He chose to publish as soon as he had a lean proof and before editing the written proof because he knows openAI could publish a proof any minute
He has all the time in the world to edit it down now