tidikanji comments on Naive Hypotheses on AI Alignment