jezm comments on The Mirror Problem in AI: Why Language Models Say Whatever You Want