@robinhanson I asked ChatGPT, it gave multiple answers, but I selected this one as best: "A human retrieves one salient answer and discards the rest; an LLM activates many partially relevant associations at once and tends to express several of them."
@DavidPinsof plausible to me that those traits are correlated with avoiding a small set of specific kinds of bad ideas, but not much more than that--that is, they seem very far from anything like necessary or sufficient for good ideas
@default_friend Ok! I came SO CLOSE to raising my hand during @Aella_Girl 'a final rebuttal to ask whether there were any identical twin studies on fetishes, had to tell myself, "BEHAVE!"--was sorry to have to run off at the end but had to be home for kids' bedtime
@RizomaSchool yup, they are open to the public, our 1st two for the upcoming academic year are 9/29, Bladerunner at Doc films 6pm, & 8pm 10/30 @LedermanHarvey on ChatGPT & the meaning of life--I'll announce the events here on twitter
@robinhanson ah, that was not clear from your post!
I thought you were just making the descriptive point, and I suspect many other readers thought the same.