Azeem Azhar

What does that layer actually look like?

What you're describing is a fundamental, almost a financial market for this information between the providers and the AI companies.

If you want to support us in bringing more of these conversations to the world, please consider subscribing to the show.

I know that there's a lot we don't know, and we're going to discover this over the coming years.

But I think it's fun to explore where it might end up.

Of course, these aren't commodities.

When I'm buying a barrel of oil or you're buying a barrel of oil,

we already know what we're going to get and we know what it might be worth for us downstream.

The trouble with an information product is I don't know.

And by the time I do know, I can renege, I can repudiate, I can say it wasn't worth what I thought it was worth.

So you almost need a mechanism that

effectively can rate the likely quality of the content and its potential value.

And then you have to allow those ratings to be themselves rated.

So there's got to be a second layer of trust that gets built on top of that.

One of the things that would do is it would get around the problem that the Spotify or licensing model runs into.

One of the problems that model runs into, and I'll just play it back, is that, you know, you look at the pool of revenue that's generated and you figure out how to allocate it to each person who's done a piece of research or written an article and so on.

that allocation formula needs to be something that people agree on.

And it also needs to be auditable.

And so OpenAI might say, well, I know I took a thousand of your things, Azeem, but actually they're only worth two quarters.

Whereas I took one of Matthew's, who's a mate of ours because he delivers our service, and his tweet was worth a million dollars.