PHARMAMEDIX retweeted
When a chatbot for health advice led to tragedy
Front page NYT today @teddyrosenbluth
gift link
nytimes.com/2026/10/27/well/…
video nytimes.com/video/well/10000…
PHARMAMEDIX retweeted
I really admire figures that can describe a concept better than long manuscripts.
This image is from an excellent paper in Nature by Vivek Subbiah about the future of evidence-based medicine and the design of clinical trials, all inspired by the rise of artificial intelligence.
This is the era of the next generation of 'deep' medicine.
"Current paradigms must be continuously challenged by emerging technology and by all stakeholders (the new generations of scientists, physicians, the pharma industry, regulatory authorities and, most importantly, patients). Disruptive innovation should lead to every clinical site being a research site, with all necessary quality checks and research as part of the standard of care.
My personal view is that every major disease needs a ‘Moonshot’ program and every rare disease should have an ‘Operation Warp Speed’—both with clearly identified, sustainable goals to improve population health and address equity, diversity and global access to therapies.
Methodological advances and future AI-based analyses of all data will provide deep evidence to realize the goal of personalized medicine— that is, to offer the right treatment to the right patient at the right time."
The paper: nature.com/articles/s41591-0…
PHARMAMEDIX retweeted
Just published @ScienceMagazine
Higher medical accuracy at lower cost
A 3-agent AI for sequential diagnosis nearly doubled accuracy compared with physicians for >300 @NEJM CPC cases and >300 ER cases AND substantially lowered cost, also exceeded performance of many AI models, physicians plus AI @HarshaNori
@erichorvitz @Dominic1King @mustafasuleyman
PHARMAMEDIX retweeted
Real-world AI goes to the urgent care setting.
A prospective feasibility assessment of patient-facing conversational Al held up very well according to patients and physicians for accuracy of diagnosis, among many other metrics
Just published @TheLancet @AdamRodmanMD @PeterBrodeurMD @alan_karthi @GoogleDeepMind
thelancet.com/journals/lance…
PHARMAMEDIX retweeted
Congratulations to Georg Nagel, PhD, from all of us in New Haven! We are proud of your remarkable accomplishment and extraordinary journey since your postdoctoral studies at YSM. Well deserved!
#NobelPrize #Physiology #Medicine
“I didn't expect this, really... I thought it's still too early to get this prize in medicine.”
Georg Nagel was in Italy when he received the news of his 2026 Nobel Prize in Physiology or Medicine. He was in Italy on a sun-drenched terrace, overlooking olive trees when he got the call.
Listen as he recalls the key research moment leading to the development of optogenetics.
PHARMAMEDIX retweeted
NEW: AI just became authorized to issue prescriptions in the US.
Nolla Health says its AI can assess patients, recommend treatment, and issue initial prescriptions, starting with acne care in Utah, with physician oversight available.
PHARMAMEDIX retweeted
Karl Deisseroth, MD, PhD, a Stanford University professor of bioengineering and of psychiatry and behavioral sciences, has been awarded the 2026 Nobel Prize in physiology or medicine “for discoveries leading to optogenetics, which makes it possible to switch on, or off, the activity of individual nerve cells in a living brain.” stanford.io/4xY9K0D
@StanfordEng @Stanford @bioe_stanford @NobelPrize
ALT 2026 Nobel Prize in Physiology or Medicine: Karl Deisseroth
PHARMAMEDIX retweeted
New Lancet Series: the way digital technologies are built & governed creates real health risks worldwide. A review of 9 million under-25s across 121 studies links digital media use to mental health & sleep problems plus reduced physical activity. Authors call for urgent, coordinated action.
➡️ spkl.io/60157VKUv
PHARMAMEDIX retweeted
A randomized trial of a patient-facing chatbot vs web based search enhanced accuracy of diagnosis of lung conditions
This is one of the first RCTs for public chatbot assessment in healthcare
nature.com/articles/s44360-0…
PHARMAMEDIX retweeted
Editorial @Nature this week calls out lack of real world data for AI medical algorithms
nature.com/articles/d41586-0…
PHARMAMEDIX retweeted
This is today’s edition of The Download, our weekday newsletter that provides a daily dose of what’s going on in the world of technology. Who’s liable when AI agents go rogue? Over the past few months, a cascade of cyberattacks by AI agents has stunned the world. In July, OpenAI disclosed that a swarm of its agents… trib.al/vedHjSb
PHARMAMEDIX retweeted
John Tsang, PhD, and his Yale lab members Jacob Kim and Ao Huang recently posed the question of whether #AI immunologists are ready for prime time. Specifically, they were referring to large language models (#LLMs) and whether they are as creative as humans in developing new hypotheses and methods for research in the field of immunology.
The answer, based on current studies, is no, not yet, says Tsang, Anthony N. Brady Professor of #Immunobiology and professor of biomedical engineering at Yale School of Medicine. While there are functions where AI can be effective, such as summarizing relevant literature, LLMs like ChatGPT aren’t able to consistently generate truly original hypotheses, scientific ideas, or experimental approaches.
