OpenEvidence CEO Urges Medical Students to Learn Classically
Daniel Nadler advised students to take paper exams and train without relying on AI, citing the risk of cognitive offloading.
Updated on Nov. 17, 2026 in Artificial Intelligence

Live Poll
Do you think people should still learn skills the hard way, without AI?
OpenEvidence CEO Daniel Nadler advised medical students to learn classically and take paper exams despite the growing use of AI. Speaking at Semafor's Silicon Valley & The World event on Monday, November 16, 2026, he said the risk of cognitive offloading is acute, particularly for younger generations who have grown up with AI access.
Why it matters
Early studies have found that deferring work to AI can weaken neural networks and lower cognitive performance, a concern for future clinicians who will rely on AI tools. The advice comes from the head of a medical AI company valued at $15 billion, underscoring the tension between AI adoption and foundational medical training.
OpenEvidence, a medical AI startup valued at $15 billion, synthesizes medical information using AI trained on millions of medical papers. Early studies cited by the company link deferring work to AI with weaker cognitive performance, though the sample sizes and methods are not detailed in the source.
The players
Daniel Nadler
Daniel Nadler is the CEO of OpenEvidence, a medical AI startup that synthesizes medical information using AI trained on millions of medical papers.
OpenEvidence
OpenEvidence is a medical AI startup valued at $15 billion that uses artificial intelligence to synthesize medical research for clinicians.
Semafor
Semafor is a news organization that hosts industry events, including its Silicon Valley & The World gathering where Nadler spoke.
The details
Nadler advised students to train classically, learn classically, and sit paper exams and tests, stating that the risk of cognitive offloading is particularly acute for younger generations. He noted that all of OpenEvidence's engineers use dozens and dozens of AI buddies, even as the company warns against over-reliance among learners.
Timeline
Monday, November 16, 2026: Nadler made the remarks at Semafor's Silicon Valley & The World event.
The Big Picture
Early studies showing that deferring work to AI can weaken neural networks and cognitive performance are now shaping how AI companies themselves advise medical training. The clash between AI adoption in clinical practice and the need for unassisted foundational learning is becoming a central question for medical education.
For medical students, the advice points toward practicing unassisted reasoning and paper-based testing even when AI tools are available. Clinicians and learners who depend on AI for daily information retrieval may need to deliberately maintain independent recall to avoid the cognitive performance declines the early studies describe.
The takeaway
Relying on AI for answers can erode the underlying skills it is meant to support, so learners benefit from building core knowledge through unassisted practice first. Treat AI as a tool to check and extend reasoning rather than a replacement for it.
Further reading
For more on how AI is reshaping clinical tools, see the Artificial Intelligence section.
Source note: This article includes information reported by Semafor.
Live Poll
Do you think people should still learn skills the hard way, without AI?










