Eric Topol says GPT-5.5 Pro scored 79/100 on medical benchmark but remains below clinical deployment threshold · Digg