Ruby comments on Why long-term AI safety is not an issue (IMHO)