Andy

And reinforcement learning typically uses human determination of what's right and what's wrong.

But this, the agent-involver, has its own methodology to do that.

So...

It's all about RL in that respect, the agent evolver system, but it doesn't say specifically what kind of sort of technical stack is being used to achieve that.

So it is an RL system for sure.

Got you.

Gotcha.

Yeah, almost everything that comes out of China is open source, but I can't pull it out of my fading memory at the moment.

But just a couple of days ago, I remember reading that one of the things that is...

Catching fire that's a delivery from one of the Chinese major firms is not.

It's a closed system.

And that's sort of intimating that at some point, China has to figure out how to monetize all these things in the global marketplace.

And so they're going to have to move in some ways towards a closed kind of platform, expressly for monetization purposes.

Because otherwise, everything's out there.

free but remember that even though the an open weight model or a system that's open uh is available you have to have some technical chops to implement that it's not like uh you know you know somebody can avoid paying uh you know all the time because you know until and unless you're a developer yourself and you have a platform that can run these things

Another million-dollar idea just voiced by Jim Giddey there.

You know, million is like pocket change these days in the world of AI.

I'll take it.

Ministral.

Mistral and Ministral.