Out now in npj Health Systems- New article from K. Narang and team highlights a few important considerations when asking LLMs to prioritize kids in the ED
nature.com/articles/s44401-0…
Newton retweeted
Excited to launch the Health AI Innovators Speaker Series at UC Berkeley–UCSF!
First up: Cameron Chen (Google), co-lead of Project AMIE, on AI for clinical reasoning in healthcare.
Dec 4, 2–3 PM PT | Berkeley + Zoom
Join us: luma.com/5dp9yd5v
Applications for re:AGENT II close Dec 11. Teams of scientists and builders will spend January 9–10 in SF building an AI Scientist, an AI Clinical Developer, or AI for Patient Access. Each track has a cash prize, including a separate Eval Prize!
Apply now. The link is in the comments.
@AnthropicAI , @modal , @benchling , @biohub , @strandaibio, @benchflow_ai, @FenwickWest , @arcinstitute, futurebio
Newton retweeted
We’re hiring postdocs at @StanfordAILab! Join the Leskovec Lab, in collaboration with the Good Lab, to build foundation models and AI agents for biology.
Help turn multimodal biomedical data into scientific discoveries.
Apply: snap.stanford.edu/apply/inde…
An investigation by Foster et al. into the risks of ASQ-positive children in the emergency department sciencedirect.com/science/ar…
8
Newton retweeted
Embedding psychiatric nurse practitioners and specialized support staff in the ED was associated with shorter stays for psychiatric emergencies, especially for patients on involuntary holds.
#EmergencyMedicine #MentalHealth #AnnalsEM
Newton retweeted
1/ Can AI help doctors make more informed decisions based on patient activity and behavior over the course of a hospital stay?
Today we introduce TRACER: a training-free multimodal harness that enables LLMs and VLMs to track patient activity using bedside video and health records.
Blogpost: ai.stanford.edu/blog/tracer/
Paper: arxiv.org/abs/2610.21266
Medical AI is built for adults, and newborns are left behind.
Out today in Nature Biomedical Engineering (@natBME): we took an AI model built to check breathing tube placement on adult X-rays and adapted it for newborns by sending it to 30 NICUs in 20 countries and territories. Learning from just 50 X-rays at each, it pinpointed the tube tip and the base of the windpipe more accurately than hospitals' own models at 29 of 30 sites, even at hospitals it had never trained on.
Why it matters: only 17% of FDA-authorized AI medical devices are labeled for use in children, largely because pediatric data is scarce and hard to share. Out of the box, the adult model misplaced the tube tip on newborn X-rays by more than 5 cm, about the length of a newborn's windpipe, in 37% of cases.
So instead of pooling data, the model travels: it's fine-tuned on 50 X-rays at one hospital, then passed to the next. We simulated this using real X-rays from 30 NICUs, an approach designed so patient data never has to leave a hospital.
Across 30 NICUs:
• Large errors (over 5 cm) fell from 37% to 2.6%
• Better at flagging dangerously deep breathing tubes at 27 of 30 sites
• Gains grew with every hospital added, with no plateau
It's been surreal working with such a diverse global team. My goal is closing the gap between AI development and real impact for patients. This is one step: it removes a major feasibility barrier, scarce and siloed pediatric data, and could help open the door to a new generation of pediatric AI.
Paper, code and figures below 👇
Grateful to @pranavrajpurkar, @emma_ychen and our team of 75 researchers across 20 countries and territories (full team below)
@harvardmed @PennMedicine @BrighamWomens @StanfordMed @GreatOrmondSt @KU_Leuven @UAlberta @UniklinikEssen @TelAvivUni @OsakaMetUniv_en @AUB_Lebanon @AUBMC_Official @SidraMedicine @ChildrensColo @SLU_Official @Newark_Beth @MedUnivSC
Newton retweeted
BodyMaps AI is now live. We're recruiting volunteers across radiology, radiation oncology, and surgical departments for a reader study on how AI assists early cancer detection, tumor contouring, and surgery/radiotherapy planning. Try it here: bodymaps.wse.jhu.edu/ — DM me to join.
Volunteers would earn authorship in our clinical papers.
@JohnsHopkins @HopkinsMedicine
Newton retweeted
Come be my colleague! We're hiring FOUR professors for computational biomedicine. I truly believe UC Berkeley is one of the best places to do interdisciplinary. As a bonus, the eucalyptus grove is a great places to take walks.
aprecruit.berkeley.edu/JPF05…
I am delighted to announce four new faculty positions in Computational Biomedicine at UC Berkeley: two at the non-tenured level and two open-rank positions (tenured/non-tenured).
aprecruit.berkeley.edu/JPF05…
Please repost & share the info with outstanding candidates in your networks
Great to see the surgical AI collaboration Dr. Amin Madani and I began at UHN two years ago, during my time as Chief AI Scientist, featured in the New York Times (@nytimes ).
