Sam Lessin on Business Models & Strategy

Essays

From the Podcast

... Yeah. It's just, it's not a terrible thing to try, but the thing for me is, like, I keep going back to this fact. I use a lot of Claude. I think Claude's excellent. It's like the core of my infrastructure brain. But if I woke up tomorrow and Claude was gone, I would care for about one second, right? And, like, that is the fundamental baseline problem in the core business model, not necessarily unique data, wetlab, vertical integration, bogey, solve all cancer in a way others can't. Like, there are still option values on having this, but you gotta support a massive core infrastructure and cost structure, and the question is, is there a core business or not?

— Sam Lessin, More or Less Ep 169

The forward deploy thing actually weirdly, you know, you look at the Palantir or whatever, in a story of it's a security problem, it's an integration problem, it's hard to switch, it actually kind of makes sense for locking a bunch of people in. It's just a very different business model if it's like what we're doing forward deploy and what we're gonna get paid on is the tokens. But actually, the reality is, and this is my biggest thing-

— Sam Lessin, More or Less Ep 169

Like, people are posting these memes of, like, of, you know, these SaaS companies being like, "Welcome, Anthropic."

— Sam Lessin, More or Less Ep 169

Have you seen the memes of SaaS investors welcoming Anthropic to the no growth table? It's really funny.

— Sam Lessin, More or Less Ep 169

The comment is, but no one's out there saying, "I'm gonna spend billions of dollars building a model I'm never gonna release, ever, and I'm, like, holding it in reserve for when shit goes wrong, and you can call me, and you can pay me $10 billion to fix your bank," which I think is the business model.

— Sam Lessin, More or Less Ep 167

And it's gonna cost you, like, a gajillion dollars, right? And so you're just like, the entire business model of the labs... First of all, they caused the problem to some degree. It was gonna happen anyway. But you're just like, "Look, I don't- can't shut this shit off," and the only thing you're gonna be able to shut it off with is, like, effectively a zero-day model, like a new intelligence that hasn't already been exposed. 'Cause the second it's exposed to the web or if people have seen it, it will get built on top of and built around, so y- it's over, right? So it has to be zero-days, but you're basically out there developing zero-day fancy models to quash the bad models that take over banks. Like, that's how you make a lot of money.

— Sam Lessin, More or Less Ep 167

But the business model is you don't release the models. You basically just, like, keep, like, the nuclear weapons, and then when a country gets subverted, there's only one person or two people you can call to fix it.

— Sam Lessin, More or Less Ep 167

I don't know. It's an interesting question of how it plays out. I mean, like, look, this does speak to my other big hot take on this, which is the only way, in my mind, that Anthropic and OpenAI become really, really valuable companies long term, not trading, not short term, whatever, five to 10 years, longer, uh, is effectively as defense companies, right? If you said to me, "Hey, what they're actually doing all day long with these super fancy computers and all this computing," you can't compete in the token marketplace You just can't, like, make any money. What you can do at the edge with, uh, 10,000 PhDs is develop a bunch of, like, zero days against other LLMs and not release them. Like, make Mythos squared-

— Sam Lessin, More or Less Ep 167

All right. So Dave, here's the funny thing. One, it is narrative capitalism because the SaaS thing was a fucking narrative, right? The narrative is you acquire a user once, and they'll pay you forever, so it's an infinite LTV, and that was the narrative of the SaaS era, which collapsed.

— Sam Lessin, More or Less Ep 161

Had one of the most beautiful public market stock charts in the history of companies until about a year ago, and then they had some transition problem, and I think, and then there was, like, a bunch of issues with, like... And then people just, like, lumped it as SaaS, and, like, the stock got crushed, but they're so good at making money. And, like, I just think that's the reality. Like, I invested in, like, the Yahoo buyout, which has been, like, an incredible deal, right, financially, that Apollo did. And so, like, I just think there's all these ways. Like, these are big, old internet things that will continue to exist.

— Sam Lessin, More or Less Ep 161

But guys, it's also a people thing. It's like Howie wants to work on AI, 'cause AI is cool. You're saddled with a thing that has hundreds of millions of dollars in revenue, but is clearly, like, a SaaS business, and, like, they're like, "Let's fix it. Let's just clarify this. Call a spade a spade." It's painful, but you're not in a place you can IPO it. I don't know, you clear the decks and you let the people... This is so human driven, right? Like, what people wanna work on, what teams wanna work on. Bending Spoons will pick it up and make a lot of money on it, so God bless.

— Sam Lessin, More or Less Ep 161

There's no heart needed. You take Vimeo, you buy it. It's a great business if you operated it properly. You, like, buy it cheaply, you consolidate customer service, you make a shitload of money, right? It's awesome. I do think, like, SaaS is crushed. It's cool they've negotiated this out. I actually don't know the details. I haven't spoken to Howie about it at all. My sense is, is that Howie gets to go operate the AI company he actually wants to operate. And as a Bending Spoons shareholder, I'm very happy. Now, as an Airtable shareholder, the write down from where everyone had it marked, the big fancy people, is pretty brutal. And I'd also say that, like, the amount of capital it took them to get to this outcome is not my favorite, 'cause there's a lot of dilution, but, you know, it's something.

— Sam Lessin, More or Less Ep 161

So here's the thing I would buy. It's not the business model they're pitching, and certainly not what they do, but I could imagine it working this way. Everyone in the world uses open source. It's pretty good. It's like having your IQ like 100 audience, IQ 110. It does almost everything you need, super fucking cheaply, whatever. But every once in a while, there's a rebellion, right? And that rebellion fucks you, right? And so who do you call when you have a rebellion? You need to call in like the fucking Navy SEALs and stealth bombers, which are super expensive.

— Sam Lessin, More or Less Ep 159

No, and all that stuff. It's just, there's a v- this is like the French make great art, doesn't make it a great business. So like, that's kind of my take. Here's an interesting thought experiment. What if the business model for the frontier labs actually is not a- actually about being used, it's more as a weapon systems? So hear me out.

— Sam Lessin, More or Less Ep 159

There are very funny, um, like, Instagram memes about people talking to the voice mode being like... Or some guy's like, "ChatGPT, build me a billion-dollar business. Make no mistakes. Be quiet, be quiet, be quiet. Do this. Do this. Like, SaaS app," you know. And, like, so those will get better, I guess. Maybe that's an impact.

— Sam Lessin, More or Less Ep 158

Tweets