ESRogs comments on Drexler on AI Risk