They Sold a Startup to LabCorp, Then Built a Value Fund | Top Mark Capital

Join the Capital Alliance Waiting List Here: https://linkly.link/2ebz3

In this episode of the Capital Alliance Podcast, Ken Majmudar sits down with Mike Nicoletti and Jason Wallace, co-founders of Top Mark Capital, a San Diego-based investment firm running a concentrated generalist fund and a dedicated healthcare & biotech fund.

Mike and Jason took an unusual path to fund management: they were operators first. They shared hard-won lessons from building (and shutting down) startups — including the painful realization that "winners stay alive long enough for good things to happen" — and how selling a clinical trial technology company to LabCorp shaped the way they analyze businesses today.

The conversation goes deep on what value investing actually means in 2026. Mike argues that value principles get arbitraged away over time — Graham's net-nets are gone, Buffett's float and intangibles edge got absorbed — so the real challenge is figuring out what comes next. For them, the answer runs through AI: collapsing switching costs, falling barriers to entry, and why the market may be mispricing great software businesses anyway.

We cover:
- From tech consulting and startups to launching Top Mark in 2012
- Why an operator background makes you a more patient investor
- How value investing principles get arbitraged away — and finding "the next thing"
- Hamilton Helmer's Seven Powers as an evolution of Porter's Five Forces
- Why AI is crushing SaaS switching costs — and why companies still aren't ripping out Salesforce
- A software company is "a sales organization wrapped around a product"
- The AI momentum market: memory stocks, data center bottlenecks, and vertical price action
- Are we living through a second industrial revolution? GDP, jobs, and the UBI debate
- The biotech funding winter, the thaw, and why ~40% of pharma in-licensing now comes from China
- Using AI in the research process — without outsourcing final judgment on the 10-K
- What a peer community means for fund managers outside the New York bubble

Follow Mike Nicoletti & Jason Wallace:
Top Mark Capital: https://topmarkcapital.com/
Telltales Podcast & Substack: https://telltales.substack.com
Mike on LinkedIn: https://www.linkedin.com/in/michaeljnicoletti/
Jason on LinkedIn: https://www.linkedin.com/in/wallace21

Follow Capital Alliance

Capital Alliance on LinkedIn: https://www.linkedin.com/company/mycapital-alliance/
Capital Alliance Website: https://mycapitalalliance.com/

Follow Ken Majmudar:
Substack: https://compoundideas.substack.com/
LinkedIn: https://www.linkedin.com/in/kenmajmudar/

Transcript

SPEAKER_01 0:00

Less than a year after we shut the business down, our potentially largest customer called us up and said, I've got your contract waiting now. One lesson we learned was the winners stay alive long enough for good things to happen. And we didn't stay alive long enough.

SPEAKER_00 0:15

As a value investor, you need to recognize that what's happened in the past maybe isn't gonna give you superior performance over the long run. The challenge is determining what is the next thing. And then they all get taken to the woodshed, obviously, because this now alien technology drops into Earth. Means intelligence sort of gets commoditized. Dario's just been outright hostile, though, to say that this is gonna kill jobs and there's not gonna we're gonna have to go to universal basic income.

SPEAKER_01 0:41

Uh prior industrial revolutions happened. We had a really step change in GDP. I don't know where GDP growth goes, but maybe we grow 5% every year for this foreseeable future.

SPEAKER_02 0:59

This podcast is sponsored by the Capital Alliance Mastermind, a private community where fund managers share deal flow, build relationships, and grow together. Join the wait list in the first link below.

SPEAKER_03 1:09

Hello, everybody. Thanks for joining us again today. Um, it's another episode of our Capital Alliance podcast. Uh, this time we have two good friends and very special guests, Mike Nicoletti and Jason Wallace, joining us from Top Mark Capital in beautiful San Diego. As you can see the difference between their background and my background in New Jersey, where I'm sitting right now in my office, it's a little different. Probably similar weather, though. It's actually a nice day outside. I know Mike and Jason are obviously they run this great fund, which they started in 2012 called Top Mark, which has two funds that we'll get into. One of them focused on healthcare, uh, and the other one sort of just more generalist focused, I guess. Is that the right way to think about the second fund? They're very thoughtful. They have a great podcast, uh, which is called Telltales that I would highly recommend, which they do uh with their mentor, Hunt Lawrence, who's who's uh a very successful investor uh and someone uh who has built a tremendous amount of value. I I guess originally his thing was that he was an energy investment banker, but he's full of wisdom. Uh and also uh Mike and Jason were in Capital Alliance, where I've gotten to know them really well over the last year or so, uh, are in Capital Alliance, and um so yeah, we've we've gotten to know each other really well. They have really impressive and thoughtful uh commentary on so many things. I'd highly recommend you follow them and follow their substack. Uh so with no uh uh no further delay, I'd like to turn it over to Mike and Jason. Thanks for joining us for this uh episode of the Capital Alliance podcast.

SPEAKER_00 2:42

Yeah, thanks, Ken. Uh looking forward to it. Uh I guess do you want me to give a little background on TopMark and how we got started?

SPEAKER_03 2:49

Yeah, that'd be great. That'd be great. Especially since, I mean, you sort of including myself. I mean, I have a I have a background. I guess you know, I studied computer science undergrad and then law before I became an investment banker and ultimately a long-term value investor that focuses on compounding. So I think you guys have sort of unconventional backgrounds as well. So maybe that's where uh what work that into how you got into what you're doing now.

SPEAKER_00 3:13

Well, let's go all the way back to the beginning then, Jason. Uh Jason and I met in high school and we were both in computer science class together, right? So unconventional. Similar to uh Bill Gates and Paul Allen. I'd I'd say that's a good good comparator. No, we're on the same project now. One of us is Bill Gates. How much is Paul Allen? Yeah. TBD. College, I studied, I you know, had this computer science background, but I ended up getting into econ, and then some professors figured out that I could write some code, so I got dragged into econometrics. So it's always kind of had an interest in the way the economy works. Um after college, worked for a tech strategy consulting firm uh based out of Sweden. Um got to live in Sweden for the beginning of my five years there and back and forth between uh Northern Europe and the US for the remainder. I then took a couple years off and raised sailboats professionally. Um and when I did that is really when I got really interested in investing. Uh doing the consulting work, I wasn't in love with the work so much, but it enabled me to pay off student loans and give me some capital to start investing. So I was consuming books on the investing, everything from technical trading to Warren Buffett stuff. So I kind of didn't have a clear direction. And that all changed when I took this leap to start racing sailboats professionally. I got uh teamed up with a guy based in who had got some funding out of New York and to do a double-handed offshore campaign. So he and I joined forces and the sailing center that we were racing out of, uh it's called the Oak Cliff Sailing Center, and it's in uh Oyster Bay, New York. It was a new kind of concept at the time, but uh the long story short is they were attempting to kind of reinvigorate the uh offshore sailing and I guess inshore and match racing uh prowess of the United States, kind of a center of excellence, if you will. While I was there, I stumbled into a group of people significantly older than me. I was in my 20s at the time, all sitting around talking about stocks. I said, Hey, that's kind of interesting. And I introduced myself and asked if I could be a fly on the wall in this meeting. And uh that led to what was sort of a grounding of fundamental understanding of how I would want to pursue investing going forward. And to tie it back to the podcast, that is actually the source of the podcast. So, my mentor, that is a meeting that he was running. He was the person I met through that. So that encouraged me to think about going to business school and switch over to something more in the finance realm. And with that, he said, Well, if you're gonna do this, if you're gonna go to business school, study finance, I think you should just start your track record now, start a fund. So he became my first investor. In 2012. In 2012, right before I went to business school.

