Yet Another Value Podcast - Fintool's Nicolas Bustamante on using AI to improve in investing

Episode Date: May 29, 2025

In this episode of Yet Another Value Podcast, host Andrew Walker speaks with Nic Bustamante, founder of FinTool, an AI-powered platform designed for equity analysts and investors. They probe how AI is transforming investment workflows, from memo creation to screening and qualitative analysis. Nic shares examples of institutional adoption, discusses nuanced challenges like bias in management conversations, and forecasts how AI could evolve investor roles. Whether you're deeply entrenched in AI or just starting out, this episode provides grounded insights into its growing role in finance.______________________________________________________________________  [00:00:00] Andrew introduces podcast and guest[00:01:50] Nic explains AI task delegation[00:04:10] Home Depot memo AI example[00:06:59] Uploading memos to train AI[00:08:13] Pattern matching with past investments[00:10:39] Small sample size challenges[00:13:10] Buffett’s approach vs. LLM potential[00:16:08] Investing skill shifts with AI[00:18:00] Qualitative work amplified by AI[00:21:19] Gumshoe research vs. AI insights[00:23:21] Amplifying analyst strengths with AI[00:25:59] AI freeing up research time[00:27:37] Future of autonomous investment agents[00:30:10] Training AI with personal track record[00:31:59] Data diversity needed for edge[00:33:38] Qualitative investing with AI portfolios[00:36:02] AI advantages in news trading[00:37:36] Losing insight through automation[00:39:21] Hybrid strategy using AI summaries[00:41:40] Identifying non-standard compensation[00:42:53] Spotting off-cycle stock grants[00:45:36] Edge cases needing human oversight[00:47:48] Tesla and extreme market narratives[00:49:22] Fragility of company valuations[00:51:16] Reliability of company filings[00:53:31] Expanding Fintool’s data sources[00:54:11] When and why to upload documents[00:56:25] Private data and unique uploads[00:58:14] Bias risk from selective inputs[00:59:38] Recording calls for richer context[01:00:23] Generating insightful questions with AI[01:01:35] Framing management conversations for AI[01:02:49] Extracting insight through competitor focus[01:03:46] Using peers to understand companies[01:04:43] Keeping pace with fast AI evolution[01:07:02] AI as necessary but not sufficient Links:Yet Another ValueBlog: https://www.yetanothervalueblog.com See our legaldisclaimer here: https://www.yetanothervalueblog.com/p/legal-and-disclaimerFinTool:https://fintool.com/

You're about to listen to the yet another value podcast with your host, me, Andrew Walker. Today's podcast is a really interesting discussion with the founder of FinTool Nick. Nick has a lot of insights into AI investing, and look, I know what you're thinking. You're saying, oh, hey, FinTools is a sponsor. This is a plug podcast. I don't think it is. I ask him a lot of different questions about how individual investors, fundamental investors, can incorporate AI and improve their AI workflow in a lot of different ways, right off the top,

like right at the start of the episode, I'm going to hit him, hey, what is the one thing that a fundamental investor can do to improve their usage of AI right now? I ask lots of different cases, use cases, downsides, everything. I think you're really going to enjoy it if you listen all the way through. So without hesitation, here we go, my interview with Nick from FinTool. All right, hello, and welcome to the Yet Another Value podcast. I'm your host, Andrew Walker. With me today, I'm happy to have on FinTools founder, Nick Wustam up.

Nick, how's it going? We did. Thank you for having me on. Thanks for coming on. Just remind everyone, there's a full disclaimer at the end of the episode. Nothing on here's financial advice. You can listen to the full disclaimer at the end. Nick, I want to get started.

So I'm super excited to talk to you. I have been talking to FinTool for about a year now. You guys suggested coming on. I am always looking for ways to improve using AI as an investor. And obviously you guys run an AI tool for investors. so I thought this could be a really good conversation. Let me start off. First question, just if our listeners stop listening after three minutes,

let's give them something to take away. If you were having, you know, the average investor who comes on and listens to this and they said, hey, what is one thing that, you know, Nick sees thousands of investors use AI? What is one thing that Nick thinks every investor could improve their AI usage? What would that be? Yeah, so we worked with a lot of hedge fund. We work with, you know, large banks. work with big consulting firm like, you know, PWC, and our product is sort of like AI equity research

analysts. What I've seen from, you know, successful hedge fund and value investor using AI is they have a clear breakdown of their workflows. So they say from ID generation to making an investment or even not making the investment, they broke down their workflow into a very specific task. Usually they did that to train their analysts, right? But now they do that and say, okay, among my, you know, 60 tasks, which one can I delegate to AI? And some of them are pretty small. And you have, yeah, I was chatting with a fund this morning. They do quarterly memo. So Home Depot, you know, new earnings. And so they will ask the analyst, okay, read the earning code, read the 10Q, the press release, the AK, and you will draft a note. You will extract

the key numbers, et cetera, et cetera. Well, it turns out if you feed Fintour with an example of the memo, so let's say Q3 and you ask, do the same for Q4, Fintour will do it automatically. So they will delegate this task to AI. And by the way, they will do Home Depot, but they will do those and they will do like more, you know, companies. So I would say it depends on your task, you're going to choose them and dedicate some of them to AI.

And I can give you, you know, more. concrete example. Screening, I think, is a good one where today you do quantitative screening. You say, hey, I'm looking at companies below $10 billion in market cap, healthcare, P ratio under 30, RIC, blah, blah, blah. But then you can say to the AI, okay, look for this. And also, I want companies that are from the lead. I want the CEO to, the CEO is mentioning like buybacks, you know,

future buybacks. I want you to check if they have the cash to do a buyback. I want you to check if they did successful buybacks in the past. Did they buy anything below the interesting value in the past? And if yes, you know, you combine, you rate the company, and you show me the opportunity. And so, yeah, it's per task and per workflow. No, so for screening, I found it really interesting exactly what you're saying, where you say, hey, find me companies that are trading below 10 times price to earnings and the COS.

I think we're trading below intrinsic value in the past three quarters, right? And that's really interesting because the 10th century turns, you could have done that on Bloomberg or anything for 15 years, but marrying it to something specific in the transcript. And obviously, this is a very generic one, but those are really interesting. But I want to actually, the memo thing really jumps out to me. So I want to ping a little bit on that, right? So you used Home Depot as your example and said, hey, Andrew is my analyst. We have a position in Home Depot.

Every quarter, Andrew kind of writes, you know, most people. are familiar with this. You do the one-page overview, right? If you're a very institutional investor, it's going to be Home Depot reported $1.50 per share in earnings. The consensus was $1.35. We thought it'd be $1.40. Here were the same source sales trends. Here's what we're seeing. Here's that. It's going to be a one-page summary of everything you need to do. I'm interested. When you talk to people who kind of start outsourcing that to AI, right? How heavily are they outsourcing it? Are they doing like, look, Home Depot's got 100 different retail competitors. Are they saying, I have a position at Home Depot, so I make the Home Depot memo, and then I have

AI make the 99 companies that I follow so that I don't have kind of the hallucination memo. Are they going so far as saying, hey, everything is getting made by AI, like my, you know, I don't get rewarded for regurgitating what's in the quarter. I kind of get rewarded for the, I'm really interested in that. I think here, sell side is a good example. You know, your analyst at a bank, you cover like 50 names, and you have to send a report to your customer. in a few minutes to like an hour about a new earning release. And it's very standard. There's like no customization.

