Lex Fridman: belief

Transcript context

…And there are, like Nathan said, also two layers to it. Someone might buy the book and then train on it, which could be argued fair or not fair, but then there are the straight-up companies who use pirated books where they’re not even compensating the author. That is, I think, where people got a bit angry about it specifically, I would say. Yeah, but there has to be some kind of compensation scheme. This is like moving towards something like Spotify streaming did originally for music. You know, what does that compensation look like? You have to define those kinds of models. You have to think through all of that. One other thing I think people are generally curious about, I’d love to get your thoughts: as LLMs are used more and more, if you look at even arXiv or GitHub, more and more of the data is generated by LLMs. What do you do in that kind of world? How big of a problem is that? The largest problem is the infrastructure and systems, but from an AI point of view, it’s kind of inevitable.…

Stored transcript either side of the excerpt. The highlighted words are the published quote; the surrounding text is unedited source, never generated.