SPEAKER_03 6:08

Pretty good, pretty good first investor to have.

SPEAKER_00 6:11

Yeah, yeah. Well, uh, I mean, part of it is I I also was relatively ignorant to what the whole running a fund would take. Um and with his hundred thousand dollars and my ten thousand that I had to my name, that's what really started the fund. And as you know well, that's not really enough to to launch a fund on. Um but I wasn't apparently it is yeah, yeah, yeah. Well, I I I think that is sort of the takeaway, is it if if you build your company, build your fund to be not dependent on a certain amount of fees, there's nothing that can prevent you from keeping fund going. And so Jason and I, we went to high school and college together, went our separate ways in our careers, and meanwhile, you know, I went to business school, started this fund, really starting to understand the way I was investing. We got reunited. But before I took we kind of get to that, maybe you should give your background.

SPEAKER_01 7:07

Sure. So my my degrees in aerospace engineering are a little unconventional, and my whole path to becoming an investor is very unconventional. So as one does growing up in Northern Virginia when you graduate college, you become a defense contractor. And my first job was calibrating government satellites. And and in that, that was the one year I used my my aerospace background. But in that job, in that role, I taught myself the program because controlling the satellites from the ground is all software-based. Taught myself professionally how to how to write software. And I really wanted to move west. So my goal was to live in Denver, which is also a big defense town, and eventually got recruited into joining the intelligence community and worked for the NSA for about seven years. During that time, it was a super rewarding experience. It was basically the Obama administration, and we were still fighting wars in the Middle East. So we're we were building a lot of data analysis systems to process a lot of the communications and data captured in the war zones. But following that, that's when Mike and I's story kind of merged again. So this was 10 years past college. I was leaving government work, which had brought me to San Diego. And we had a mutual friend that was leaving Palantir and he had an idea for a startup. And he reached out to Mike originally and was asking him if he knows anyone kind of in that space that's entrepreneurial and would be looking for something to do, and put that third person and I in touch. And we eventually decided we need Mike's help to build all the business models and help us pitch VCs in Silicon Valley. So we roped him into flying out from the East Coast. You were living at Newport. It's the only time I've ever called in sick for work. Yeah. So we we put him on a red eye out to the West Coast. We went to San Francisco, pitched a couple VCs, ended up getting some funding, and really forced him to quit his job and join the join the startup community with us.

SPEAKER_00 9:04

That was long enough ago when people started companies in the same place, like everybody moved to the same place, right? Yeah, right. So nobody wanted to move to Rhode Island. That was pretty clear. Uh-huh. That the consensus was we could we could make San Diego work. And what was that a defense tech firm or with the startup?

SPEAKER_01 9:23

Yeah, it was we were using we were very early. We were using machine learning, computer vision specifically, to analyze like safe city camera data, okay, helping police look for you know self-crimes. Got it. But it was extremely expensive to run that at the time. Palantir this guy left Palantir for this? Correct.

SPEAKER_03 9:42

Okay, so Palantir has been around for quite a while. Because what year would this have been? 2014, I think. Yeah. Yeah. Okay. So you had the fund going already.

SPEAKER_00 9:51

Yeah. So the fund was already going at this point. Um I I I started before business school. So a couple years of business school, um year working, and then and at this point, like underperforming pretty solid solidly uh for those first few years. Because I I'd like to say I was still trying to find my footing as to what type of investor I was going to be. Yeah. And it all started to click when we started this company. What Jason and I liked about or like about venture is the venture capitalists are incredibly good long-term thinkers. And which is sort of counter to what maybe your traditional value investor would say about venture. They think it's just gambling or or whatever. It takes a very different portfolio approach.

SPEAKER_03 10:39

But it's an interesting perspective. I I would I would say, like, definitely on themes and general kind of trends and things, they must be very strong. But at the same time, my observation, I'm not that involved in venture, honestly, but uh is is sort of like, oh, you just bet the jockey. So it's like, you know, because even the idea is probably going to evolve, it's probably gonna pivot. So it's like, don't worry about that too much, you know, as long as it's like just big TAM, good jockey, you know, go ahead and power law outcomes, right? Like and power laws, right, exactly.

SPEAKER_00 11:10

Exactly. So that's great if you're the investor and you have access to the right deals, but it really sucks if you're the jockey.

SPEAKER_03 11:17

Right. And yeah, I mean, if you would say that probably given the power law outcomes, I mean, it really comes down to access, right? Because there's probably some general consensus of like what jockeys they want, which is obviously someone who's got some drive and experience, and you know, it's fine if they failed, it's it's even better if they've succeeded in the past. You know, they're probably looking at that team more than anything, no?

SPEAKER_00 11:44

Absolutely. But but the end result for the team is that if the idea doesn't work out, you're back to zero and you're starting to get out of here. Of course. Yeah, yeah, sure, sure. During these first uh, you know, these first few years of building that company, I know Jason's interested in investing. So, of course, I show him what I'm doing nights and weekends, and we start working on stuff together. And this vision of how do you marry the things that we like about venture capital with what we like about value investing in a way that enables us to run a concentrated public equities portfolio started to take form.

SPEAKER_03 12:17

And so what year did you guys and before we get there, uh you I just want to get on the record. Like, so where did you guys go for college and business school?

SPEAKER_00 12:25

And oh yeah. Uh well, we both graduated from Virginia Tech. Um Virginia Tech, okay. Right. And then I went to business school at the University of Rock.

SPEAKER_03 12:33

Computer science major. Computer science majors?

SPEAKER_00 12:36

Uh no, Jason, you were Aerospace Engineering. And I was I was uh economics.

SPEAKER_03 12:42

Got it. So you are an actual rocket scientist, basically, Jason. Right, right, is what you're saying. 100%. Yeah. Okay. I have this, I have this uh one of one of my uh favorite quotes, not favorite quote, but I just something I say just almost half off the cuff is nothing is rocket science except rocket science.

SPEAKER_01 13:00

Yeah.

SPEAKER_03 13:03

You've got it all covered. Got the non-rocket science and the rocket science covered. Yeah, exactly. So uh so that's great. That's very helpful. And and so it's been 14 years you guys have been running the fund. What has that experience been like? What have you learned along the way? Uh yeah.

SPEAKER_01 13:20

Talk about successes and mistakes or challenges. So when we started the the startup and we realized we wanted to work on investment strategies as well, we would we would show up to the office early. You know, most startup people will show up between 9 and 10 a.m. So we would get in a couple hours earlier into the office, work on our investments, building our strategy, and then when the rest of the team showed up, we would start our day job.

SPEAKER_00 13:47

Which is probably a sign that we you know we're better suited to be investing, anyways.

SPEAKER_03 13:52

And and how long before you the startup, whatever happened with that, and then you know you guys now are 100% focused on the fund, right? Or funds, yeah.

SPEAKER_00 14:02

Yeah, we we did we did a number of startups. So the first one did not did not take hold. We it's a combination of being too early and at that point in time, so 2014, 2015, uh, 2016, it was it was very difficult to sell into it, it it's very difficult to sell into government today. It was very unpopular back then to even consider doing a startup that sold into the government. Right. Valentir was basically the only one that had proved it out. And at that point, it wasn't yet proven out.

