“The medical profession is unconsciously irritated by lay knowledge.”
- John Steinbeck in East of Eden
(Loved this book btw)
- John Steinbeck in East of Eden
(Loved this book btw)
@jonreismanMD
Peer review for journal articles has always been a highly imperfect process—but AI is pushing this already-flawed system to its breaking point, Anna Wexler argues. She explores how this technology may be ruining academic publishing: theatlantic.com/ideas/2026/10/…View quoted post on X
INSANE. Microsoft has published peer-reviewed research in Science showing its AI diagnostic system outperforming physicians across two medical benchmarks.View quoted post on X
The results:
- Emergency-department cases: 80% accuracy for AI versus 47% for doctors.
- Complex NEJM cases: 83% versus 36%, even when doctors could use reference materials.
The system uses OpenAI's GPT-5 to simulate a panel of physicians that debates diagnoses, orders tests, and weighs costs. Microsoft says performance improved further with GPT-5.5 without changing the orchestration.
These were simulations using historical cases, not actual patient care.
Microsoft is now partnering with Beth Israel Deaconess Medical Center for a prospective study in which patients will interact directly with the AI.
Microsoft AI CEO Mustafa Suleyman told The Verge he expects "extremely impressive" production deployments of intelligent systems in 2027 and predicts medical AI capabilities indistinguishable from human performance within the next couple of years.