Howie_Lempel comments on On Deference and Yudkowsky’s AI Risk Estimates