SPEAKER_01 14:32

But one lesson we learned was the winners will stay alive long enough for good things to happen. Yeah and and we didn't stay along alive long enough. But the good thing happened less than a year after we shut the business down was our potentially largest customer had a several million dollar deal for us that he called us up, I don't know, nine, ten months after the business closed doors and and said, I've got your contract waiting now.

SPEAKER_00 15:02

That's yeah, that's what start it's it's really a it's a duration mismatch, right? So yeah startups raise capital for to have 18 months to two years of runway. And then but the the timeline to develop a government contract is probably four years. So this particular contract we had been working on for a while, and uh and sure enough, it did Did he not realize that you guys had like closed down like nine months earlier? Yeah, well, we we had another uh uh we had a lifeline uh from a city uh to be to to bridge us there. And unfortunately for a combination of reasons, they dropped it um because they got sued by the ACLU for certain uh surveillance practices unrelated to us, but anything that had anything to do with surveillance got dropped. So it was sort of like uh you know, this that that put us in a position where it was just no no longer able to make it go forward. And and that's the thing, as Jason said, any startup is generally seated in a future reality that has to occur. But in order for the startup to capitalize on it, it has to stay alive long enough to get there. And when we go through these funding cycles, like you know, during COVID, you could get money from anybody for for any idea in in venture, and then everything turned in 2023 and uh a lot of companies had to fold. It's not many entrepreneurs have lived through multiple cycles and be able to actually see that and recognize it.

SPEAKER_03 16:34

And so what you guys um that one got.

SPEAKER_00 16:37

Yeah, we close yeah, we closed the doors on that. The next one we did was what so what year was that?

SPEAKER_03 16:43

16. Probably 16.

SPEAKER_00 16:45

Within a year or two.

unknown 16:46

Yeah.

SPEAKER_00 16:47

Okay. The second company we built was in the clinical trial space. So similar technology, machine learning applied to clinical trial data with the idea of helping build better clinical trial protocols. It was early. Obviously, doing that today is relatively trivial. Um, we were using natural language processing to basically tokenize clinical uh clinical trial protocols and provide recommendations for pharmaceutical companies that were running trials. That company did end up selling, but not at like a great exit. Um, it was came to a situation where we were, again, you had to keep uh, I guess it was a good learning that we had from the first one. Is like we knew we had to keep the company alive in order to be around. And we got to the point where we had too many people and not enough incoming revenue. So, really, Jason fired himself, I fired myself, and we left the my co-founder on, and she was able to keep it going to the point where they were able to sell it a few years later.

SPEAKER_01 17:49

Okay. Any other startups together? So when I left that, and and San Diego is really a town of defense work and biotech. So it's kind of natural we migrated into the biotech industry. And so the next startup was really around facilitating the the conduct of a clinical trial and capturing patient data. So this was mobile apps to allow patients to record outcomes during the clinical trial without having to go to a clinic, without having to fill out paper, because at the time most of the industry was still run on paper. And and this was now we're in 2019, going into 2020, COVID pandemic starts. We quickly got inundated with an incredible amount of business. Everyone patients could not go to clinics anymore for anything that was not emergency, conduct their clinical trials. Everything had to shift digital. So we had we have more work than we could than we could fulfill. Um wow. And that was really a a wild time. It was a roller coaster. Everyone, you know, is sitting at home, sitting on their hands, not doing a whole lot, where I'm working 20 hour days and not sleeping and and eating burritos all day. So detriment to my health. But um we were involved with several really interesting trials, COVID-related. The most interesting was Regenerons, a monoclonal antibody study. So this was the IV bag of antibodies that you could get if you had a case, if you had a hospitalization for COVID, and and it really jumpstarted your immune system. So that was that was a wonderful drug. I think I don't think anyone in that study that was hospitalized for COVID ended up passing away. What ended up happening to that drug? It wasn't distributed wide enough. It was approved. I think it was the first treatment that actually did receive approval. It just wasn't, they didn't have the reimbursement mechanism. The way clinical back to the way clinical trials are designed, they pick a very narrow set of patient population that they know that they have a good idea at least that that they can get this approved. So then the FDA approves it, and then you can basically only market it to that patient population. So they didn't take the emergency measure and say, like, hey, this works. Let's give it to anyone. That didn't occur. I think that would have been a huge benefit for the nation if that did occur.

SPEAKER_03 20:17

Yeah, interesting. Um, and what happened to that company? It was Regeneron.

SPEAKER_01 20:22

No, no, I mean your your political startup. Yeah, we we uh we ended up selling that to Lab Corp. Got it.

SPEAKER_03 20:30

Okay. So that was a successful exit, relatively yeah, that's great. Yeah, so let's pivot back. So obviously, you know, Buffett has this saying about being a businessman and an investor and how they're symbiotic with each other. I guess uh you guys want to reflect on that.

SPEAKER_00 20:45

Yes, starting as an operator versus an investor, I think investors tend to have less patience. I think that's even even true maybe in value investing circles, in that it takes a long time for a strategy to come together. A startup in particular will take a very long time. But if you're turning around or changing the strategy of a large company, I mean that they move slower, the results are not going to be clear early. So I I do think there is something to that is having an operating background is probably, if not a prerequisite, a really beneficial thing to have as an investor.

SPEAKER_01 21:23

Yeah, for sure. And I think the two of us having different backgrounds and in very different ways we look at a problem solving and analyzing a business is really a benefit as well. It's not only having a partner, so you have someone to bounce ideas off of all the time, but coming at the at it from two different angles.

SPEAKER_00 21:40

Yeah, and and frankly, our returns when we started working on stuff together went way better. Yeah.

SPEAKER_03 21:46

So talk about the current funds now where how they're how they're positioned, where you're seeing opportunity, what you've learned about running funds.

SPEAKER_00 21:56

Uh sure. So maybe we'll start with talking about capital partners since that's That's the main fund. That's a generalist fund. Uh it is our is a concentrated public equities fund. For a long time, I didn't want to call it a value fund. And I'll explain why. But what we what Jason and I kind of figured out is these high-level thematics provided a sort of top-down approach to deciding what companies we were going to research. And we followed kind of cash flow-based fundamental analysis to select the companies that we liked. So I didn't really consider it a value approach until I met more people that did different types of value. Value really, I think, at the end of the day, and I wrote a paper on this at the end of our Q4 letter, it's over time, principles get integrated into markets and they get arbitraged away, essentially. So if you if you go back to Graham and Dodd and Securities Analysis and the intelligent, you know most of those things that he teaches in those books, well true, are pretty much non-existence today. Like you don't you don't get net nets unless you go to a foreign country with less good implementation of um property rights and whatnot. So what happened over time? Well, Buffett, obviously his disciple, evolved to uh recognizing the value of float and insurance and then utilizing that, recognizing the value of intangible assets on the balance sheet. Is Coca-Cola investment, for example. I think that as a value investor, you need to recognize that you know what's happened in the past maybe isn't going to give you superior performance over the long run. The challenge is determining what is what is the next thing. Again, these things evolve over very slow arcs, almost like one theme dies uh with you know with the person that comes up with it. And so for us, I think having the technology backgrounds has been helpful because it's enabled us to look at companies that some people might have looked at and said, oh, the PE is too too high, or it's a technology company, I don't trust it because I lived through dot com. Uh and at the same time, I think we can have a rational outlook towards like the nature of software today versus um you know valuations have always also have come down, but you have to recognize the fact that the very disruptive force has entered that market with with the large language model models and artificial intelligence. So um, so yeah, what would you add to that, Jason? I like to fill the Charlie role, so nothing more to add.