Maybe the customization is, oh, you know, the estimate with this. And so you feed the documents, the initial memos, you think tool and it will duplicate. You will review the output saying, okay, this is good. Maybe you're going to add a line. Usually what they do is they add like either in the summary or the conclusion saying, hey, we're still bullish and we maintain all price target to blah, blah, blah. I'll get to blah, blah, blah, and then they publish.

So that's a very simple. For hedge fund, what I've seen in the memo is less standardized. Every PM, every analyst, they have a custom memo. That's also why AI is popular is because there is no such thing as like one memo for every company. You might have like one memo per company or per industry. And also, depending on your thesis, you will look for a different thing, right? So sometimes they look for, you know, margin pressure.

mention of inflation. And so that's very useful when they give us, you know, an example of this memo that they upload and the AI can duplicate it. I like that you said an example because one thing I've tried historically and I haven't had success with. And I'm interested if you've had people. Like I do a lot of writing, obviously. And I have tried uploading like, hey, here was my thesis. Here was a successful thesis I had on a company that played out. find me a similar company, find me a company with a similar thesis, I haven't had much luck in terms

of that like pattern recognition mirroring. And now maybe I'm doing two porky examples, but have you had people kind of had success with saying like, you know, if you're a mutual fund, or even hedge fund, let's say you're concentrated hedge fund. You might have made 50 investments over the past 10 years if you're super concentrated. Maybe it's 500 investments if you're semi-concentrated. Have you had people say, hey, here's my top 50 ideas, like generate me five ideas to research that kind of look like this. Or another way I've thought about it is, hey, here's my track record. Here's my track record. Here's my current portfolio. Tell me which things in my current portfolio look like my best investments and which things share qualities with my worst investments. I haven't had much luck,

but just because dumb, dumb, me drooling on my mouth has and doesn't mean no one has. So have you seen people a lot in there, but have you seen people have any success with that? Look, it's because it's super complicated. Even us, we do that only with super enterprise customer. We ask them, hey, give us your portfolio. Obviously, if it is like some hedge fund functions here that we have in San Francisco, they are HMI capital or whatever, those guys are like super concentrated. And you can look at the 13F for like firms like TCW, right?

They have like 200 billion under management. So they give us their portfolio. They give us their coverage, you know, universe of companies. and then what one from them is also additional information. And that's why now and every solution is similar, right? We ask them to connect their data. So we go to the SharePoint and we download the data. And in this data, they have like memos of like, hey, here's this on this investment.

Here's this why we invest in that investment. And then we do the work of looking at the dates, looking at the stock price, and trying to gather as much information as possible. to identify what's a winning investment for them and then trying to duplicate on the whole stock market. But it's not something that you can do with, you know, consumer, long-language model. It's very intensive and very error-prone.

So just like just super hard. So this actually, I had a question on this later in the conference for later in my notes, but I'll just ask now. I was talking to a friend the other day and talking about using AI. And he kind of hit on something. And both of us actually were at the same time. You're like, hey, if you're Renaissance, right? If you're Renaissance and you're making 100,000 trades a day, AI can be super useful for, like, course correcting you and studying your history, right?

Like, you've got so many data points. Whereas my friend was like, hey, my average turnover, let's just make the numbers easy. I hold eight stocks and I hold them on average two years. That means on average, I'm making one investment per quarter, right? And he's like, I can't now, he's not saying AI is useful, but in that way, he's saying I, don't generate enough data for AI to really be able to go back and say, hey, you know, like everything you're doing is so esoteric or so niche or so at the edge. He was saying, AI just can't help me in that way.

It doesn't mean he can't help in a lot of other ways, but he was saying, like, I want AI to make me a much, much better investor. And he was like, hey, it can be your super Google, but it can't, like, really improve you like it could a quantity fund. What would you say to that? Yeah, very familiar with the Renaissanceius case. You may have noticed, but French guys like us, like me, we tend to work either in AI or in mass-oriented stuff.

So look, when Jim Simons showed up and he said, I'm going to apply machine learning and statistics to the stock market, I'm going to predict prices, and I'm going to identify price discrepancies with machine, and, you know, I'm going to trade on that. And people are like, you know, the traders that were in those floors at the bank, they were like, no way. I rely on my intuition. I blah, blah, blah. Obviously, it didn't work for gym for maybe five to ten years. And then, you know, now Renaissance is, you know, 90 billion compounding at 30 percent a year. So ultimately, it worked. I think you might be low on the compound in there, by the way. I wish. I wish I could have invested. But so look, it works for this quantitative

data side. And now we have the quantitative data side, which is more like our customers, value, long-term-oriented. And obviously, those guys, they say, hey, I rely on my intuition a lot. It's an art and not a science. But large-language model, now they have an understanding of text, data, nuances. And so ultimately, it will work if you give the long-language model enough data. So you start with the SEC filing, the earning course presentation, the expert call network, et cetera, et cetera. you input your criteria and at some point the AI will run offline and, you know, find investment

opportunities. And that might be different that the investment opportunities you add in the past, but it will be the same, right? If you do small cap, it will be small cap oriented. If you do profitable company, it will be only profitable companies. So I'm 100% sure it will happen. I understand that people have a lot of skepticism, like back in the days when, you know, the math guy wanted to apply ML to, you know, the stock market.

And I can tell you, as a software engineer, I lost my job to AI. And you have to understand that two years ago, AI was writing maybe five to 10% of one code and it was buggy and we were pissed off at the AI. And then you went to 30% and then 50%. And today, it's like 100%. AI is the best software engineer on the planet. And it's inevitable for, you know, every profession. Let me ask you the next question then.

I've written before and I've said before, I worry. Look, 50 years ago, you could have made really good returns if you, I'm rereading the Snowball right now. And a lot of what Buffett did was literally go through Muti's manual and say, hey, this company trades for 10 and they earn $5 per share every year and they have $20 per share in cash. What a deal. And you bought it. Now, obviously the man did a lot more than that.

But literally, those are the stories in the book, right? You could have made tons of money doing that 50 years ago. Renaissance comes along, and Renaissance obviously does this on steroids with a lot of other stuff. But that quantitative stuff is dead, right? For the past 20 years, you could make money doing qualitative stuff, things that weren't in the numbers. If your thesis was this is trading for eight times price to earnings, I think I'm going to make alpha, you were dead. If your thesis was, hey, this trades for 30 times price of earnings, but the earning number is actually eight times once I make adjustments or because you could make a lot of money. I worry, quantity is dead.

