Will Radiologists Build Their Own AI? | Dr. Peter Chang

In this special SIIM-CAIMI series episode of Scanning the Market, host Jay Gurney sits down with Dr. Peter Chang—radiologist, software engineer, co-founder of the Center for Artificial Intelligence in Diagnostic Medicine (CAIDM) at UC Irvine, and CAIMI co-chair. Dr. Chang breaks down a startling shift in medical software development: why 99% of his code is now AI-generated and what happens to healthcare innovation when the barrier to building complex software collapses. Moving past the mechanics of coding syntax, this discussion examines why narrow, expensive point solutions will face margin compression, how the value chain is pivoting toward orchestration and native workflows, and why clinical governance must evolve to evaluate non-deterministic, agentic systems. Key Topics Covered in This Episode: -99% AI-Generated Code: How coding agents shifted from toy tools to production-grade accelerators capable of building medical vision models over a weekend. -The Crucial 1%: Why human engineering principles, architectural steering, and domain intuition matter more than syntax and prompt engineering. -Democratizing Medical Software: What happens when clinicians can pull open-source models, tweak repositories, and solve niche clinical problems directly. -The Death of Overpriced Point Solutions: Why charging hundreds of thousands of dollars for narrow, single-finding models is becoming economically unsustainable. -Shifting Value to Orchestration: Why the algorithm itself is becoming a commodity while routing, data flow, inference scaling, and PACS/reporting integration capture value. -Dynamic vs. Deterministic Governance: The friction between static FDA clearances and real-world clinical software that auto-heals and updates continuously. -Evaluating the Human-AI Team: Moving beyond isolated model benchmarks to evaluate how specific clinicians perform when paired with AI assistants. -CAIMI 2026 Preview: Exploring the AI Builder Showcase and creative agentic workflows in Philadelphia.