00:47:44 · Khe Hy
And so I'll just go into Codex on my phone and say, schedule a meeting with Brett this Friday at 01:30 p. M. And then if I want, I can have it send you the invite as well. And so, like, but I need to be it needs to be on my Codex, right, because the permissions are local to my machine. They're not, like, in the cloud. So that would be one. Another one I have, this is super cool. I call it the CRM tickler file. And so I have a CSV file that has all the people in my network and a frequency in which I would like to just, like, stay connected with them. And so then I have a skill that actually, well, first I have that hooked up on Codex to iMessage because like 70% of my work, like, I don't think I've ever emailed Matt. It's like always been through to you as well. So it's like, what's the point of a CRM if it can't read your iMessage today in this day and age? It's kind of pointless. So I have it hooked up to my iMessage and my Latour Gmail. And every Friday it runs, it basically looks through both sources and sees like who falls outside of the window of like, you know, who do you need to get in touch with? And then it sends me an email of a list of the people that are like past their window. And so I could... But let's say it runs every week, but let's say I wanted to run that on like a Wednesday night and I'm on the road. I could just say, run the CRM tickle skill.
00:49:19 · Matt Stockton, Brett Caughran
[00:49:19] S1: That's cool. Yeah, I've never, I have not gotten into the use the AI from the mobile thing yet. And I'm kind of resisting it because I'm already I'm already spending too much time on this stuff as it is. Right? And so it's interesting, though. I mean, think that's a big part. You're not the first person that said the tech... The texting aspect was really important to them to be able to kind of manage and run run these things. But, yeah, it's it's interesting. [00:49:48] S0: Matt, what are your thoughts on the I've heard a lot about the Chinese models. There's been this sort of wall of like token, you know, token burn, token maxing, token efficiency. Uber ran through their tokens for the year and four months. [00:50:02] S1: But it [00:50:02] S0: seems like people are pointing to sort of the fast follow of the Chinese models as a sort of antidote to that, with sort of routers on top of that. What are your thoughts on that space?
00:50:18 · Matt Stockton
[00:50:18] S1: Yeah, it's interesting. You know, I I have dabbled with running local models on my machine just just as a way to just see how it works, right? Like download some weights and run them. And I don't have a super beefy machine, right? Like, the models aren't great. But the smaller models and some of the open source models, they're getting really good. I mean, if you look at the benchmarks, and you kind of have to take the benchmarks with a grain of salt sometimes, but the open weight models are very effective. And my belief is that most people's work right now, like that they're actually doing, probably doesn't require like the frontier model anyways, right? Probably a lot of the tasks that people want to use is very serviceable by these smaller models. So I think it has legs. And I think, you know, these these smaller models will continue to get better. I think, you know, companies, if they want to cost optimize, this is going to be a route that they go, right? Like, you can already see, like Shopify put out an interesting post a couple months ago. They took a workflow that they were running at scale on a frontier model and kind of, I don't know if they fine tuned a model, but they did something to run it much more cheaply, right? And like, they got a 100x, I think, cost improvement there. So I think that will happen. I think it's tricky because it's much easier, I think, to just like use the frontier model and run after it. But I think as companies have things that they need to do at scale, and it's it's repeatable and they've proven it it provides value like that cost optimization, angle will become more important. What are your thoughts there?
00:52:16 · Brett Caughran, Matt Stockton
[00:52:16] S0: I don't have very detailed thoughts. I just hear people talking about it more. So it seems to make sense to me. I've been a fan of, you know, tools like Perplexity Computer, which is sort of like an abstracted OpenCLO that has access to 18 models. And I find it just to be more consistent than I do a Claude, which is really important for the workflows that I do. And I think people are trying to struggle to figure out Excel as well too. And this common response I get is that, shouldn't try and update a model with Frontier Intelligence. A lot of that should be done in a coding environment. So how do you start to think about actually deploying more deterministic use cases at potentially local low cost models and use frontier intelligence for a stack of things when you really want think through a debate or apply judgment and sort of think deeply with the strongest intelligence. But to deploy frontier intelligence across a full stack of the sort of agentic investor process doesn't feel like, one, it doesn't feel like the most effective methodology. I don't want a stochastic nondeterministic tool in the quantitative realms of my investment process. So I think that's part of why I've been more excited with agents. Like you see a little bit of glimpse of that of like the LLM as the orchestrator of these various tools, calling sub agents, spinning up coding environments, etcetera. [00:53:45] S1: But
00:53:45 · Matt Stockton, Brett Caughran
[00:53:45] S1: But [00:53:46] S0: that's about as deep as my knowledge goes in that space. Feels like the spinning up of sub agent. That's when I started to... Gave up on Cloud Code when someone's like, Well, to actually do that, you have to spin up sub agents and do all this. I'm like, I don't know what I'm doing. I'll just find a tool that can automatically spin up sub agents, which perplexity computer does. And Codex has gotten much easier to spin up many, many sub agents in Codex as well. [00:54:14] S1: Yeah. Yeah. [00:54:18] S0: Well, has been a wonderful, Matt. Almost certainly this conversation will follow Sutton's bitter lesson curve. We'll probably be laughing at the conversations we're having here in twelve months. Inactionistic were those conversations we had about sub agents and Chinese models, etcetera. But thank you so much. So we'd love to have you back on and sort of stay in tune with the frontier here. Please keep tweeting because it's part of my educational stack as well too. For sure. You're right [00:54:55] S1: on Twitter. It's like, if you want to be in the loop, you have to be on there, but you also have to ignore a lot of [00:55:02] S0: things that [00:55:02] S1: are on there. [00:55:04] S0: Yeah, exactly. [00:55:05] S1: And you got to put the phone down once in a while.
00:55:05 · Matt Stockton, Brett Caughran, Khe Hy
[00:55:05] S1: And you got to put the phone down once in a while. [00:55:07] S0: Oh yeah. I tell people my rule is just like, for... I hope Andre Karpathy can still tweet at Anthropic now because I'm like, Just wait for Andre Karpathy to write about something and then like spend a couple of days thinking how that applies to finance and investing. That's sort of been my like ironclad rule of staying ahead on AI and finance. [00:55:30] S1: Yeah. You got to find the people that are talking about the right things and just pay attention to them, I think. [00:55:35] S0: Yeah. Great. Well, thank you so much. This was really educational and entertaining. [00:55:41] S2: Yes. Thanks, Matt. Thanks for having me on, guys. Appreciate it.