Renaissance skilled AQ or whoever. I worry that the quality of AI is going to kill it. Can you, like, are professional investors going to have a role in the world in five, seven, ten years? Look, you know, the reason I started Fintow was pretty much at the Berkshire meeting because, you know, I work in AI for a decade. I went to the Buxhire meeting one time right after selling my previous company. My purchase company was a legal AI company, a sort of, you know, alpha sense for the legal industry. And again, I hear Buffett saying, yeah, if I do small cap, I can do like 50% a year, blah, blah, blah, just scanning through opportunity.

This year, I show up to the meeting, and Buffett is like, well, I had this booklet of like Japanese companies, and I read it all, and I look at what makes sense from a business perspective and the management. Obviously, very quantitative assessment that is impossible to do before the long-language model. And now the AI is going to do it. And I think it will still be possible to generate alpha. It will be sort of like obviously harder. Like the market will be more efficient. And a lot of it will be done by AI.

The same way, us as software engineer, what we do today is like we are the architect. We are like the meta-thinker. but the AI is writing 100% of the code. It's almost like when you interview someone, it doesn't even matter if the guy knows how to code because the AI is doing it all. And one year ago, I was with my friend in San Francisco, the guy went to MIT and then he went to Stanford.

And it was like, no way AI is going to write my code. And it was like, obviously, right, it's way too hard, like similar to some of our customers today. And then it happened. And then it's like, okay, what should we do now? And basically you are this orchestrator of the AI systems. This actually dives nicely into a thought I've had in my mind. Do you watch, I think you and I've talked basketball before.

You follow basketball a little bit? Yeah, yeah, especially in San Francisco, we have, you know, great team. So one thing I've had in my mind is Steph Curry, you know, top 15 player of all time, multi-time MVP, all that sort of stuff. if you took step curry and rewound time 50 years and put him in the NBA i don't even think he would be an NBA player right his greatest skill is a shooting there's no three point line so he can't stretch defense like that and also step curry has a lot of problems with his ankles and if you put him 50 years ago when he's playing in flat top converses with the medical science then i don't know if

his ankles can even hold up so again one of the 20 best players of all time he's not even playable 50 years ago and in today's NBA he's probably the most valuable player i mean it's him Ron and two other players are the most valuable players of the past 15 years, right? In contrast, I think of like the legacy power forward who wasn't big enough to be a center but was, you know, 610, couldn't shoot but could rebound. Like, not Tim Duncan's one of the best players all the time. He could had a mid-rate, but you know, kind of like the Twin Tower style, a lot of these plotting centers, plotting power forwards.

30 years ago, they were super valuable played out of the NBA today. Can't even get a spot. Roy Hibbert, 2010, super valuable verticality, couldn't even play in today's NBA. The reason I ask is this. You framed something where, I think we can talk about the lawyers, but 30 years ago when portfolio manager, quantitative skills might have been the most important things for portfolio manager. AI, in the same way that medicine and sports has evolved, AI is going to make some skills a lot

less valuable and some skills a lot more valuable. What type of skills do you think AI is going to be like kind of a leverage point where if you were an analyst or portfolio manager today, you might be saying, hey, with where the puck's going in five years, these are the skills I really need to be focusing on. Yeah, that's a very good question. And I will say today, the only two jobs where you can really see the impact of AI and AI's artificial general intelligence, like this super smart system is, in my opinion, if you're a software engineer, you see it, it impacts your job. It's very hard to get the job with the junior guy, et cetera.

And if you're a lawyer, and I saw that from my previous company, because lawyers, it's just words. You ask the AI to do NDA, you know, 100% accuracy. And it's, and when it happens to you, it happens so fast. Because every day you have new models, new capabilities. And, you know, even for us, we're like, okay, how do we hire a software engineer? What is the job if it is not coding?

And for PM and Alice, I see that with, like customers all the time. If you're a large bank and you employ tons of juniors and their job is to summarize AK in a nice way, standardized way, now, you know, everything will be done by AI. So obviously the bank, they say, okay, Jim, you cover 50 names. Well, now you can cover 100 names. But also there's a, hey, Jim, maybe you can do more complex work. So let's say you were looking at, you know, Chipotle.

And maybe you can look at also the five companies that are around. And maybe you can produce, like, more complicated analysis. It's not like the impact of inflation on cheap-outly business, but you're going to do that for, you know, like Donald, young brands, etc., etc. And the analysis you're going to produce will be just, like, way better. And so that's one way you can frame it. It's just like you're doing more qualitative work over time. Let me propose a different hypothesis.

And this is one I've talked to people about and I've gotten pushback in different areas. So 20 years ago, 30 years ago, whenever, quantitative skills, right? Your portfolio manager, long-term capital management blew up, but you kind of wanted a long-term capital manager, right? Somebody who could do bath in their head quickly, and that is in part because there weren't even a lot of computers back then, right? The Excel and the modeling has been outsourced, but as, so you kind of get into reading into 10Ks and understanding and getting to niche cases. I want to come back to niche case in a second. But I have wondered, like, if AI, if I could upload the most successful investments of everyone for the past 30, years, put into AI, and AI can start spinning out.

I've wondered if that sort of stuff, like anything that's in the filings, all of that edge, all of that alpha is gone. And where the alpha for the next generation is almost, hey, can you go meet with management in a room and read their body language better, right? Or are you better at kind of like, I wonder if the gum shoes type stuff where, hey, go to the franchisee meeting and talk to 100 franchisees and then plug into an AM on and say, hey, I felt pretty depressed at that franchisee meeting. And then the AI says, oh, the stock is forecasting franchisees are 7 out of 10,

and you felt 3 out of 10, it's a short. So do you think that, like, gumshoe work and that personal work becomes more important with AI? Or you could tell me, hey, AI can get on the earnings call and read someone's body language better than anyone else in the world. So it's actually worse. How would you think about that? I think, you know, having a data age is always the source of,

generating alpha. And I will say, from my experience working with a lot of PMs, all the public, you can generate a lot of information with public data. So you can do it in two ways, just like by better looking at it. So, you know, Buffett is known to be able to compute the owner earnings pretty fast for a company, right? But you can say, hey, compute the owner earnings, get rid of the stock-based compensation. and bullshit and all that stuff

and rank all the company instead of having an analyst doing one or two company the AI is doing that on the whole stock market and then you have in my opinion like data source

that are not sufficiently used I was discussing with one of our it's a big firm, big PM investing in tech and they say hey there are so many podcasts where the CEO of Microsoft is going out there

and it's going to comment on capex on you know the new AI capabilities, or is LLM a commodity? And in that case, you know, they are losing money, et cetera. The guy cannot listen to this podcast. Each of them, you know, the next three-man podcast is like six hours. Do I wish? It will be like four hours.

So typically, for example, what we did at FinTool is we downloaded all this podcast. And then we asked the AI, okay, identify everything that is relevant for an investor. Every mention of CAPEX, every new product launch, every mention of how competitors are doing. And I think you can generate alpha that way. But more to a point, I think we'll see a sort

of like Citadel but focusing more on the qualitative data side. Like, as you said, like looking at a video, analyzing the management pattern, and all they, you know, excited or not and just trading on that.

