Has anyone thought about how to proceed now that AI notkilleveryoneism is becoming more relevant/is approaching the Overton window? — LessWrong