DanielFilan comments on Towards more cooperative AI safety strategies