Corey Grant

Right.

But we use the bad, bad word.

It was negative.

Anyways.

So I'll tell you is that Crimson X gone the way that they did it.

And this is much like AI.

And I use this example all the time is it would show you an example and say, Hey, this PC is, you know, bad, man, it's cooking.

It's great.

Or whatever you would then say, is this positive?

Yeah.

Is this neutral?

Is this negative?

And you would tune about 20, 30 examples, and then it would apply kind of machine learning across that to be able to say, hey, here is what we believe based on what you said is this.

And that's really fundamentally what AI is.

It is you basically having someone, data science in the background, basically say, this is what this is.

And then you're looking at the output of that and you're kind of validating.

It's kind of the same thing when you use Copilot.

It says, was this correct?

Was this a good generation?

Every time you say that, that's just kind of a feedback in the loop to your question that ultimately goes back to the data scientists to make AI better.