We stress tested many frontier AI models for multimodal medical reasoning (including GPT-5, Claude 3.5, Gemini 2.5 Pro). They’re not ready. Faulty reasoning, use of inappropriate shortcuts, hallucinations. Published today @NatureMedicine nature.com/articles/s4159…

@EricTopol @NatureMedicine Those are not frontier models

@SandelloRed @NatureMedicine Yes they are

@EricTopol @SandelloRed @NatureMedicine FALSE. Just like you misled the public about the safety and efficacy of the Astrazeneca vaccine and helped prevent it getting EUA'd in the US in early 2021, at a critical time when we had millions of doses stockpiled and people were desperate for vaccines.