Let me, I'm not sure where to go with that. I guess when you find people using FinTool or AI in general as an amplifier, what skill do you find they're trying to amplify the most? It depends of how AI enable they are. So we have, you know, we open an account, let's say, you know, 50 seats and we have the guys that are always on CHAPGPT and perplexity. They will pick it up and ask thousands of questions, right? And we have also some people, they don't even know how to use AI, right? They don't even know, you know, about CHERGPT. They don't know how to prompt.

I'm just laughing because I was talking to a very successful investor the other day. And he was like, hey, Andrew, I listen to one of your podcasts. and you said, if you're not using AI, you're going to get left behind. I was like, yeah, I really believe that. Like, I can do stuff in 15 seconds that used to take me a day. And he was like, I've never opened chat, GBT before. I was like, oh, buddy. Sorry, continue.

I was just laughing because you hit the nail on the head. 100%. I mean, you go to this meeting. You meet with, you know, a very famous investor, you know, investing legend. And you're excited. And you realize that, you know, he has never used AI. And obviously, the guy that is 30 years old in the room is a bit like the guy who set up the meeting.

So I will say several levels, very simple. They have a question, hey, she both lay, how many stores did they open, you know, per quarter over the past eight quarters? And, you know, because it's a KPI, they won't find it in Bloomberg, you want to be in faxed, and they have to deep dive into the earning calls and the AK release. And the AI was, you know, create a nice table with a source. It's kind of like 101, right? It's a simple question on a company.

And then you have a hard question where you say, hey, same store sale, but you're going to compare that to the five company in the industry. You're going to read every earning calls and you're going to analyze what the management say about that. So like multiple company and merging like the numbers plus qualitative assessment. And then you go on and on and on and you have like this massive queries. where people like scan the whole stock market for a bunch of preterias, both qualitative and quantitative, like, oh, you know, get rid of the stock-based compensation and do this and do that.

And the workflow takes, you know, like 20 or 50 minutes to answer the questions. I'll just make it personal. And I might even clip this out because I suspect that most of my friends who I talk to, and again, most of my friends who I talk to are plus or minus five. years from me. I talk to investors much younger, but most of them plus or minus five years run, you know, similar concentrated value or event styles and stuff. So I suspect most of my friends are using FinTool, chat chip, TEPT, everything in the same way, and I think this might be useful for them. The characteristic you said at the beginning is kind of how I use this for, and also I might

just clip this specific piece of the podcast out and put it on Twitter because I think this will be very useful for people. The way you described it is kind of how I use it, right? I probably spend 30 minutes to an hour of my day, every day in chat GPT and FinTool, and basically I am using them like a super Google, right? My famous one with FinTool is I was trying to pull what Caesars had said about acquisitions, how well they had done over the past seven years. If I had done that on myself, I would have had to go through 50 transcripts on my own looking, would have taken me at least the full day of work. FinTool did it in 30 seconds, five minutes to go through all that. Chat GPT. I'm using it like a super Google. I'm using them to say, hey, quickly build me same store sales for Wendy's McDonald's Burger King over the past 10 years. It's fantastic for those. What do you think I could be? But again, I feel limited because it's just improving. It's grabbing data and it's summarizing it much more quickly, which is great. It's freeing up tons of my time. What else could I be using to, instead of just using it as a super Google or it's free of time, how could I be using it to make me smarter? Or what else could I be using it?

using to, like, improve my job? I think ultimately, you feel limited and you're limited by the technology. Finance is pretty hard to nail because it's a combination of words, bus numbers. And so all these, you know, finance software, like FinTool, they look very basic, right? Data extraction at scale, looking at the earning calls. But as time flies and we have, like, better models, we can do, like, just more complex analysis. you know, analyzing company ultimately, the two breakthrough will be offline and parallelization. So parallelization, you ask one question and you get an answer. What if FinTool can go and answer

100, 200 questions at the same time? You want a company primer on whatever name, and instead of asking the five questions, FinTool will return like business description. How do they make money, the financials, how do they pay the company, the CEO composition, benchmark that with their peer group, et cetera, et cetera. And offline is something that exists only in software engineering, and it's very recent, is, you know, you work from 9 a.m. to, you know, 7 p.m. What if, you know, FinTool can do research from 7 p.m. to 9 a.m. the next day. Because once FinTool knows how you research company, what's important for you, and how ultimately you make a decision, which

requires you to explain a bit to the AI, I'm excited by this opportunity because the CEO comes from a competitor, is a great capital allocator, they are talking about spinoff, they have plenty of cash on the balance sheet, and just like the agent, you know, will work on the background and try to find you opportunities. And I think it has to do something very complicated. a kinto morning star with the score, it needs to score the opportunity. Say, hey, competitive mode is a five out of five. Management team is the three out of five. And we go from you asking question and being limited by yourself, right?

What type of question can I ask and your time to the AI just like pushing your information? And you go and you say, okay, this is interesting. This is not interesting. And the AI learns and push you even more relevant information. That's a future of AI system where you don't go and ask a question, the AI push you rather than info. That's an interesting. So let me just dig into that future a little bit.

I mean, that's an interesting future because what it would rely on is your past interactions with AI and your past investing, right? So are you saying in the future as AI continues to scale, is there going to be almost a moat in, hey, I can upload the past 20 years of successful investments here? I can upload them to AI, it can learn from that. So I've got a moat versus a 21-year-old who's listening to this, who's like, I want to break into investing. When they start working with AI, they're not going to have any successful investments to point to.

So they're going to be way behind because they don't have that historical data. Does that question make sense? Yes, it makes sense. You can also argue that there is a absolute, you know, good investment, right? If you look at the track record of like Apple and Coca-Cola, you can deduct. from the past, those were good investment. And then you can try with the AI to ask, yeah, okay, why they were a good investment

and rank, you know, Coca-Cola. How much of that is, how much of that is end-of-winning, though, right? Like, yes, Coca-Cola and the 80s and Buff makes great investment. Philip Morris famously best-performing stock of the past 50 years or whatever. Like, how much of these are end-of-one where I was like, hey, like, I could make a very simple argument for it. Hey, Philip Morris, addictive product, good brand, big modes, you know, you're selling cigarettes, distribution, addictive products, low PT, great. Yes, it's awesome. Or I can make another argument

like, hey, Philip Morris, find me something that trades for an eight times P because they've got the threat of the government bankrupting them over their head. Like, that was very much end of one. And historically, it looks great. But things could have gone a different way. And the government could have demanded a, you know, Fannie and Freddie in 2008 style pound of flesh where all the equity belongs to us now. So how much can AI earn from that? I have a follow-up question to that, Yeah. So, you know, when I discussed with my friends that are working up to the high-frequency trading firm, they have like terabytes of data. And obviously, terabytes of, you know, the price and the volume and the options and all that stuff. And for AI to be on the qualitative

side, it be omniscience, like a bit like the Warren Buffet, it needs a bit of everything. It needs the news, you know, the newspaper at the time, the story at the time. And the more data, the more context, the AI will be able to, you know, understand what, what is a good opportunity. Because often, you know, Mr. Market is just crazy. And it's crazy because like China tariffs. And maybe the AI can read, you know, what the White House says and what some experts are saying and say, okay, I think like humans are panicking and I think it's a buy. Right. So it needs a lot of data. And it's still the early days. I'm laughing because you say tariffs is like, Man, try and trade an AI on win and what Donald Trump's going to tweet.

