Max Harms comments on Corrigibility should be an AI’s Only Goal