Bill Gurley

Many of the players in the market are reselling a wrapper on top of a foundation model, on top of a hosting service.

And many of them, I think people believe, are at negative gross margin.

So you might, in buying something from a wrapper company, be getting compute

Cheaper than you would have got it from the model companies, getting it cheaper from the hosting company.

And that revenue is being counted three or four times with negative gross margin until we get to a point where unit economics matter.

And it can't matter in an all out war market that the gavage tube funding creates.

You have to go for market share.

You have to.

I think that window is in front of us in terms of how that settles out.

But I have no doubt, even if you want to step away from the foundational models, like the work that Brett Taylor is doing at Sierra, I have no doubt that's real success.

and will impact every one of those companies that he touches or they touch and will change those companies materially.

I just don't have any doubt of that whatsoever.

I guess that's a long answer of saying, I think so many of these things are a rational reaction to what's happening.

My answer to that question is decidedly personal, and it relates to what you just said.

I'm probably doing 40 or 50 searches a day on AI platforms, which is more than I ever did Google searches.

It's almost all...

a form of very quick learning, super quick learning about either particulars I forgot, things I don't know about, and it's every day.

And I think to myself, for those people that are inherently self-learners, the speed at which they'll be able to get things accomplished and move up the ladder is breathtaking.

And then I think outside of LLMs, from Tesla FSD to other types of problems that are being solved with traditional AI, those are super interesting to me as well, maybe more profound.

I do worry that LLMs have a limitation.