SoerenMind comments on Our forthcoming AI Safety book