The AI Product Going Viral With Doctors: OpenEvidence, with CEO Daniel Nadler

OpenEvidence CEO Daniel Nadler explains how a free, doctor-direct app reached 10-25% of U.S. physicians by training small specialized models on peer-reviewed literature instead of the public internet.

• OpenEvidence reached 100,000+ monthly active U.S. physician users within roughly one year, growing almost entirely through word-of-mouth.
• Medical knowledge doubles every 5 years by conservative internal estimates; doctors cannot keep pace through medical school training alone.
• New England Journal of Medicine approached OpenEvidence after senior editorial board members became power users — other well-funded AI companies had been rejected after offering large payments.
• OpenEvidence uses an ensemble of smaller specialized models trained exclusively on peer-reviewed literature, with no public internet connection, outperforming large general models on in-domain medical tasks — a finding published as best paper at the leading ML-in-healthcare conference in 2023.
• The platform is free to all physicians, including community doctors in rural and underserved areas who cannot afford enterprise SaaS pricing, and is used at Walter Reed without a government procurement process.
• Nadler estimates OpenEvidence could reach 1 million lives saved by around November 2034, based on 300,000–800,000 annual U.S. deaths from medical errors.
• Nadler defines high IQ as neuroplasticity — the speed of acquiring new skills — and credits this trait, not pedigree, as the common factor in elite team members from Harvard and MIT labs.
• Broad foundation models are commoditizing at the application layer; Nadler argues the high-value companies of the next decade will be domain-specific application builders, not foundation model providers.

2025-03-04 · Watch on YouTube