ChristianKl comments on Why do misalignment risks increase as AIs get more capable?