Chamath Palihapitiya

You're going to be working with NVIDIA to build models, to open source them.

What is the big announcement here?

And Mistral, obviously, based in France, you're the leading AI company there.

What's it like running the company and building a large language model in Europe?

Obviously, there's regulations and all kinds of considerations.

Privacy, the French are known for protecting privacy.

In the United States, we're known for taking it away.

How is the landscape there and what do you have to deal with there that maybe you wouldn't have to deal with in America?

What's the pros and the cons?

And so your belief is specialized verticalized models, healthcare, finance, engineering, different verticals will win the day or a global model will win the day that does everything?

This training data, using experts to come in and refine a model, most people don't know this business that well, but this has become a very large part of the industry.

Obviously, Scale AI was doing it.

They went to Facebook, lost a lot of the customer base who didn't want to send their data, I guess, over to Meta.

We're investors in a company called Micro One that's doing pretty well in this space.

There's other folks doing it.

Explain to the audience...

what you're doing specifically for companies and how this training works in a verticalized way, and then how you silo that data.

Because if you're working with one customer in aerospace or FinTech, they might have a need set, but they may not want that training to go to a competitor.

Yeah, this seems to be once the entire open web, what was available legally, gray market, et cetera.

I wouldn't have you comment on that controversy, but we kind of exhausted what's in the open crawl, yeah?