Samuels and Auren Hoffman were discussing companies that wind down failed businesses and sell off their internal data, and whether that could become a revenue stream for others. Samuels's company brokers real-world data for AI training.
…public. That's how they make money, yeah. Yeah, and like this selling the data. I mean, I actually, it's an interesting thing for like non-profits to think about where this is like material could be like a material revenue stream for them. Um, if they, if they're licensing some of this data,
oh, interesting, and yeah, that could also probably like there's a whole other set of ways you're interacting in a nonprofit that might be different than a for-profit or something like that.
Yeah, I mean, that is one of one of the jokes we have internally is the issue with buying Spearlin's Spirit Airlines data is like you're going to teach the models to be really good at operating a failed airline, right? Right, totally, right, right. So, you might uh run into the same thing if you're figuring out how to
charge the uh consumers for literally every single thing that they do, including a bag of peanuts or something. Yeah, perfect, perfect, just using our AI. Uh, like about 20 years ago, one of my friends was doing early version of AI, and the way he was training the data was the Enron emails. Um, I…
Totally. And that I think I've heard of other folks subsidizing. I mean, I think it sounds like he was on the earlier side of doing that, but I've heard of other folks doing that. Again, like it is this meta question of what are you ultimately, what are you teaching the models to ultimately do if…
No, I mean, in some ways, you are a store, right? You're going sourcing interesting goods and then you're selling those goods multiple times. And part of what you do is just being able to do procurement and sourcing it. And then, how does like a data store end up being very valuable? And it's, and…