SPEAKER_03 24:40

Well, I would say like going back to that point, that's sort of where value does make, you know, meets this new idea of like, well, you have to you have to figure out what's next, but at the same time, if you're wrong, you tend to get taken out to the woodshed, right? Whereas if you find something that isn't popular as reflected in multiples, there's probably less downside if you're wrong, and potentially maybe things will either turn around or mean revert or just it's too cheap.

SPEAKER_05 25:13

Yeah.

SPEAKER_03 25:14

So we're at an interesting time right now because we're in uh we're recording this, it's June of 2026. Obviously, this year at least, and probably for the last year and a half, but it's just become more and more pronounced is the the AI trade, and SpaceX is basically the other, you know, recently it's basically I think it's public for two weeks now. Uh but the AI trade is pretty much where all the money seems to be flowing. Uh, so you have companies at 10x or more in just within within a year, thinking like the memory companies. It's basically everything related to like a story about a bottleneck in AI data centers seems to be flying high. Some of these sort of vertical patterns, when you see a company that's been the same for 10 years or 15 years in a range, and then all of a sudden 10x is an historically cyclical business like memory. Definitely, you know, the question is like how but it's a very much a momentum market. So any thoughts on what we're seeing in the world of uh like large tech companies and momentum and AI on the memory part, I'll play the I'll play the pessimist real quick.

SPEAKER_01 26:22

It's become such a bottleneck that I think the software engineers can solve around it. It's they're trying to maximize the the context window of these AI language models and hold as much data into memory as they possibly can. But the end goal is that the context window is all the data that you have. So you can't load that into memory ever. So I think I I I have to imagine that the the Frontier AI labs are working incredibly hard on how to get around this need for huge amounts of now very expensive memory. So I I I think it's still a cyclical business just due to that.

SPEAKER_00 27:04

Yeah, the the this time it's different, is rarely, rarely right. But how long this memory cycle lasts? I mean, there are some physical constraints, right? Like it will take a certain amount of time to build a new fab. But what do we know about the capital cycle? The memory companies will print more free cash flow in the next year than anyone. Um SK in particular will print more free cash flow than any any individual company that will attract a lot of capital to go after it. And uh I think the thing to understand that's kind of different than like a cement factory, right? Is that there's other ways to solve the problem than just adding more memory.

unknown 27:46

Yeah.

SPEAKER_01 27:47

And they're adding capacity. At the end of the day, this is still just the undersupplied commodity. Whereas I think the AI, like the chips, the processors are highly specialized, and there's that's more on the on the bleeding edge.

SPEAKER_00 28:00

And it's and the real value investment here would have been to invest at the bottom of the memory cycle when they had negative margin margins.

SPEAKER_03 28:06

Which some good value investors did, but they did they sold it after you know they've jumped up 50% or 100% or whatever, yeah. You know, as value investors typically do. Yeah. And it would have been right uh the other nine times, but this time they left 10x on the table or more. Yep.

SPEAKER_00 28:23

Uh now now your question was more about like what else around big tech. You have it you have a couple things that are moving simultaneously. You have open AI and anthropic, who are kind of the two most important, plus I guess you could add uh XAI as a third. And Gemini. Yeah. And Gemini. So between those four, two of them, the largest of the two, or the most uh vocal of the two, Sam Altman and and Dario, have I think done a really poor job with public relations. And that you watch one of the the uh the AI, the Stanford uh commencement address that Soundar gave, and people walked out. We have this fantastic of a technology that is enabling people to have more agency than they've ever had before, and yet it's very unpopular. It tells you that the people that are leading these companies that are doing this are doing a really poor job of communicating to people how this can help them in the future.

SPEAKER_03 29:20

No, I'm not I'm not sure that I I would necessarily go. Well, first of all, like what tech founders have done a great job with being a you know public relations not many good at it. I can't I can't think of any, really. Steve Bomber, I wouldn't maybe, maybe Steve Jobs.

SPEAKER_00 29:38

I don't know.

SPEAKER_03 29:39

I mean, probably I mean, yeah, if you could think of anyone that maybe could be a candidate, it could be Steve Jobs, but even Steve Jobs, it was more like his product was so great and he was an evangelist for his product and a great salesman. But as far as like actually PR or like trying to manage perceptions of people, I don't even know if you think but sales.

SPEAKER_00 30:06

Oh, and that's we're gonna have to go to universal basic income. Oh, and this is such a threat that uh, you know, the way they handled Fable and and Mythos uh really has boxed, I think this has boxed them into a corner and it's put them in a position where the trusted source, which is Amazon's AWS, Microsoft, Google Cloud, are in a much stronger position. So uh, you know, I I think that Anthropic's gonna do great. I think OpenAI is gonna be do great. But who holds the compute? Right now, it's those three companies.

SPEAKER_03 30:41

Well, I I wonder though, right, because you have you have the the energy power layer, you have the physical like data center layer, and you have the model layer, and then you have the applications that can be built on top. And it and then on in the model layer, you have these sort of I guess four closed models, mainly four, and you have a bunch of open source models um that are you know right on their heels, I think, is is fair, but much less expensive, and you get much more control, you can host them and all these things. So is it clear to you that the economics of being in that LLM layer are actually gonna work out great for anthropic and chat GPT?

SPEAKER_00 31:24

Um, software software is and this it is software at the end of the day, right? You can you can fit the weights on a USB key on this thing. So yeah. That said, all of us are using these models, right? We use it heavily in our research, and anthropic models are better than deep seek models. They are more than casuality, you mean but again, the value that I derive is generally so high that I don't want to worry about the difference in the cost. And right, one thing we know from the last couple of years of the stuff being in market is that token costs go down and they go down at a relatively predictable rate over time. So having the best model is great. I think the pressure that you're seeing right now, certain companies in the last couple of weeks have talked about how they're going off of Anthropic and moving workloads over to open source models. That all makes lots of sense. And at some point, Anthropic will probably offer a lower, much less expensive model in order to retain that business within their ecosystem. In the the same way today, they offer haiku and sonnet as as those means, but they'll likely have to get more price competitive on that.

SPEAKER_03 32:35

It's it's it's actually a really interesting case study. I mean, I'm not I I think AI is probably different, but you know, if if Chat GPT is like the Hertz and Anthropic is the ABIS, seems like Avis really kind of got a jump on Hertz, you know, like quite a bit. What do you attribute that to? And then going back to the model, I'd love for you guys to comment on what where do you think there's still opportunity in the layers.

SPEAKER_00 33:00

As far as anthropic versus open AI, I think the pretty clear mistake OpenAI has made is focusing way too much on the consumer product. If you remember from their history, they launched ChatGPT as a demo of like what are the capabilities of this technology. And so many people signed up for it that they were forced to start thinking, oh, this should be a consumer product. And what Dario and the Anthropic team realized earlier than the OpenAI team is that one, software is a language, right? It has rules and structure. In fact, it's an easier language than English. And two, the most likely adoption and who's going to be willing to pay the money for this is going to be businesses because you're replacing something that might take more time with a person or you're making a person many times more productive. Who would want to pay for that more? The individual who's planning a vacation to Hawaii and wants to schedule flights and stuff, or the company that's paying multiple hundreds of thousands of dollars a year for information workers per person that can get three to five X of scale on those people. So I think Anthropic just recognized that first. OpenAI was earlier and more distracted and spread their cards out too thin. But it's a gap that probably can be closed. Absolutely. I mean, I think I would think OpenAI's latest model, and if you use Codex, it's just as good at this point.

