michael_nielsen

Yeah, I do think that "psyching yourself up" thing is an interesting edge case, kind of like suspension of disbelief to enjoy a movie...

I know lots of ppl who do this well, & it might be crucial if our motivational systems can't run on EV alone! But it feels like an edge case

i.e., temporary irrationality over the question, "How promising is my current project / line of inquiry?" feels like an edge case

It's v. important in lots of science. In that example, Feynman is talking about believing something false for years, as his "edge"

Huh maybe we're just talking about different things: "innovation" is one project and "evaluation" is another, and I'm thinking of the latter

I'm not sure. I think the examples you give, perhaps AI risk especially, are ones where we're still very ignorant.

And I claim there are many impactful decisions made on the basis of similarly inaccurate models