00:05:49 · Brian Neider
[00:05:49] S0: For a lot of the businesses that have been around for a long time, and we invest in software businesses and tech enabled service business, Internet businesses, they need to figure out how far in front they are from an AI standpoint and understand how AI can have some type of an impact on their business. And we see a lot of businesses that have operated for a really long time. They might be a really good business historically, but for whatever reason, they kind of wanna bury their head in the sand because dealing with the AI sort of threat and opportunity, by the way, ends up being a pretty difficult thing for them to do. And so a lot of times companies that we see that are slightly older in nature or have been around for a long time, the most cagey things that I see from them are the resistance to want to talk about how AI is impacting their business. So it's a little bit less so on kind of like the numbers. There's definitely that aspect, and I'll get to that in a second. It's about not wanting to talk about the elephant in the room, and I think that that's that's a difficult one. For a lot of the AI first businesses, so businesses that are already sort of using AI or they have AI related products, it's a totally different story. And that one, I would pin more on kind of like the numbers. So they'll have a super strong vision on where the world's headed and how they're gonna have impact in this, you know, brave new world. Oftentimes, we'll find that these companies are very late stage in terms of revenue, so they're more than willing to share their revenue. They're more than willing to share number of tokens that they're that they're using. But their unit economics are totally upside down, and so you have situations where you have really large cost structures, reliance on other LLMs to to run the businesses in the first place, and we see a lot of, like, normalization of the gross margins. So you'll have situations where costs are set aside almost like on a pro form a basis with the assumption in the future there might be, like, more use of open source models or costs from the existing LLMs coming down and the token cost decreasing, and everything's gonna be fine in the future. All that might be true, but oftentimes when you start poking around, the CEOs get a little bit cagey.
00:07:59 · Brian Neider, CJ Gustafson
They might even present some of the data without footnotes, and it's really hard to know what's true today versus what's laden with assumptions for the future. And that becomes a difficult conversation because you're trying to piece together what does this look like right now in the here and now versus what does this look like, you know, in the future. And that's a completely different ballgame. [00:08:21] S1: Right, wrong, or indifferent. AI helps us. AI hurts us. We're insulated from AI. How important do you think it is just to have an opinion? [00:08:29] S0: Well, it's super important to have an opinion because I think that it's just the way in which the world is moving. And I think that, you know, the idea of burying your head in the sand just is never gonna be a good thing. So I think a lot of times we have a situation where we'll see a company that says, basically, we're doing the same as we've always done and things will be fine. And it's okay to have a point of view to say this isn't gonna change in the way in which everybody thinks it's gonna change, but you should have some stand... Substantiation behind that. [00:08:54] S1: Got it. And and on the second group whose unit economics may not be where they want it yet, are they showing something to you that has weird adjustments to it for what they think the future run rate state is going to be, or what's going on there?
00:08:54 · CJ Gustafson, Brian Neider
[00:08:54] S1: Got it. And and on the second group whose unit economics may not be where they want it yet, are they showing something to you that has weird adjustments to it for what they think the future run rate state is going to be, or what's going on there? [00:09:07] S0: It's usually something of what... Like what you're talking about. So it might be something where it says, look, here's our actual financial statements, but maybe it doesn't even present the actual p and l. And then there's some type of pro form a p and l where the revenues actually match perfectly fine, but there's some type of assumptions laden in the... Usually it's in the cost of goods sold where they're making an assumption. They might not exactly be foot... Footnoting it. You get the metrics. They say our gross margin is x. And then when you get the actual financials, it doesn't match at all. And then you start asking the questions, and that's where it gets a little bit difficult and and awkward. [00:09:45] S1: So so they have the real revenue, but then they're saying, oh, but by the way, we think Moore's law is gonna kick in here, and our gross margin actually is gonna be 40. It's gonna be 65. [00:09:53] S0: Something along those lines. And, you know, it's like, well, okay, over what time period is this gonna happen? And why did he not denote that this was kind of like the future state? Why is this showing as 2026 revenues?
00:09:53 · Brian Neider, CJ Gustafson
[00:09:53] S0: Something along those lines. And, you know, it's like, well, okay, over what time period is this gonna happen? And why did he not denote that this was kind of like the future state? Why is this showing as 2026 revenues? [00:10:04] S1: Well, I think it begs the question, is is there a difference between selling your company well and being intellectually dishonest? Where where exactly is the line there?
00:10:13 · Brian Neider
[00:10:13] S0: It's a hard one. I mean, you wanna sell your vision. You wanna be bullish, but holding back material data is something that we... Like, for us, we we have a hard time about. It's a lot in kind of how the market is receiving you. So if there's investors that are super bullish in your space and you could get deals closed, like, without sharing too much data, that's just the market speaking, and don't hate the player, hate the game. There's a lot of situations where companies might say, look. Like, this is the data pack that we're sharing. There's a tremendous amount of demand for what we're doing, and you could take it or leave it. That's a a difficult game to play, but a lot of investment firms are willing to play it. And that's one aspect. I don't look at that as being intellectually dishonest. I just sort of look at that as the companies are sharing what is necessary to actually get the deal done, and there's actually just so much demand that they don't need to be doing anything other than what they're doing. There's other situations that I think are a little bit more acute where maybe there's not that demand for the company. Maybe it's not one of these situations where, you know, a company says, okay, we're raising and they give a small data pack to 10 different investors and have a billion dollars of demand by the weekend. Maybe it's like more of an ordinary course type situation where companies willing to share data, investment firm gives them a diligence request list, says, hey. Look. These are the 10 things that we need or 20 things that we need. Could you share? A lot of times what we see is that there's management teams that are consciously not sharing certain aspects of that data. So they'll say, look, we gave... You asked for 10 things, we gave you eight of them. None of them were the single most important thing. And usually the one or two things that they didn't give you oftentimes can be the single most important thing. And there's usually something hiding behind that. And when that happens, you start... You know, you need to question it. And I look at that more as oftentimes more dishonesty rather than, you know, the market's super hot and we're just gonna be getting something done. We only need to share what we're we're gonna share.
00:12:13 · Brian Neider
Sort of like selectively sharing in a situation where there should be a little bit more trust, and that's where we have a difficult time.