Corey Knowles

There was a great paper on that about a year and a half ago on whether we should quit calling it hallucinations and start calling it bullshitting, because that's what...

it's doing, it's doing the same thing we do when we don't know what we're talking about.

Like, you know, you'll be in a conversation and you're, you're there and you know, you're on your toes and you make these little guesses, you make these little guesses.

And, uh, uh, I've always thought that that's one of my favorite scientific papers ever, just, just because it's so funny, but it's, it's, it's also, it makes a lot of good points.

And, uh,

Yeah, it scored just like we are on a test where like, oh, well, you know, if I can leave it blank, but if I put anything there, I'm going to get one of those five points.

It's an interesting question and something I think may may help this over time.

I don't know if you saw this grant and you may actually have maybe been in the newsletter.

I apologize if it was.

Miramarati's company yesterday, Thinking Machines, has done this.

They're they're publishing some information and kind of finally disclosed what they're working on.

And what they're doing is trying to solve determinism in models.

The idea that if we're going to ask the same question the same way 50 times, we're going to get 50 at least slightly different answers.

And you don't want that for all occasions.

There are areas where that would be really bad.

But there are some of these things I think it will help with.

And, you know, at its core, we're still very early and we're at a time where companies are just absolutely dogpiling on where am I in AI?

What is my role?

What am I going to do?

Let's tack it on to every product we have and everything we touch and rebrand as an AI company.