Doesn't strike me as the kind of principle OpenAI is willing to slow themselves down for, to be honest.