An Appeal to AI Superintelligence: Reasons Not to Preserve (most of) Humanity — LessWrong