Eye On A.I. #317 Steven Brown: Why Modern Medicine Needs AI-Assisted Decision Making
11 snips
Jan 25, 2026 Steve Brown, founder of CureWise and Stanford-trained physicist/computer scientist who built multi-agent AI after his own rare cancer journey. He discusses agentic AI analyzing messy medical records, multi-agent consensus to surface missed tests, large context windows and RAG, and using AI to educate and empower patients for better clinician collaboration.
AI Snips
Chapters
Transcript
Episode notes
AI Agents Revealed Missed Tests
- Steve used multiple AI agents to analyze his past medical records and modeled agents after his specialists.
- Agents converged on recommending specific tests (free light chains, possible bone marrow biopsy) that earlier doctors hadn't ordered.
Advocate For Yourself With Doctors
- Become an expert in your condition so you can ask the right questions during short doctor visits.
- Advocate for yourself because constrained clinician time often means care follows standard checklists.
Use Context Windows Over Fine-Tuning
- Large context windows let models use hundreds of pages of personal medical data as prompt context without fine-tuning.
- Organizing relevant recent records into that context is more effective than blanket model training.