aog comments on Drexler on AI Risk