SPEAKER_03 34:25

Yeah. My son actually works at OpenAI. I think you guys know that. And uh he there was a period where he he was like, Yeah, cloud's better. And then he's like, this was about a month or month and a half, two months ago, maybe like April, he was like, uh, codex is actually kind of better now. Yeah. For at least for programmers and stuff, you know.

SPEAKER_00 34:46

The trouble is is that people ourselves included, right? We've we've now moved everything over to Cloud. Yeah. And you get comfortable with it. And we now you have switching costs, right? Yes, yes. So so back to like what is a value investor? One of the frameworks that we really like for looking at a particular business is Hamilton Helmer's Seven Powers. And if you haven't read the book, phenomenal book, very easy to read. We always give it to our analyst or uh isn't it similar to I mean, there's some overlap with Porter's Five Forces, though? It's really an evolution of Porter's Five Forces. Yeah. Evolution. Okay. One one way to think of it is uh Michael Porter will give you the layout of the neighborhood, and Hamilton Hell Helmer will inspect the house. Got it. Okay. So you're like, I want to be on this this part of the neighborhood because this is this part of the neighborhood's good. How do I assess if this company has a stronger moat than well?

SPEAKER_03 35:39

Take a look, like let's let's bring in software as an example, SAS. You would look at it with set with the Hamilton framework and say, hey, this is an excellent business, you know, high switching costs, recurring revenue, blah, blah, blah. And then they all go taken to the woodshed, obviously, because this now alien technology drops into Earth. Uh so I know you guys have looked at it. Uh, I've I've looked at it, a lot of other people are looking at it. But what's your take on all of these companies, these great companies, Adobe, uh, you know, Salesforce, very profitable still. You alluded to it earlier, but let's let's let's uh let's let's double click on it.

SPEAKER_00 36:21

There's a lot of threads to pull here. Um, since we mentioned switching costs, let's start there. The main thing that's happening with switching costs is that switching costs are going down very quickly, very, very, very fast. One of the things I did in my first job out of college uh in tech strategy, we would advise on the post-merger integration of telecom companies. And that would be like a billing system, an order management system, all the stuff that you need to run a telecom company. And it was ungodly expensive. It was really hard to do well. Our company, the reason that I got hired, and the reason that it became a very good company and got acquired relatively quickly, is that we took over a project that Accenture was failing to deliver. They brought our company in, we delivered it on time, and it's just one of those things that's it's not easy to do well, and adding more people to the problem doesn't necessarily solve it. What does AI do in this? Well, if you listen to the most recent Snowflake earnings call, for example, a lot of their customers are switching from legacy databases to Snowflake, namely like Teradata, but their ability to transition a company from Teradata into Snowflake is much, much faster than it used to be. And that's because they've built their own harness, it's custom to what the way that their software works, and they can do a much more efficient transition. So switching costs because of AI.

SPEAKER_03 37:50

Yeah, because of AI.

SPEAKER_00 37:51

Yeah. Now the same thing is true on the other side. That means you have more potential competitors at the same time. So barriers to entry are going way down.

SPEAKER_03 38:01

So what which software businesses do you look at and be like, hey, you know, this is this is unfairly punished right now? Any any come to mind?

SPEAKER_01 38:10

Yeah, there's a few. Without getting into specifics yet, I'd say uh uh running a software business is only a third, only a third of the company's engineering. And of that third, you have product designers, you have people doing research on customer research. So maybe half of that team is actually software engineers writing code. And then senior engineers write code 50% of the their hours, their work hours. So the amount of man hours spent writing code at a software company is actually very small. It's really just a sales organization wrapped around a product. So that's one lens we've been looking at lately is is you know, just because you can replace the 10, 15% of the business that is actually writing code doesn't mean you're gonna replace all these software companies. They have there's so much more to it that and we learned this from from being operators in software businesses. Is you can have the product and it can be a great product, but the company can still fail. It's it's a lot more than just having software and being able to iterate through new versions at a rapid pace.

SPEAKER_00 39:25

And anecdotally, like we've we talked to a lot of startups. Most startups are still paying for some sort of SaaS, right? They're still using HubSpot, or they're still, you know, there's some of these kind of core things that they're still paying for. So we're not yet seeing that Salesforce, too.

SPEAKER_01 39:41

Yeah, we talked we talk to startups and we talk to very large software companies, and none of them have said they're they're ripping out SaaS products out of their stack and and rebuilding them themselves. None of them are doing it.

SPEAKER_03 39:52

Yeah. So why is the market not giving them more credit?

SPEAKER_00 39:56

One way that we look at it is do people love or hate this product? And people really hate paying for Salesforce. It's very expensive for what it delivers. And you could get an equivalent or near-equivalent product for one-tenth of the price, not mentioning those before AI, right? So again, switching costs are going down, and yet I wouldn't bet that a lot of companies are going to rip out Salesforce. Salesforce is going to figure out how to make sure they deliver more value to their customers so their customers don't leave. And you're seeing if you look at the job postings for some of these companies, forward or forward-deployed engineers, which was a term coined by Palantir, actually, where they basically embed a software talent. It's basically a engineering sales role. Embed it with the customer, figure out what the workflows are, and make the product work for them. So we have a theory that the horizontal SaaS companies are going to become more vertical, right? Because they're going to go into their largest company customer and make the product work really, really well for them. And then they'll figure out how to wash and repeat that business. But that brings up like another view on this is that how much different is a software company than a consulting firm?

SPEAKER_03 41:14

No, I was going to ask, yeah. So like you on the one hand, you have Constellation, which is a collection of verticals, and you have things, things like Accenture, FTI that are consulting firms, right? Yep. Yep. And then you have things like SP Global and uh Visa and MasterCard and things which have these like networks, and they're all, I think, affected by fears around AI. Yeah.

SPEAKER_01 41:38

So a specific company that I like, and we we don't have a position in it, is Toast. And it got sold through the SaaSpocalypse as well. But they're really a hardware tool in restaurants. So I think that was one that I view as unfairly punished.

SPEAKER_03 41:55

If we fast forward, where where is this gonna go in a year or two or three? What is what can we say about what the world is gonna look like?

SPEAKER_00 42:03

Two things. One, the human mind does not do well with exponentials. But if you were to look at what AI was capable of 12 months ago to now, if you just extrapolated that one more year, we agree. Yeah, forget the exponential curve. The quality of the output is only gonna get better. And what does that really mean? That means intelligence sort of gets commoditized. These things will still like the it go back to Hamilton Helmer's seven powers, these are actually more important now as a result. Your cornered resource, your switching costs, the things that make your customers want to come to you. And at the end of the day, it's value delivered versus value captured.

SPEAKER_03 42:47

Well, I for one am very excited that I'm alive during this AI era because I used to dream about it. And it's it is like an alien technology dropped dropped in just suddenly in like 2023, and now is just exponentially better. Uh, I remember like trying to pull up even like images in Chat GPT. They were pretty terrible like a year ago.

unknown 43:10

Oh, yeah.

SPEAKER_03 43:10

And now they're incredible, only getting better. So yeah, it's really incredible. I've I I'm a big fan of Substack. I know you guys have a Substack and I have one as well called Compound Ideas. And uh I've been writing recently about Musk and stuff, but also like what is what is it potentially, what are the implications potentially of a post-abundance world, right? Because a lot of our assumptions are built around this idea of scarcity, and then capitalism exists to create incentives so scarce resources can be allocated better. But if everything drops to the cost of energy, uh, because energy powers cognitive work and it powers physical work eventually with robotics, then you don't necessarily need an economy that allocates scarce resources. Now, this is not an imminent thing, I think it's probably 30 years out or something, but then again. Again, you know, again, like you said, human human brains can't handle exponential, so I don't know if it might be sooner, you know.

