A large language model (LLM) matched or exceeded hundreds of expert physicians in diagnostic and management reasoning tasks across six experiments, a new study showed. The LLM's advantage was most ...