Good luck with that. So we've mentioned Renaissance several times. Quantitative has been taken over by computers. There's no doubt about that, right? Like quantitative is dominated by computers, learning models, AI, whatever you want to call it, dominated. Qualitative to my knowledge has not, now there are things, you know, you can get a quantitative fund that runs on value package to everything, but concentrated qualitative has not, I haven't even seen anyone try to do an AI. Have you guys thought about back testing or launching like,

hey, here's our qualitative AI concentrated portfolio? How do you think of fun like that will work? Or do you think there would be any success there? What is very hard is when you have like machine learning models and you want to test the models, you need quickly to see the output of the test. So if you train like a recommendation system for, I don't, Tinder, you want the right, swipe left and look at the data. But when it comes to like a long-term, you know, investment portfolio, you're not going to wait 10 years to see like if this would be back testing, right? Like you could back-testing into, and again, you get very small sample sizes because even

I said, Nick, back-tested to 2005. So we've got 20 years, eight-position portfolio, and you're holding everything for three years. Cool, that's roughly 17 times three. So cool, we've got, what is that, 54, 51, 51 data points. Like, that's not that great. But it is interesting because if we're saying, hey, the future AI is going to be increasingly better qualitative, launching an AI fund right now would kind of be the model, you know? Yeah.

So I think it's a bit hard with large language model in the sense that if you show them just like past data and you try to find some correlation, the LLM baked in in their training set, they already know the future because you can say to the LLM, look only from 2000. and 8 to 2015 and, you know, give me an answer for that parameter. But the training data, they know the story. They already know 2023 or they have a guess, you know. You train on Apple and you say, look only at this, but but the LLM knows that Steve Jobs will come back. And you have your prompt where you say, hey, you know, just consider that time frame.

So I think that the long term is very hard. That's why I think quantitative finance is focused on the short. I did see a large trading operation in New York where they were kind of like trading the news. And so it was like a bunch of like kids 25 to 30 years old. I mean, in Bloomberg Terminal, all of them. And they were like reading the news and say, okay, there is like a gas leak here. And then they were trying to find companies that are, that have like exposed to this, you know, gas thing. and then they were like quickly looking

and you know the balance sheet and stuff and like trying to short or you know I do think like your LLM can do a better job for that you read all the news I have no doubt about that though you do worry but I have no doubt it could read a news article and do exactly what you're saying faster than that run though I will say I do remain

impressed by how often you'll see something and an hour you'd be like oh that might not have been that good for this company and it'll take kind of an hour for its ability to the price so maybe there's still room for humans or maybe AI can do it better. I feel like we should think about launching like, hey, here's an AI eight stock portfolio. It like finds the best qualitative things. I think that could be interesting.

Yeah. Let me ask you separately. So speaking of concentrate portfolios, I was talking to another friend and I get similar conversation where I was like, man, I'm not saying all your thinking should be outsource AI, but I was giving them my pitch. Like, it really, it frees up a lot of time. You know, a lot of things you do by hand, it will free up. And he had a twofold pitch.

One, and I agree with him. I actually kind of cleaned this model up. He was like, look, there's the old thing in banking. I build every model myself. And it's not because the Bloomberg models aren't impressive or whatever you want to download. You build every model yourself because then you're building your understanding of the company, right? And there's just something unique about plugging in. Next year, revenue growth will be 5% and kind of flowing it through all the income statement.

and seeing how it impacts the company versus just having it presented to you, right? So his thing was once you start using AI, you kind of move away from the building monitor yourself. So that was number one. I'll let you respond to that. And then I want to hit you with his, like, maybe more powerful point on number two. I think it's very true. I think that's the unknown is like for us, you mean the more we take notes,

the more we do the work, like the manual work, then we learn. but then in the i was i was chatting with the customer this morning they were analyzing the compensation of the ship of place CEO so they say okay you know they look at the def 14a they look at the he has a cash incentive of like two million dollars and there was like okay the two million dollars 75 percent of that is like company performance factor and 40 percent of that is like comparable sales uh comparable restaurant sales and then cash margin and stuff. So, you know, they spend a lot of time.

And they have come up with this analysis on the comp of this guy. And we're discussing together is like, yeah, but look, the AI can do this, can do also the previous CEO, can benchmark this composition with this new composition, it has a stock and can look also five different companies. And so, yes, you can do that manually for like two hours and learn about his compensation, or you can have an output and start learning about it. is the composition standard, is everything we are the composition of the CEO before and the CEO now. And ultimately, I think you will have to digate that to AI and try to learn on more complex information.

And at some point, you just have to have the taste, the pattern recognition, the understanding, where you say, hey, AI, gather me all these trends, and then I'm going to think hard about this. I'm going to stop spending my time on just like, you know, data extraction. No, my solution, and not that my solution is perfect or anything, but my solution has been I love using AI for the broad stuff. I mentioned the Caesars earlier, or I believe that Adam Reni for Fintel now, I love saying, hey, go summarize the past five years of executive comp and trends at this company and maybe some of the company's peers, right?

Because that information, that takes multiple hours digging through these processes. These proxies are massive statements, and I feel like they're intentionally a little bit oxidated, and AI can return it like that. And then what I tried to do is look at that, learn from it, and then the company I'm really focused on maybe spend a little bit more time just on their most recent proxy, right? So I'm getting the nice AI summary, and then I'm trying to understand the most recent one because the most recent one's the one we're on, and it's the one that really matters. And I can use the insights from the previous ones and the peers to do that.

So that's just been my solution. I'll pause there if you have any thoughts on that process or anything. And I do have one more question I wanted to ask. Yeah, I think it's exactly right. And back to the conversation I had with a customer about Scott, the CEO of Chipotle. So he did all his analysis. And I think we did the same with AI. And he missed something.

It's because when you ask, like, Fintu, I compare his composition to five other CEOs. Fintu will flag, hey, there is something that if the guy reaches like 200% of his target, the rest is not paying cash. It's paid in RSU. And, you know, that's something you want to know, right? But it was probably buried in a footnote or whatever, and you kind of, like, missed it. And it was a key information. So, yeah.

Just, well, this is off the cuff. But one of the things I always am interested in is executives with non-standard compensation. So probably the most famous one would be Elon Musk in 2018 or 2019. The Tesla's unadjusted stock, I'm making an abrupt numbers, is 100. And they give them tons of stock options that say, hey, if this goes to 300 in the next eight years, we're going to give you basically $50 billion. But, you know, if it doesn't, it's zero, so you're taking a huge upside bet. And he gets it.

