John G. Halstead comments on A list of good heuristics that the case for AI X-risk fails