SPEAKER_00 44:05

We looked at prior industrial revolutions. Um actually, Jason, maybe you should talk through this because Jason wrote two very good papers on, and we're probably due for a third one at this point, on the similarities or at least what happened back then. And I think that can provide an outlook for why and what's happening for in the future.

SPEAKER_01 44:25

Yeah. Sure. Yeah, so I started by comparing the first and second industrial revolutions to where we are today. And the first industrial revolution was at the turn of the 1800s, the invention of the steam engine, the locomotive, and that really foundational base of mechanical work. And a lot of what was done for that next hundred years was a person operated a mechanical machine. And it and it really boosted their productivity. But it was like building a base of knowledge on how to build machinery. And then at the at the late 1800s was really the second industrial revolution, and we automated a lot of those mechanics, those the the machine, you know, we we we invented assembly lines and and highly, you know, the the production of the Model T and all that. Um so where we are today is in the 1990s, we had you know this industrial revolution, personal computing, information revolution, information revolution. We we digitized basically all of human knowledge over that 10-year period. And you know, we invented the the mobile computing, it was but it's basically it's humans operating computers. And here we are today at maybe the second information revolution, you know, and and we're having we're automating a lot of that now that we've digitized everything, we've learned from it, we've built the building blocks to have this second industrial revolution that we're starting now, is our feeling. And when the prior industrial revolutions happened, we had a really step change in GDP. And after the second industrial revolution, we're growing GDP at 3% annually, and it it plateaued. So we had a step change, and then there was a new course of growth for humanity. And we feel like a similar thing's likely to happen in the next five, 10 years here is we'll have another step change. I don't know where GDP growth goes, but maybe we grow 5% every year for this foreseeable future.

SPEAKER_03 46:27

So I totally agree with you. I think productivity will go up and economic output will go up. But I guess the question in my mind is more around if human beings are not needed for that productivity or economic output. If companies, for example, don't need to be run by management and organized to take inputs of resources and whatever, because basically DAOs, you know, digital autonomous organizations that can be organized with AI and robots can do a lot of everything you need, then you know, it do we have a system of capitalism the way it is today?

SPEAKER_00 47:06

I I don't know. I don't think any of these ideas are necessarily new. And the you know the concept of the loom was very scary to people that were textile workers back in the day. Right. That is if I remember right, that's where the concept of the lemmings or the um the Luddites came from. So much as it's easy to stoke fears about these things, there's nothing about the past that tells us that this will be anything but net positive for people and standards of living.

SPEAKER_03 47:39

Yeah, certainly overall, the the course of technology has been positive for material wealth. Uh higher productivity can be distributed in whatever fashion.

SPEAKER_00 47:48

But even if it's unfair unfairly distributed, the standard of living at the lowest end of society gets massive. I mean, the bottom 1% in the United States versus the top 1% 150 years ago. I mean, what yeah, which would you choose?

SPEAKER_03 48:08

No, I I've I'm absolutely on board with that and I've written it. Although the as a counterpoint, I think people tend to measure status and happiness relatively, not in absolute terms compared to somebody that was alive 100 years ago. Uh, but that that is absolutely, I think, a correct point economically. Yeah. Uh but the difference now is with technology, like for example, you when social media came out, I did not I would not have intuitively guessed that it would create a lot of mental health issues for a whole generation of young people. That was not intuitive to me. However, that having been said, we kept it away from our children. You know, and this is my kids were born like in the early 2000s. So I think we were relatively early. So we didn't really see that it would serve them. But but you know, but a lot, so that's not intuitive. So it can also, it can be true, like for example, it's increased communication and a lot, it has a lot, a lot of benefits. So it can it can be simultaneously true that the world is much richer and also that there's greater inequality or some other mental health or other problems that are caused by not having meaning or not having purpose.

SPEAKER_01 49:15

There's precedence for that. So during the prior industrial revolutions, if you gauged and we read um you know books on the topic and stories from that era, and you gauge the population's sentiment, and it was very negative. And they thought the economy was crumbling, and it's kind of parallels a lot what you hear today, that the economy is not good, but then you look at the numbers behind it and it's it was accelerating.

SPEAKER_05 49:39

Yeah.

SPEAKER_01 49:39

So I think we're likely to see the case same kind of situation.

SPEAKER_03 49:43

And and you know, it took a while to sort of flush out, but yeah, it reminds me of that quip that a recession is when your neighbor loses your job, and a depression is when you lose your job.

SPEAKER_00 49:54

Yeah. You know, uh another way to look at this is technology evolves at a much quicker pace than biology. And we know this as investors because the markets are often very irrational. And it's our lizard brain, as we refer to it, as it's your lizard brain that holds you back from making good, rational decisions. But yeah, even the best of the best make those mistakes. Yeah, absolutely. When you throw technology, whether it's social media or whatever, at a child who's a developing reptilian brain, don't be surprised if they end up having consequences, positive or negative, that were not anticipated.

SPEAKER_03 50:36

Yeah, that absolutely makes sense. Um, I think that's a good segue to I want to talk about healthcare and biotech. Uh I think one of the super, at least, I mean, I'm not a healthcare biotech investor, and I think in that respect, you guys are relatively in the minority as far as I know, people, is that I don't think a ton of value investors tend to feel comfortable with healthcare and biotech. But I I think especially given that uh AlphaFold just sequenced all the proteins and how basically I think we're on the precipice of just incredible advances and a pace of advance through due to AI and cognitive processing. Where what do you guys think of where the opportunities are? Now, is it still the case that biotech is is sort of unpopular? Because I I guess I just read an article that the index has come back a long way.

SPEAKER_01 51:26

It has come back a long way, but it spent years being very unpopular. Okay. At least I think the last three years have funding's really dried up. So there was a huge bout of through the pandemic years, and then that all kind of dried up. And and MA is really the engine of the biotech industry. If you start a biotech, you're hoping to advance your drug to phase two, phase three clinical study, and sell to a big pharma company. So if that sale's not there, the scientists frankly don't get rich and decide they're going to start another company. The VCs don't see the return that they can then roll into the new crop of companies. So the funding all froze for several years, and we're just now seeing that thaw. Because those big companies stopped buying the minnows? Yeah, acquisition slowed down, funding slowed down. A lot of money got out of I think there's a couple reasons, but you know, one of them is AI. A lot of the venture money shifted to AI. It wasn't popular to place money with life science, healthcare, biotech VCs.

SPEAKER_00 52:36

Well, should we talk about China and licensing? Because another thing that's happening, and this is actually more recent, is what, two years ago, almost nothing was in licensed from China. But today it's in the 40-something percent range.

SPEAKER_03 52:50

Meaning that 40% of all the things that like US or global companies are in licensing are coming from China. Yeah. Chinese research.

SPEAKER_00 52:59

And what what they've done is they've really commoditized commoditized the chemistry. And what does that mean? Is they've taken patented molecules and modified them slightly and created a market for it. And that's been very good for the large pharma companies like Lilly. Not necessarily.

SPEAKER_03 53:16

So who are the who do you guys think are the beneficiaries and who gets hurt by the coming, or maybe it's here, revolution in AI and biotech and healthcare? Loses.