And then, you know, the Delaware judge says, no, and this is all pre-split. So the stock's way up. But that was a very non-standard package. I just want to, how have you found, you can say Intel specifically, AI, whatever it is? How have you found them analyzing non-standard package? And I think this will flow nicely into my next question. Yeah. So what we do, and we try to run this analysis offline,

we for instance, run models and look at the change in compensation metrics. So they say, hey, he has an annual cash incentive, and here are the factors. And we are asking basically FinTool to look at all the default in A and say, hey, did they change the factor over time? And then you can do one more step. Is that correlated with the company, maybe netting? income decreasing or the stock price crashing and trying to maybe identify that as an early red flag so new death 14a aha new compensation it was like 50% of the net income now it's only 10

what you know what does it mean what do they know and so yeah that's we try to do this sort of like complex analysis also like the non-standard comp like sometimes they have like a they don't have a I can't, but then they have, like, private security budget, private jet, you know, school for the kid type of thing. That's standard at this point, to be honest. Yeah, exactly. Where is my, I should I spend it for my private jet? I can, I'll point you to guys some real fun ones if you want.

No, one thing, I should probably work with you guys on this. One thing I'm trying to work to build better is one of my favorite signals is, you know, most companies will grant RSUs and options once per. year. So, you know, February. Every February, all the execs get RSUs and options. And when they do, they file for them for us. But every now and then, a company will do off cycle grants, right? They'll say, hey, we're giving them in April this year as well or something. And that is a really interesting signal, right? Because they normally do that because they do it in April because good news is coming in May and they want to get everybody paid. And I've like, every investor loves

those signals. But it's very hard to pick up on. And I will tell, I'll give a small but like I tried it on chat GPT and FinTool, and ChatGPT's was terrible. It just like gave me only and FinT tools gave me 20 examples. And I'd say of the 20, like only four or five were actual good, but chat GPT managed to give me zero and FinT actually got me some examples of them. So it's something I can probably modify, but I think it's just a really interesting use case where everything sees just a form four, but if you can find like the right context and stuff, AI can solve it.

Let me ask my other question. So my friend who I was talking to, the other thing he said, I can see this. He was like, look, Andrew, again, quantitative for AI, quantitative is so good, right? The Renaissance, large. When you're running a concentrated value fund, the edge is in the nuance. The edge is in the edge cases, right? And this is an example I came up with, but I think you'd agree. Cliff Sosen with Carvana.

Carvana is a stock that was down 90%. it had every short seller in the world has published a report on it it was way over levered you know the profits were going down it was shrinking like if i had given you that set of if i had given you that set of facts every base case everything would have said that is a zero right yeah that but it was the edge case where and you know this is end of one but it was the edge case and what he was saying is edge cases right oh carvana also was the only person who could make that business model profitable. There are lots of other cases, but he was saying, look, you want the nuance, you want the edge case, and that's where the real money is made. And AI is never going to be able

to detect that edge case, because what it's going to do is it's going to read Carvana's 10K, and it's going to say, hey, I've read 2010Ks like this and everyone has failed. Everything that has all these red flags, the base rate is terrible, so it's zero. Now, maybe that's saying Carbana was like your negative EV lottery ticket that actually paid off, or maybe it was But he was saying you're never going to find that edge case. What would you think about that? So every Friday at Fintu, we do a presentation on a company. And I think two months ago, I chose Carvana.

So I think it's a great example where I asked Fintou to look at the accounting, compare with other companies in the space. And then I put the Eidenberg research report and all these research that was like super bearish. and obviously tons of red flag right the the father runs a company he loans money to corvana it's blah blah blah and so yeah it was full of like red flags and i think two things first that's why we need the human in the loop someone with the you know that is highly paid for the insights and and that's true for now for lawyers because like AI is writing everything but you need someone with like a deep understanding and someone to make the bet.

And second of all, that's why also finances really hard is because it's fatale. It's because this carbonate thing, it might have been a zero. It might be a zero now. Maybe it's a fraud. I mean, we don't know, but yet the stock is rising. And if you were shorting, well, you lose a lot of money. And so I think that's what we need the, you may need the loop is because it's fatale.

And if it has consequences to be wrong. And sometimes, you know, in one second, you can be wrong in this game over. No, and look, it's not carbon Tesla, right? Yeah. The red flags and the short sellers who've been burnt on Tesla are unbelievable, right? And the base case, hey, we're starting up a auto manufacturer. Cool, all of those go bankrupt. Like the red flags with the, like it just goes.

Now, again, this comes back to, I think my friends would have said, hey, if you built up a bucket of companies that had all of the carbana, all of the things I just said are all the Tesla characteristics, 98% of them would probably massively underperform, but it's the two that work. And if you're running a quant where you're spreading the bet over 100, you can do that. But if you're doing qualitative where you're saying,

I'm picking one, like, it's very difficult to use the AI there. So, yeah, I don't know. Yeah, because you don't have a basket, right? Because you can make the case where you create a sort of like ETF and you show the ETF of all the companies that AI has identified. But if the premise is I have like 10 stock and a super concentrated long on the investor, then it's hard. And for Tesla, I try with Finto.

I say, hey, give me the real value of the car business. How many cars are they selling? And get rid of this like EV credit stuff. And, you know, it comes up with maybe a valuation number. But then you need to say, okay, yeah, well, you need to factor the fact that Elon Musk is Elon Musk, that the credit thing is good and blah, blah, blah. and ultimately you make a decision. I've done that math, but as my friend Dan once said,

look, the issue with going short Tesla is you do the math, and you're like, okay, the entire global auto industry is worth $300 billion, and Tesla's trading for a trillion. So it's worth triple the global auto industry. Like, that's your bare case. And then the bull case is Elon Musk is taking us to the moon. It's like you're just talking different stories at that point. Yes, go ahead.

No, I was thinking about when I read the Fintill analysis, there was like a part about the super goodwill that was Elon Musk and also the long shot. I think it was like Optimus robots and the sort of like Nvidia competitors that they are trying to build. And, you know, even if the AI was like, you know, it's Donald's goodwill for the company, but, you know, maybe not, you know, 700 billion.

But, you know, that's why you need the human in the loop. the end. No, and just not on Tesla, but like, again, it's on the edge cases. There's a company that me and my friend debate all the time, and he's like, look, this company has built this killer model. They've got this killer mousetrap, razor razor blade. The returns of invested capital are going to be crazy. It's a medical device company. And I'll be like, that's cool, but you're, we're valuing them at a billion dollars. And it took, like, there are indeed to develop this device was four million dollars. Like, it's really hard for me to understand how you can make a billion dollar a billion dollar company with a four million dollar r&D product and like i understand

more it goes into their salesperson every bit like it feels like somebody else could come and copy this product and sell 50 percent cheaper but it's kind of where like look tesla for years the short one of the many short things has been hey their r&D makes no sense versus when they're training globally but i guess somehow they're making it work you know just a couple more quick uh this might be the last question, but one thing, I think FinTool might be about to change it, but FinTool right now, it's only focused on SEC filings. And I go back and forth with like, hey, is it better when I'm evaluating a company to do it in Vintel and not have to worry about the hallucination where like I've had chat GPT when I ask it, hey, what was this company's

earnings three years ago? Like it'll pull a different company's earnings or a pull a earnings number from seeking alpha, which was not the company's earnings, right? Yes, I know that. What's the advantages in disavit, and it doesn't have to be FinTool specific, but what's the advantage and disadvantage of saying, like, hey, let's only use company filing. So, like, I've got something where, you know, if the company is lying there, they will be held legally liable and go to jail for it. Like the information there can be generally trusted unless it's an outright accounting fraud,

versus, hey, let's go broader and incorporate things on the internet when I'm kind of studying this company and using AI. Yeah, I know you're right. I mean, one of our biggest customer, at first of all, like, hey, don't bother I'm using chat GPT. And I think three months after, they realized, like, nonstop hallucinations, right? And using the surge. And we actually, we have a benchmark. There's a benchmark called Finance Bench.

