Corey Grant

Exactly.

And that's not hard.

I mean, you're, you're feeding in, you know, representative examples of that.

And you have very much a human in the loop to say yes, no, maybe so.

And you're ultimately training it out.

And no one is going to use or even know what I'm talking about.

But the very first machine learning that I use personally is,

was back it would have been 2009 there was this company at dell i was leading so weird social media listening back when like it wasn't a thing where we were going to like boardreader.com and looking for people complaining about dell while we worked with the company

which was called Radiant 6, it ultimately got bought by Salesforce, where they would scrape every social media platform.

So it was much like Google, but it would scrape social media to be able to then query against, you know, like, hey, someone said Dell, we could respond to them or whatever.

But there was another tool called Crimson Hexagon where the biggest issue we always had with Radiant 6 was sentiment.

It would be like, what is positive?

What is negative?

And there was no algorithmic thing to it.

It was like, here's the words that are bad.

Here's the words that are good.

And if they use good or bad, we know if it's going to be good or bad, right?

Well, someone would say, man, that's a bad PC.

That thing's cooking.

Probably positive.