today we were talking about the argument that talking about AI risk publicly was actually a really bad move, because various corporations and governments’ response to being told about AI risk was “hey superintelligent AI? that sounds cool we should try to build one!”

and now I’m wondering if AGCbtIitSH?! is actually part of Eliezer’s brilliant plan to discredit himself enough that everyone will stop taking the idea of superintelligent AI seriously

#we'll see if Cthulhu gives them EA grandchildren#r a t i o n a l i t y