It's the leading benchmark for equity research where Alice put, you know, the question and the answer. And then you run LLM model. So you run chat GPT with surge. You run Cloud. You run FinTool. And then you have a score of accuracy. Right. And I think Finto was like 98%, chat GPT, I think it's like 40%. Perplexity might be like 50 something. So our approach is we have to start with the ultimate source of truth and be very good with that,

even if, you know, it means the product might be, you know, slower and limited in data sources. Now, you know, we added after, you know, investor presentation and earning calls. that we added all the YouTube transcript that we consider it as a source of truth because if Zootlebird on the podcast is saying, you know, talking about his vision of AI cap expanse in 2030, you know, it's him.

So it's a source of truth. And now we're going to release the next two weeks, the web. And this is a tricky one because, yeah, if it searches like seeking alpha and multiple and return a busshot data, it's game over. It would be wrong, right?

So the way we have to build it is it does search the web for very specific information. And then you have it's a bit slow because then we have to verify the accuracy of every piece of information. And then integrate that into the answer because the best way to lose a customer, to lose a PM is the guy asked about like dilute your account, whatever, whatever. And you show an answer with either like an hallucination or a wrong number. he was like lose trust and, you know, be very unhappy. And so that's why also finance is hard. And that's why most like finance chat retrieval software or a bit slow is because if you want to do it, right, it's like so many different steps.

Let me just in terms of my use case, it's my personal use case to make it very selfish. I generally am not uploading much to FinTel, JETGAPT, name it. name it when I'm doing something. Now, I will point them to, hey, I'm looking at GE, particularly the GE's 10K, like, tell me X, Y, Z, or I'm looking at Pfizer, and I want to understand their oncology program. Bizer just had an oncology investor day. You can find the transcript here. Go look at that. How much should I be uploading my personal stuff onto chat GPT into whatever it is when I'm working with Aon? also the perspective is as a software vendor we do everything so you don't have to upload the data

so for instance we will process you know 20,000 podcasts for you and you will never you know download the video or download the audio whatever so you will never do that having said that they are private internal memo maybe Excel file stuff that you own and that might be good for the LLM to see. And then, you know, sometimes I don't see what customers, you know, upload, but some of them they tell me, you know, they have this interesting blog post that they think it's valuable. And we're going to not catch that blog post. Or they have this piece of like saleside research published by one of the friends, right?

Then they're going to upload that. Your clients are uploading? Are they getting better results, do you think? When you upload, I mean, for instance, I was chatting with a guy, there's this famous blog called Semi Analysis, which is they study the NVIDIA of the world, and they do very, very detailed research. They've been on the podcast a few times, yeah. Yeah, and this guy is a genius, and yeah, his analysis is so much worth it. And so when you upload and then you ask questions about liquid cooling data center, well, yes, is analysis, you know, matter in terms of market size, understanding of the technology.

Maybe he has a piece of, yes, the Abilene data center is close to completion and it has like 100,000 GPUs and they want to expand to 200,000 GPUs. So then you have the information. So they are going to buy GPUs. It's a buying pressure, you know, I want to say that the more information, the better. No, it's just, I'm surprised you said that because a lot of that information is in the public domain, right? So I'm surprised like uploading, it sounds like what people are trying to do is more say, hey, I'm uploading something in the public domain, whether it's podcast or that post, and Git FinTool or AI or whatever it is to really focus on it when it does the analysis. Am I thinking about that correctly?

I mean, public at the end of the day, we're going to capture, but let's say you have an expert call. Let's say you call the management team, you call IR, and you have the transcript of a call. It's not available. It's only on your computer. This is the sort of information you will upload. Should I be legal requirements on everything? Should I be recording my calls with management teams, whatever it is. Should I be recording them and then uploading them when I'm looking for processing?

I have discussed with so many customers where they use recording software. They use like granola, they record on Zoom. And then on their SharePoint or OneDrive, they have all the transcript. So, again, if it's legal, I do try to get recordings of my calls with Mnivis. But it's generally for me to go back and listen to my notes. And one thing I worry about, like, I've had this issue with chat GPT before, where for a while I was doing a lot of legal analysis in it. And I noticed, hey, if I upload six, if you and I are in court, I upload six points, you and I

I upload six filings from you and five filings for me. And I asked Chat, CPT, who has the better argument? I noticed it was always saying you because I had put six documents from you and five documents for me, right? And I worry if I do a management call, like, it's going to incorporate too much of my bias. So if I do, I can't help it, right? If I'm short a company, I hate a company, my questions are probably going to be more

pointed than if I'm long a company. Or if I'm long a company and they announce a really bad quarter, my questions are probably going to be more, and I'm worried like it's going to incorporate too much of my bias. Now, obviously, there are things there are said that, not that I'm getting MNPI, but they're just different or phrased different than they were in public, but I kind of worried about giving too much bias when I was uploading that if I'm asking the AI to interpret it in any way. Yeah, you're sure, right. I think like the chat GPT, look, Chad GPT is a horizontal product.

It's worse like 20 bucks a month. They are trying to cover like many, many use cases. Most of them are like consumer type of use cases, right? And that's why in every vertical, you have, like, big software vendor that leverage the same type of technology, but they do a better job at understanding the information, avoiding the biases. So if you go to legal, you'll find the Harvey, the Legora, which are like big software vendor now. And yes, when you upload like a brief and you compare, it won't say, this guy is better because he has like, he uploaded like more data, you know, compared to church it. And then the next next step is how you can reduce the bias. And I think I see that in earning call.

If you don't do anything and you just feed an earning call and you ask, you know, the LLM, is it bullish for the company? Is it bearish? Well, if the stock is down, the CEO is always like pumping up, yes, we're going to make it. We have this new transformation. I am cutting costs. And so you need this extra layer.

So the AI can understand, okay, be rational, forget about his language, look at the numbers, put that into context, look at the five earning calls, look at, you know, other companies, and then answer. If you have five more minutes, I'd love to just pull on this just a little bit more. So when I'm, let's say I'm going to start recording my calls with management teams and uploading them to FinTool, chat, chat, jeep, whatever it is for AI analysis. How should I be thinking about my conversations with management teams in ways that would make the upload? maximally beneficial for AI? First of all, when you're meeting the management team, we have this like workflows in FinTool, which is sort of like prompt library.

