Nick Heiner

Yeah, so this is a big open research question.

By default, the models are collapsed.

They will say very consistent things.

For instance, Claude loves Marcus Chen.

I don't know who Marcus Chen is, but if you ask Claude to, for instance, generate a thousand fake resumes...

And you ask it to do it sort of in sub-agents.

So each Claude does not know what the other resumes are.

You're going to get 700 Marcus Chen resumes.

Yeah.

Yeah, same.

Now, of course, the CEO of our company is named Edwin Chen.

So, you know, no conspiracy theories there, but...

Maybe, maybe.

Right, exactly.

But yeah, so that is sort of the challenge is like the model gets some idea in its head that like the archetypical person is named Marcus Chen and like the second most likely person to exist is named Priya Patel.

And yeah, it is hard to like get them out of that.

And like right now you can do it with prompting and sort of other techniques.

But yeah, I think sort of big open research question to see like, I mean, this is what temperature is for, right?

Like the temperature inference setting.

But of course you crank that up too much and it's going to stop following instructions or get zany in other ways.