habryka comments on Six Thoughts on AI Safety