And one of the most popular is like read the past three to four earning calls, look at the number and generate questions for the management team. And it is used a lot. So I guess like customers just like ask Intel to generate very thoughtful questions. Do you know why I choose it? Because every investor has a, situation where they're going, you know, you're going, you're flying to a conference. You might as well meet as many companies you can while you're there.

You have four companies you know and you get six companies thrown in your schedule and those meetings are the worst. And I will tell you, having been at big firms before, people will pay huge money to get accurate, to get good questions to ask someone like that are more than walking you through the 10K just in that situation alone. So I totally get it. It happened to me Monday night. I was with a launch bank.

They do a bus tour of every. every, you know, tech public companies. And, you know, PMs, a very big firm, top analysts. And they went to L.A. and they had a meeting with Snapchat. And here they had a meeting in their self with DoorDash and, you know, great companies. And they were like, and the girl was like, what are we going to ask on Snapchat? And it was like, so lost. And in his head, he has, okay, well, stock-based compensation is through the roof.

Rovenue is kind of flattish. And it was like, I need the AI to ask thoughtful question. Otherwise, you know, the meeting, when I'm going to ask, anything, it will be weird. So, yeah, I see the use case. Oh, but back to my original question. So I'm going to have a conversation with the management team, right? And I'm going to report it. How should I, heck, I might start doing with this with the podcast because obviously I hope I'm having smart guests on on companies.

But I'm having a conversation with the manager team. How should I frame the conversation in a way that I could then upload it and make it max beneficial. Do I need to be steering them to, like, providing numerical answers to everything to make it easier? Do I need to be asking them to keep everything, like, tightly defined? How can I ask, get a good transcript or recording to give it to AI? I think now you have, like, several, like, recording software. I think Ranola is the leading, like, recording software. They will do a great job at just, like, getting the whole, the road transcript, creating notes. And so I guess at the end of the,

the day is how can you ask like small question to extract some sort of additional information? That's it. What type of questions would most feed the AI with something? Like is it getting numerical answers? Or would it be like, hey, you guys said same source sales were down 2%. Can you break it down in four different ways just getting like statistical breakdowns of things that they've already said? What would be the best? But something I learned talking about the AIR person at Datadog. She's extremely smart. She was, I think, a PM before, and she said, I am answering all the question with all the

publicly available information. But there's a trick is sometimes people don't ask about the company. They ask about the competitors. So, you know, that they are in the dollar tree meeting and they say, hey, do you think like dollar general, they see a sort of like inflection point or impact on inflation or as a way to get an answer, and they say, oh, yeah, they see a ton of, you know, pressure from inflation, and then you can deduct that there, too, they have, they struggle with that or their inventory. And I don't know. I think the heart of asking good question to management

team and having them answer the question is, I don't know, it's an art. No, look, what you said there is interesting to me, because I had not thought of, as they're saying, hopefully they're not giving you MNPI when you're calling like the IR team or the management team. But I hadn't thought of that. Like, hey, forget your company.

McDonald's, forget you guys. I want to talk about Burger King. Like their numbers looked a little bit different than yours last quarter. Let's talk about Burger King and what they're seeing that's different. And then, you know, as when, if they're saying, oh, it's got to be inflation.

Burger King's seeing a lot of inflation, then you can probably guess McDonald's is like, has inflation on their mind. So that's really interesting. Nick, this has been super useful. I probably need to have you guys spend more time. Dack actually showed me, you mentioned the workflows, and I've been spending a lot of time with some of the historical comp workflows.

Again, that's where I've really gotten a lot of benefit from. But I probably need to spend more time on it. Any last thoughts before we wrap this up? No, well, a last thought on AI is that it is extremely early that I think most people form an opinion on the things. technology as it is today. And because as human being, it's very hard to grasp what an exponential is. Sometimes you try something, three months later, it's completely different.

But you have the opinion of what it was like three months before. You're hitting the nail in the head. It's funny you say that today because it was last night or today, Google released their new video editor. And people are like, here's the video that they were pumping out three years ago. And here it is today. and everybody was mocking it three years ago and today it looks like

like literal best movie quality things and I do have some friends when I talk to them I'm like hey I'm spending an hour in AI every day whether it's chat TPFint or whatever and like an hour like I tried that two years ago and it was worse than Google I'm like man two years it's just crazy

how much better it's gotten over that time yeah that's why you need to stay curious and it's also hard you know as I'm always pushing my team like hey use this new AI tool, then they're like, hey, but it sucks last month. And then you say, okay, I will give it a try. It's like, oh, wow, I, you know, I'm doing so much with it now. And it's just exhausting. Just like keeping up with AI news in general is exhausting. But at the end of the day, it's worth it because if you can have an edge, if it does your job,

you know, that's good. The worry I, like, A, I want to get mad at my friends when I'm like, I'm spending a lot of time in AI. And they're like, that thing, it's so dumb. It's like, dude, I literally just had a conversation with you where I'm like, hey, I'm worried AI is going to replace every quantitative finance. And you're like, AI is not as good as Google. Like, yeah, I think it's probably a little bit better than Google. But, you know, it's interesting. Like, as you said, stay curious. But I'm worried it's one of those things, you know, if you weren't using email 10 years ago, you were just dead, right? Like emails, table stakes. And I'm worried it's one of those things

where, unfortunately, as Buffett said, with the textile mills. Like, you invest in new textile mills, and it sounds great, but everybody makes the investment. So it's kind of, table stakes just to compete, nobody earns excess returns. I'm worried AI is one of those things. Like, it's table stakes. If you don't use it, you get your face ripped off. And if you do use it, like the returns just come down even further because it makes the market's more and more efficient.

And it just gets tougher and tougher over time. The gap is so huge, though. I mean, I work into, I go into big firms and I don't want to say names that I, we are not allowed to use like chat GPT at work. So we don't have it. I cannot even use it on my personal computer. It's some, we have some, sometimes. we get call from CEOs that's like, I need AI in my firm, but most of them, they don't even

know what it is. So I think the, at least for now, there's like a huge gap in the market. Sometimes we see some customers, I don't know, like Kennedy Capital, those guys are, you know, small cap oriented. They are in St. Louis' misery, okay? They are so AI enhanced compared to some of the folks I go see New York. And it's just a matter of like, you know, they had like two, two, three great PM there. they were like, okay, we need to get on this thing.

Well, I'm going to be following up and making you make an introduction to me to Kennedy Capital because I want to pick their brain and start talking to them. But Nick, this is awesome. Look, again, Vintel was like the one that opened my eyes to, oh my God, you start putting things in and things that took you a day before it just like spits it up like that

and you can spend a lot of time on other stuff. So appreciate you coming on. We'll have to follow up in the near future. Cool. Thank you. A quick disclaimer. Nothing on this podcast should be considered investment advice. Guests or the host may have positions in any of the stocks mentioned during this podcast. Please do your own work and consult a financial advisor.

Thanks.