Read more here: medicine.yale.edu/news-artic…
PHARMAMEDIX retweeted
It’s incredibly hard to see through the current AI hype in healthcare. Myriads of potential uses cases, a lot of companies with commercial interest and changing regulations dominate the field.
To see what is really going on and what we can realistically expect from AI to deliver in practice, I mapped the rapidly expanding universe of AI use cases in healthcare from early-stage “on the horizon” innovations to “safe bets” that are already backed by strong evidence.
I analyzed them on two scales: little evidence to evidence-based (meaning there are studies and peer-reviewed papers proving their efficiency and safety); and low risk to high risk (meaning patients’ lives might be at stake in case of an error).
This yielded four groups:
1) Speculative and risky (little evidence, high risk)
2) On the horizon (little evidence, low risk)
3) Handle with care (evidence-based, high risk)
4) Safe bet (evidence-based, low risk)
PHARMAMEDIX retweeted
AI is changing peer review. But how much should reviewers rely on it?
AI can help manage rising review volumes but faster doesn’t automatically mean better.
Tiffany Leung explores how reviewers can use AI thoughtfully without compromising the quality and integrity of peer review.
Read more: blog.jmir.org/navigating-ai-…
#PeerReviewWeek #PeerReview #PRW2026 #JMIR
PHARMAMEDIX retweeted
The legal implications of medical AI, a review of 31 cases involving AI tools
nature.com/articles/s44360-0…
PHARMAMEDIX retweeted
The case for science of the AI mind
gift link, by @camhberg
wsj.com/opinion/we-need-a-sc…
PHARMAMEDIX retweeted
AI May Be Breaking Mathematics — And One of the World’s Greatest Mathematicians Is Worried
What happens when AI becomes so powerful that it can solve mathematical problems faster than humans can understand them?
That question is now causing serious concern in the world of mathematics. OpenAI reportedly used around $15 million in computing power in an attempt to rapidly solve a case related to the legendary Navier–Stokes equations, one of the seven Millennium Prize Problems. The achievement may sound like a dream come true. But mathematician Terence Tao believes the real story is far more troubling.
Tao has warned that the AI race in mathematics could be moving too fast. His concern is not simply whether an AI-generated answer is correct. The deeper problem is what happens when machines produce complicated mathematical results without giving humans a clear path to understand, verify, teach, or build upon them.
For centuries, mathematics has advanced through human ideas, mistakes, intuition and explanations. A brilliant proof does more than produce an answer — it can open an entirely new way of thinking. If AI begins solving difficult problems faster than mathematicians can understand them, that process could start to disappear.
Tao has even raised the unusual idea of keeping certain mathematical problems away from automated solvers. It may sound extreme, but there is a mystery at the heart of the debate: if a machine discovers mathematics that humans cannot truly understand, have we advanced mathematics — or simply learned how to produce answers?
AI may be able to find the solution. But mathematics has always been about something deeper: understanding why the solution exists.
And that may be the part we cannot afford to lose.
Source: Choi, W. (2026). Fields Medalist Warns AI’s Math Rush Undermines Discipline’s Process. The Chosun Daily.
ALT AI May Be Breaking Mathematics — And One of the World’s Greatest Mathematicians Is Worried What happens when AI becomes so powerful that it can solve mathematical problems faster than humans can understand them? That question is now causing serious concern in the world of mathematics. OpenAI reportedly used around $15 million in computing power in an attempt to rapidly solve a case related to the legendary Navier–Stokes equations, one of the seven Millennium Prize Problems. The achievement may sound like a dream come true. But mathematician Terence Tao believes the real story is far more troubling. Tao has warned that the AI race in mathematics could be moving too fast. His concern is not simply whether an AI-generated answer is correct.
PHARMAMEDIX retweeted
Lessons from one of the first [and the largest] randomized trials of AI in medicine @NatureMedicine
A trial of >100,000 women undergoing mammography with or without AI support for interpretation, by Kristina Lång @lunduniversity
Most importantly, can it [AI] improve patient outcomes?
nature.com/articles/s41591-0…
PHARMAMEDIX retweeted
BREAKING NEWS!
The FDA just updated its list of FDA-approved/cleared AI-based medical technologies. Now they list over 1614 devices!
As always, I analyzed them for you.
1) Medical specialty still has a clear winner.
Radiology 1230
Cardiovascular 154
Neurology 76
Anesthesiology 30
Gastroenterology-Urology 27
Hematology 22
2) Type of submission
510(k) 1553
Premarket approval 21
De novo pathway 40
Check it out: fda.gov/medical-devices/soft…
PHARMAMEDIX retweeted
Fitbit founders James Park and Eric Friedman launched Luffu Link this week, a screenless wearable for caregivers, kids 13+ and aging parents. It tracks activity, sleep, calories and heart rate variability, uses AI to flag only important changes, and can send a voice message plus location with one button—no phone needed. Preorder is $250, with shipping expected in early 2027.
theflowspace.com/physical-he…
