Sam Clarke comments on On presenting the case for AI risk