Karl Yeh

Same thing Yian has his own thing.

But I don't know if his is less superintelligence, more world models.

Old models, yeah.

Yeah, there is one thing.

So maybe this is going to finish the discussion.

So I wrote a post about this last, like yesterday.

So last week was the very first time in our discovery session for a client that we had one department, very anti-AI, like all five.

The manager was anti.

One of the employees was anti.

The other two were kind of just following along, but they're all kind of anti.

And most of the time, what we found is people don't trust it, but it's not like very anti, right?

Like it's not like,

And, you know, the very first time somebody pulled the report, we're like, hey, from MIT, you know, critical thinking decline.

I don't know where they're pulling the 40 percent hallucination rate because I was like, what model, what prompt, what task?

How did you measure?

Like, was it grounded in statistics?

Was it grounded in data or was it just pulling random?

Like, was it free model?

Would you pay for it?

Like, there's too many variables there to be like, right.