A (small) critique of total utilitarianism — AI Alignment Forum