Is it time to talk about AI doomsday prepping yet? — LessWrong