Listen on:
Episode overview
Most healthcare AI doesn't fail on accuracy. It fails when it tries to leave the lab and enter a hospital.
Artem Trotsyuk knows that gap firsthand. He earned his PhD at Stanford building a smart bandage for wound care, then watched the path from paper to patient turn into a harder problem than the engineering itself. Now part of a governance group at Stanford Health Care, he evaluates AI tools before they reach clinicians, weighing financial impact, ethics, and workflow fit against what patients, developers, and doctors each actually want.
In this episode: why vibe-coding gets you 80% of the way to a healthcare app and stalls at the clinical door, how hospitals actually score a tool's accuracy and risk, why the evidence keeps favoring augmentation over replacement, and how community hospitals without big budgets can get access to the same rigor.