Sodium comments on (Non-deceptive) Suboptimality Alignment