General-purpose large language models outperform specialized clinical AI tools

General-purpose large language models outperformed specialized clinical AI tools on medical tasks like diagnosis and treatment recommendations, according to a study in Nature Medicine.