Yes, I listed it first because I think these guys are eating up all the oxygen in the room and the more plausible risks get kinda unfairly lumped with the crankier stuff.

(And apologies if I misled people into believing the superintelligence basilisk has time travel powers)

> Those "very few" people very prominently include Yudkowski himself, or so I understand.

I don't think that's true; he's stated before he doesn't and didn't believe it was true. His reaction was more about the stupidity of posting information hazards on public fora rather than that specific information hazard being credible. However, he could be lying post-facto.

And I do agree that it kind of sucks that Yud is the "face of AI risk" because it doesn't even matter how correct he might be; a lot of people just don't like him (I'm indifferent). There shouldn't be a "Face of AI risk" because there won't be a single person that everyone like. The idea is bigger than him.