Surgical procedures led to a 16% increase in financial hardship and an 8-fold rise in out-of-pocket spending for US working-age adults, especially the uninsured and those with private insurance. ja.ma/4oLKuH4

Releasing a new "Agentic Reviewer" for research papers. I started coding this as a weekend project, and Yixing Jiang made it much better. I was inspired by a student who had a paper rejected 6 times over 3 years. Their feedback loop -- waiting ~6 months for feedback each time -- was

✍ Busy publication day - two important AI and healthcare articles out today that I want to draw your attention to: 1. This one, the second in a series in the The BMJ discusses "How AI affects patient agency" - it's a patient-focused take on the advances we're seeing in

AI makes student feedback radically better if you structure the prompt. In an experiment (N=129), students who used AI to give feedback to their peers: • Gave more specific praise • Identified clearer gaps • Offered 3x more actionable next steps

The sad truth is that very few folks want to do peer reviews. Good reviews are extremely valuable but many are pedestrian. The whole process takes way too long. So glad to be a part of a journal that is fundamentally rethinking this whole process! NEJM AI

Huge fan of multi agent systems, agent based modelling, and social intelligence - these frames still seem really absent from mainstream AI discourse except in a few odd places. Some half-baked thoughts: 1. Expecting a model to do all the work, solve everything, come up with new

Now I would add that professionalism demands the use of assistive artificial intelligence, because only then can you best know what you are doing or the effect of what you do. Raj Mehta Ethan Mollick Adam Rodman Sylvie Delacroix

Remarkable story of high school team who used CRISPR to make a rapid diagnostic test strip for Lyme disease, speeding dx from 2 weeks to days

x.com/i/article/1996…

AI's clinical safety is not the same as AI accuracy. NOHARM shows LLMs can score well on knowledge tests yet still give harmful clinical advice, mostly by leaving out key actions. If you evaluate AI in medicine, measure safety as its own metric. #MedEd

Summarizing 2025 in AI in a tweet 1) No sign of a slowdown in exponential pace of gains 2) Jaggedness remains the main issue of AI 3) Early days for deployment, but many companies reporting positive ROI 4) GenAI became an industry, with industry-level impacts 5) AI is still weird

First, do NOHARM: towards clinically safe large language models arxiv.org/abs/2512.01241 & bench.arise-ai.org via David Wu, MD, PhD et al Richard Lehman Raj Mehta

I have read this article a couple of times and, for those that care about primary care access for rural communities, I urge you to read this analysis which was published in the Annals of Family Medicine. Here is what the researchers found: there were 11,847 rural FPs in 2017 and 10,544 in

Many European countries skip universal hepatitis B birth doses, giving only to high-risk infants & starting routine vax at 6-8 weeks. Why do U.S. Hep B birth dose recs differ from Europe? 🤔 I’m glad you asked! 🧵👇🏽 1/4

We love to say "costs fall X% every time capacity doubles." This paper looks at 87 technologies and says: actually, no. Past learning rates are not reliable predictors of future learning.

29%—That's a lot of weight loss, folks Triple GLP-1 drug receptor retatrutide Compares with avg weight loss with Zepbound 21%, Wegovy (Ozempic) 15% nytimes.com/2025/12/11/hea…

Something I told 13 yo: Many if not most philosophical controversies are artifacts of trying to using natural language at a higher resolution than it will support.

Bombshell: Oliver Sacks (a humane man & a fine essayist) made up many of the details in his famous case studies, deluding neuroscientists, psychologists, & general readers for decades. The man who mistook his wife for a hat? The autistic twins who generated multi-digit prime

The original “doctors will be replaced with machines” paper was written in … 1979 Only the machines were linear regressions. And they still outperformed clinical intuition.

Artificial Intelligence in Detecting Statistical Errors: Implications for Authors, Reviewers, and Editors jkms.org/DOIx.php?id=10… via Olena Zimba et al Richard Lehman Raj Mehta