Roman Leventov comments on Steven Wolfram on AI Alignment