Julian Nalenz comments on The “low-hanging fruits” of AI safety