For me, this connects to a bigger research question behind our work on MedSAM: how do we turn medical images into useful clinical information?
With MedSAM, our group tackled a foundational piece: outlining organs, lesions, and other structures across medical images.
Surgical AI takes that ambition into a different setting: helping surgeons interpret what they see as an operation unfolds.
The NYT story features GoNoGoNet, which highlights predicted “go” and “no-go” regions during surgery.
Green and red.
Then surgeon Timothy Jackson suggested a third color:
Yellow.
That captures the next research challenge beautifully.
Delineating a structure, recommending an action, and recognizing when there isn’t enough evidence to recommend one are distinct capabilities.
We’re making progress on teaching AI to see. The deeper challenge is making its guidance worthy of trust.
Read the NYT story: nytimes.com/2026/10/07/world…
MedSAM: nature.com/articles/s41467-0…
MedSAM2: arxiv.org/abs/2504.03600
Out now in JACEP Open: Devisetti et al evaluate whether LLMs can perform as well as clinicians at calculating clinical decision rules from patient histories. jacepopen.com/article/S2688-…
15
Newton retweeted
Rerunning our newly accepted NeurIPS paper with the latest agents, we've found 3 significant improvements to fitting interpretable models, available now through imodels: csinva.io/imodels/blog.html
NEW paper from Microsoft Research.
(bookmark it)
The entire interpretability literature is built around human readers. As more analysis gets delegated to agents, the right target of interpretability shifts. This paper is a recipe for designing tools that agents can actually reason about.
They introduce Agentic-imodels, an autoresearch loop where a coding agent (Claude Code, Codex) iteratively evolves scikit-learn-compatible regressors that are simultaneously accurate AND readable by other LLMs.
Interpretability is measured by whether a small LLM can simulate the fitted model's behavior just by reading its string representation. Predictions, feature effects, counterfactuals, all from the __str__ output alone.
Run on 65 tabular datasets, the discovered models push the Pareto frontier past every classical interpretable baseline (decision trees, GAMs, sparse linear), and improve four downstream agentic data science systems on the BLADE benchmark by 8% to 73%.
Paper: arxiv.org/abs/2605.03808
Learn to build effective AI agents in our academy: academy.dair.ai/
Newton retweeted
I built an interactive echocardiography tutorial with Claude Code.
Chapter 07: a 3D heart, the standard imaging views, the probe position, and simulated ultrasound video.
Newton retweeted
Introducing Synthetic Hospital: an open, fully synthetic longitudinal EHR benchmark with verifiable ground truth!
1,268 patients, 5,602 encounters, zero PHI. Physicians could not reliably distinguish its charts from real ones.
📄 arxiv.org/abs/2609.30027
💻 github.com/sparkcpark/synthe…
✍️ sparkcpark.github.io/posts/f…
Newton retweeted
Thrilled that #paper2agent is published in @nature today!
Scientific knowledge is traditionally stored in passive papers. Paper2Agent transforms papers into virtual authors that answer questions, apply its methods and collaborate w/ other paper agents to make new discoveries 🧵
Newton retweeted
At the OpenAI Foundation, we believe breakthroughs against disease will come from pairing intelligence with more observations of the world – in other words, more data.
Today, we're announcing over $125M in grants, launching our second science program:
openaifoundation.org/news/pu…
Newton retweeted
Verified U.S. clinicians can get free GPT-6 Astra (Pro) access today via ChatGPT for Clinicians.
Sign up: chatgpt.com/plans/clinicians…
Newton retweeted
Replying to @Stanford
@Stanford Science Fellows 2027 — $98,000 Postdoctoral Fellowship, designed around foundational research in the natural sciences, rather than a narrowly defined project or employer-assigned research programme.
Details at stanfordsciencefellows.stanf…
Apply by 16th Oct 2026
Newton retweeted
🔴 CAREER UPDATE / I AM HIRING 🔴
After several months of being extremely annoying to everyone who knows me, I have big news: this fall I'm joining Brown University as Assistant Professor of Biostatistics, Data Science, and Computer Science!
Newton retweeted
I'm hiring! 🚀
Our Frontier Health team at @GoogleDeepMind is looking for a Research Scientist to build foundational AI & physiological world models for human biology.
Help us transform healthcare from reactive observation to proactive intervention. 🧬
🔗 Apply: goo.gle/3SPNis2































