Hackathon results suggest AI agents could make verifying research far more routine

Read original article