How Do Multimodal Large Language Models Perform on Clinical Vignette Questions? | JAMA Author Interviews | Podwise