I had the privilege of hosting Nigam Shah, Professor of Medicine and Biomedical Data Science and Chief Data Scientist for Stanford Healthcare, for a conversation that traced an improbable arc — from an orthopedic surgeon redirected toward doctoral study the very year the Human Genome Project seized the world's attention, through two decades at the frontier of computational medicine. He guided me through Stanford's quietly foundational role in this field, dating to the first AI supercomputer deployed on a medical campus in 1984, and through his own contributions: rendering the chaos of unstructured electronic health records legible for bedside decision-making, and in 2018, launching what became one of the earliest real-time AI consultation services — a lineage he traced back, with characteristic precision, to a 1972 experiment at the New Haven VA. Perhaps most illuminating was his framework for thinking about technological change in medicine: an analogy to the evolution of money, where the deeper transaction — human beings caring for one another — remains fixed even as the mechanics accelerate around it. That distinction, he argued, separates genuine augmentation of care from mere digitization of the status quo. What lingered longest after we spoke was his unsparing clarity on incentives. He didn't equivocate in observing that the preponderance of AI deployed across American healthcare today serves the machinery of billing and prior authorization rather than the patients it ostensibly exists to serve — and that no technological sophistication can compensate for incentives left unexamined. We explored ChatEHR, the tool his team has embedded into clinical workflow at Stanford and which now serves roughly 3,000 physicians daily, alongside MedHELM, his rigorous framework for evaluating medical language models against the realities of clinical practice rather than the artifice of exam questions, and the constellation of companies he's founded in the space between insight and implementation. I closed, as I always do, with the phrase that anchors this show — Ichigo ichie, "one time, one meeting" — a fitting note for a dialogue this rich, unlikely ever to be repeated in quite the same way. About Nigam Shah, MBBS, PhD, is Professor of Medicine (Biomedical Informatics) and of Biomedical Data Science at Stanford University, where he also serves as Chief Data Scientist for Stanford Health Care and Associate Dean of the School of Medicine. His research applies machine learning, knowledge representation, and artificial intelligence to analyze multiple types of health data, including electronic health records, claims, wearables, and patient-generated web content, in order to understand disease and improve clinical care. He earned an MBBS from Baroda Medical College in India in 1999, a PhD in Integrative Biosciences from Pennsylvania State University in 2005, and completed postdoctoral training at Stanford in 2007, joining the Stanford faculty in 2011. Over his career, Shah has authored more than 350 scientific articles and has been recognized with numerous honors, including the 2012 Stanford School of Medicine Faculty Award for Outstanding Teaching, the 2013 AMIA New Investigator Award, and the 2016 Department of Medicine Divisional Teaching Award. He is also an inventor on eight patents and patent applications and has co-founded three companies. He was elected to the American College of Medical Informatics in 2015 and inducted into the American Society for Clinical Investigation in 2016, and more recently, was inducted into the Association of American Physicians in 2026. His work focuses in particular on making machine learning models clinically useful and on bringing AI into medical practice safely, ethically, and cost-effectively Support the show
