> when asked by authorities

Key point right here.

You let people post what they will, and if the authorities get involved, cooperate with them. HN should not be preemptively monitoring all comments and making corporate moralistic judgments on what you wrote and censoring people who mention Mickey Mouse or post song lyrics or talk about hotwiring a car.

Why shouldn't OpenAI do the same?