hazelzhang comments on No Safe AI and Creating Optionality