Michael Roe comments on What actual bad outcome has “ethics-based” RLHF AI Alignment already prevented?