TrE comments on Self-assessment in expert AI predictions