Corey Grant

And you have to go hire a bunch of people or contract it out.

And there's just not enough data scientists in the world for everyone to start being a model builder.

Like it's just there's just not enough.

Like I can't go build a foundational model.

So when I say train, I'm talking like foundational model, you know, open AI meta Google level.

What you're kind of referring to is fine tuning.

And I'll give you a perfect example of something I did on a Pro Max.

And it's a stupid use case, but I think it's so cool because it highlights the fact of what you can do on a workstation.

As I was working with all my partners, can't really name them, but they were working with basically a gravel quarry.

And the way the gravel quarry operates is,

is they sell rock by size, right?

So it's three inch limestone or pebble or whatever.

And the way that their model worked with their contract with this other company was, Hey, we're going to buy a load of rock and it has to be within 10%.

I mean, it has to be 90% three inch rock approximately, or we get that load for free.

And if it is above 97, they pay the full thing of the load.

Right.

So very old school industry.

Right.

Like not a lot of AI in it.

Yeah.