Shmi comments on [SEQ RERUN] Above-Average AI Scientists