SPEAKER_01 53:26

I think there's a lot of winners. Yeah. I think I really like the space in general right now for several reasons. One, you brought up Salpha Fold, we're gonna be able to design new therapies a lot faster. These Chinese biotechs, like we need to figure out the protecting our IP challenge. But the way they got started is providing low-cost, early stage research to US biotechs. And and that's a real benefit if you want to cheaply start you know pursuing an idea. So yeah, there's a challenge there with China, but we'll figure it out, and maybe we can do things.

SPEAKER_00 54:07

It's not all bad. I mean, like having a low-cost, high volume chemistry basis of expertise is great for everybody.

SPEAKER_03 54:17

Yeah. But when you say chemistry, I don't know. I think of like nitrogen and you know, oxygen and things. But you're you're talking about a pharmaceutical chemistry?

SPEAKER_01 54:28

What do you mean biologics?

SPEAKER_00 54:30

Both. Uh yeah. Yeah.

SPEAKER_01 54:31

So I mean a lot of it is combining chemicals to make uh a chemical compound that you're trying to make. You need to figure out how to make it at scale.

SPEAKER_03 54:42

So are you saying that China will just like kind of be to chemicals the way TSMC is to chips? You just outsource it to them and they make they make it for you?

SPEAKER_00 54:52

I mean, that's obviously happening in some sense now. I mean it's okay. I don't think it's gonna be everything. But remember, this revolution happened in the 80s, right? Like Vertex Pharmaceuticals and Regeneron were probably the first of the early we're gonna design the molecule. And that was very hard to do back then. I mean, they had they had complicated computers that would like try to help them do this stuff, but it was definitely more a dark art. And it's evolved today where you know, not only can everybody do it, but China's figured out that they can do it at scale and scale the operation.

SPEAKER_03 55:31

I see. And so where where do you guys see the opportunity in the next say three, five, 10 years on the healthcare biotech side? You know, I'm talking big pharma companies, biotech innovators, you know, there's venture like in the whole ecosystem.

SPEAKER_01 55:46

Yeah. Well, I'll tell you one way the whole stack's going to benefit is a lot of the recent changes the FDA is making along with federal regulation, and they're trying to cut out a lot of the downtime and wasted time in running a bringing a drug to market. So it's a right now, it's a 10-year process. And once it's in clinical stage, so you have to file to have your new drug applicate, investigational new drug, then you run your studies. When you conclude your study, do a data analysis, which can now be aided by AI, then that package gets sent to the FDA. They sit on it for a year, make a decision at the end of the year, and then tell you if you can go forward or not. So there's this back and forth that happens several cycles. And they're trying to cut down on how long, you know, the data analysis process to submit it to them, how long the FDA takes to look at it. Can they use tools to speed that process up? Can they streamline the number of clinical studies you have to run? So we're compressing this whole timeline and we're compressing the preclinical before it gets to humans, new regulations around how they're going to govern animal studies. So once you develop your drug, you have to patent it and that starts a clock. And then that clock determines how valuable the drug is after it reaches the market. So the average drug once it once it's uh commercialized is on patent for seven years. And if we can compress the dozen years that it takes up front, and if your seven years that it's on patent can go to ten, even just adding three extra years probably doubles the value of the drug. Because the the first part of commercializing it is you're going to get insurance coverage for it, you're building your sales force, you're educating physicians. So by the time you're at year seven, that's your peak sales. So if you can extend your peak sales a couple more years, you've doubled the value of the drug. So I think we're gonna see, you know, we're at the early stages of this, but we're gonna see the value of all these drugs dramatically increase.

SPEAKER_00 57:50

Now there's there's another dynamic to this, though, is the US healthcare system is on a one-way track towards insolvency. And so just like in technology, you you need to be very careful about what you choose to buy. The same thing goes in the life sciences, because if your value created for your product is less than or not significantly more than what you charge, then don't be surprised if say we so say we go to a Medicare for All. Don't be surprised if the US starts negotiating very heavily, as this administration has already started to do, in order to reduce those costs. So currently, you know, what is the the market today is could be very different. But if you focus on companies that deliver a significant amount of value given the cost, um, I think you can choose some very good companies. Yeah.

SPEAKER_01 58:47

Now a headwind until all this happens is these big pharma companies have a lot of patent cliffs coming up between now and 2030. So they're this is why the MA is starting to kick back on right now, is they have to buy drugs to fill their pipeline, their their sales pipeline.

SPEAKER_00 59:05

I think this year there's already been more acquisitions this year than all of last year. Right. Right.

SPEAKER_01 59:10

I think we're on track for one of the largest years of MA.

SPEAKER_00 59:14

That was your prediction, right?

SPEAKER_03 59:15

This is the largest MA year on on record. So so we have another five or ten minutes left, and I want to touch on at least two other things. So one uh one of them is uh how will we as long-term investors be able to harness AI to make better investing decisions? I think you guys are doing some innovative things. I know Brian Lawrence is as well. So maybe touch on that, and then I want to uh talk about Capital Alliance a little bit too.

SPEAKER_00 59:40

So um I think investing similar to software creates a lot of leverage over one particular individual. Right? You can you can get a lot out of a little because money scales, and so does software. So, with that, for us, we see our ability to continue growing this fund for a very long time without necessarily adding heads, which will be good for us. And the long term, it'll be good for our investors because our we'll be able to take down our management fee, and that all will enable us to run a much more efficient operation. But efficiency aside, how do we make better investment decisions? Break up the tasks that one has to do as an investor. Many of them now you can just automate some portion of it. Now, if you're brand new and don't know what to look for when you buy mistakes or in order to get the type of result that you're looking for, I think that that requires a level of experience. But I think we're in a pretty unique position because of our technical backgrounds, uh, in order to build the systems to support those processes.

SPEAKER_01 1:00:44

It feels very value add, but you brought up Brian Lawrence, and you know, we he we talk about the new things we're doing with AI, and so does he. But that's at the end of those calls, he likes to say, but has it made you any money? And we can't, you know, put our finger on it yet.

SPEAKER_03 1:01:02

Yeah. I mean, it's yeah, it's really it's interesting because uh I saw this chart of like how already the companies that are using AI are outperforming the ones that aren't. And obviously, in like small business sector, there's probably gonna be a lot more that aren't than are, just by the nature of small businesses. Um so I think I think it's interesting. Like, if at least based on your experience, and you guys are probably more advanced than others, what what are some ways that a fund manager can really start to leverage AI? You know, in a practical sense, like, you know, should you get clawed code and have it do XYZ, or should you be coding custom applications, or should you be using a service that's out there that you know you just maybe don't know about?

SPEAKER_00 1:01:48

Eventually there will be services, you know, there'll be the Bloomberg of AI or something like that. But for now, I think of it a couple different stacks, right? If level one is just chatting with Chat GPT, right, or Claude, or choose your favorite chatbot, you can get a lot out of flushing your ideas out with a chatbot. Chatbots tend to be incredibly sycophantic. So you need to take steps, update your system prompt with instruct it to be critical, instruct it not to compliment you, instruct it to challenge you. That would be level one. I think level two is when you can how do you pull all of the information that you need into one spot? And I think now pretty much all the chat the chat applications have like projects you can dump files into, whether you're using Notebook LLM or you're using Chat GPT projects, these all enable you to basically say, I want to take all this stuff and I want to work on it with the help of an LLM. I think that can be really helpful. Once you move into Claude Code or ChatGPT Codex or one of the others, you can move into defining skills. So like these are things that you will do over and over again. For example, many of us use a lot of checklists, right? When you're looking at a company so that you don't miss something, you build a checklist over time, and maybe you add to that checklist over time. Those are perfect templates to move into skills and have the AI do the first pass analysis, if you will, on it. Going beyond that, we will in kind of where we are is we have a central repository of data that Jason and I and our AI agents share and feed into. So it's a combination of new data comes into it every day. Um and we contribute to that every day. And then we've built workflows that run autonomously in loops around that data that kind of spans the gamut from idea generation, knowing what the AI now knows about us, what is the most what is a really good company that we should look at. So we're using it from the perspective of can it can it make us more efficient, make us make better decisions because we're gonna look at more companies and we're gonna look at better companies than we otherwise would have. Another thing we we have is like a we have an internal podcast that's just for Jason and I. And every single day that shows up on our podcast player, and what does it do? It basically looks at our database and says, this is all the new stuff, and this is how it relates to your existing positions, this is how it relates to your watch list. So instead of spending hours doing essentially that task every morning, we're getting at least some chunk of it done in a podcast on the drive to work.

