort of the infosec. Then you have the issue of you're a tech analyst covering 60 stocks, getting that person off the desk away from earnings, away from model building. Sort of be like, Hey, you have to completely revamp how you do everything with these new tools. It's sort of like, leave me alone, nerd. Like I'm trying to basically, have a print in three days. Yeah, exactly. So changing the sort of engine in flight analogy. So it's going to take a it'll it'll take a few years to go through this full transition, in my opinion.
01:08:32 · Alistair Smallwood
[01:08:32] S0: I agree. I agree. And I think that a huge part of it is around educational, the educational element of people really understanding how to use these tools effectively. The massive issue, as you full well know, is that if you're in a risk taking job, you don't have any time. And the people that we have seen really make leaps over the last twelve months are those guys that for whatever reason have been in the garden between roles. And we've even seen users within the platform where I've enabled them in month one of their gardening leaves. And the quality of usage that I've seen from month one through month nine has been absolutely phenomenal in terms of real core understanding of how to use these systems properly. And ultimately, AI is a skill like anything else. It's a ten thousand hours job, and you do have to spend the time to understand and stay up to date with what's moving.
01:09:21 · Brett Caughran, Alistair Smallwood
[01:09:21] S1: Yeah, yeah. And I think at all firms now you're having sort of this early adopter cohort of analysts who have spent the time on nights and weekends that are like, holy crap, look what you can do. The problem is that transmission mechanism from those five people at a 100 person firm, how do you adopt that strategy to the other 95? And like the challenging point is like you're on an exponential here. Like to me only with Opus 5.5 a couple weeks ago did I feel like I could now scale an investment firm. Like I could build a single person investment firm now. Like I'm confident of that based on the system I've built that could actually generate alpha and protect risk. So it's like not moving too fast, making sure you're sort of moving and deploying at the right time when the technology is ready, which I think we think it's, if not today, like close to it. But then it does require this sort of mosaic theory of scale and system type like it required, like it is sort of a different lots of different approaches to doing the job. [01:10:23] S0: Exactly. And the more time that you can spend doing that research and development around those potentially new methods, and they will just plug in as an additional input to the ultimate system that you build is super powerful. And there'll be a huge number of things that we're still yet to discover. Hopefully, over next few months, we can share some of those with you. [01:10:40] S1: Awesome. Great. Well, thanks for your time, Alastair, and good to spend this time with you.
01:10:40 · Brett Caughran, Alistair Smallwood
[01:10:40] S1: Awesome. Great. Well, thanks for your time, Alastair, and good to spend this time with you. [01:10:46] S0: Perfect. Cheers, Brett. Thanks so much. Bye bye.