SPEAKER_03 1:04:47

Just so people can visualize the effort it took, is this is this you have your own server, you use cloud code to create these workflows and applications? Is that generally how it works?

SPEAKER_00 1:04:58

Yeah, cloud code has been the workhorse. We we've we've experimented with open claw. In fact, our early earliest trials were with open claw. I think for our purposes, we're a little more technical, so we could handle Cloud Code a little better. I and I should say, like Jason and I have a good delineation in this because you've heard the concept of vibe vibe coding. Vibecoding is great for somebody with lots of ideas and uh no technical knowledge. Since my coding background is so old, I can I can vibe something up and then show it to Jason, and Jason can say all the reasons why it's bad, or at least make it more robust.

SPEAKER_01 1:05:44

From my perspective, I delay doing things because I know someone else is doing it and they're gonna open source it or release it, you know, in the next month.

SPEAKER_03 1:05:53

In terms of creating new applications or whatever, just like wait for someone to release it. Yeah. Just as a practical matter, um, I think. Mentioned it to me, Mike. I just wanted to get another reminder of it that like if you want your AI to be able to access, you know, market data or fundamental data or whatever, I think you mentioned the way you guys do it that you thought was pretty good.

SPEAKER_00 1:06:13

Yeah, there's there's a bunch of different APIs out there. Um well name particular ones because just Google financial data API. But you can also just create a skill. And we have this that will go to the SEC website, okay, get the document and manually scrape it.

SPEAKER_05 1:06:31

Oh, okay.

SPEAKER_00 1:06:32

Which is, I think it's regardless, it's still important. Because if you were say your old workflow was on Bloomberg, right? At least for the type of investing we do, where we aim to do a whole bunch of work up front and own the company for a very long time, you may do all that research on Bloomberg with consolidated financials and and and a pre-populated model, but before you pull the trigger, you're going to go back to the source documents. So, yes, plugging into an API makes that first part easier, but it should never eliminate the the side of this is really coming to understand the company and being able to be critical of your own, your own work or your AI's work.

SPEAKER_01 1:07:12

Well, and we purposefully pulled out some automation too in the past where we force ourselves to read the 10Ks and the 10 Q's. And we don't want to outsource that. Yeah. Last few minutes.

SPEAKER_03 1:07:24

Just wanted to so Capital Alliance obviously is a community of fund managers who get together multiple times a year. You guys are part of that right now. And uh the reason we could create the community is like we we think that relationships compound and in-person multiple times a year is is how relationships can compound the best. Um, plus, in a world where we're all connected by digital technologies, true in-person community is sort of harder to find. Um, and just wanted to get your thoughts on sort of um the capital lines community.

SPEAKER_00 1:07:55

Yeah. I so think about like the type of people that start investment funds. It's generally, at least for us, is you have this level of curiosity and some sort of drive to find the truth. That skill set tends to favor more introverted people. Tends to, not always. It tends to favor more analytical people, it tends to have fewer super outgoing salespeople. And yet, for most funds, running a fund involves a good chunk of the latter. So I think for us, we have a lot of benefit that we get to work together every single day. But there's a lot of single fund managers that are solo by themselves all day long. Either way, we've gotten a lot of benefit out of talking to other people that are doing similar things to us and sharing what we've learned. I think, Ken, you were you're 100% right that somehow people are becoming less connected. So providing a venue to enable us to get together has been really powerful. And I think it's I think it's something that a lot of people should consider.

SPEAKER_01 1:09:01

Yeah. The the members are from all over the US and a few in Europe. And yeah, if you're we live in San Diego and there's very few people here that we know do anything similar to us. So to have a community where you get together a few times a year and everyone's doing roughly this the same type of investing is yeah, it's it's it's hard to build that otherwise in your local community if you're outside of the New York area. Yeah.

SPEAKER_00 1:09:29

I'd also say it's just fun. Like I in the early part of my career, you'd I have to go to these conferences, and I hated going to these conferences. But like this is really fun in the same way that kind of the value X stuff is just really, really fun because everybody has this shared common ground.

SPEAKER_01 1:09:44

It's really friends of yours, right? So you're looking forward to showing up to a conference because it's all your buddies. Exactly.

SPEAKER_03 1:09:50

Yeah, that's that's sort of the idea, is that um obviously there's like the formal programming and there's the ideas we share and the help we ask for and all those things. But I'd say like half of the value is the is the dinners and the lunches and the informal conversations and the unexpected, serendipitous things that cut that that sort of start happening. And then, you know, certainly because I have like experience with another group that inspired Capital Lions just in a different domain. Um and now I've been in that group for eight years, and it's if I look back, I I couldn't have imagined all the good things that came from these relationships. I wouldn't have even known that I would have formed those relationships. It was just a matter of showing up, engaging, being open. And like because it's in a domain that I I don't really spend most of my time in, that you know, whereas I do spend most of my time in investing in fund management. That that's kind of what inspired inspired the creation of Couple Lines. So yes, I think we're at the end of the the broadcast. Any uh any last words? And if also if people were intrigued or want to reach out or follow up. Uh, I know you guys are uh you have room for new investors as well. Uh so how does some somebody get get in touch?

SPEAKER_00 1:11:03

I guess the first first step would be subscribe to the podcast. It's called Tell Tales. We're on every single Wednesday. And then we also publish our our partnership letters every quarter. And you can get those by subscribing on the website. They tend to go out quite a bit later than they go to partners, but we we just sent out the Q1 letter. And if you go sign up on the website, we'll send that to you right away. And uh and yeah, to start a conversation. I I think our goal with with new investors is to find people that are understand what we do and are are well aligned because this is a long-term game. So we like to play these long-term games with long-term people. And so as a result, as Ken knows, we don't really try to do sales. It's more about developing relationships with people who think similarly long term.

SPEAKER_03 1:11:53

Yeah, couldn't have couldn't have said it better.

SPEAKER_01 1:11:56

Do you want to add anything, Charlie Munger? No, I think I think Mike Mike said it. It's just reach out to us. We're available. We we enjoy talking to RLPs and and anyone interested about these topics, like like we just had the conversation with you. Great.

SPEAKER_03 1:12:11

Well, I've I've always been super impressed with how thoughtful you guys are. Love telltales, try to catch it whenever I can. And uh look forward to uh you know seeing where things go from here. Cool.

SPEAKER_02 1:12:27

Appreciate it, Ken. Yeah, thank you. Thanks for watching this episode. Click the screen to catch our past episodes with Ass of Saria and Matthew Peterson. And to apply for the Capital Alliance Mastermind, hit the first link in the description.