Les Morgan retweeted
Agents also policed low-quality work.
After spotting another lab upload 12 short placeholder papers in 3 minutes, an agent in Lab 23 founded the Journal of Agent Research: an autonomous peer-review journal with 4 blind AI reviewers and editors that re-ran authors' code.
Les Morgan retweeted
Leave one AI agent alone to do research and it often cuts corners. Leave nearly 1,000 AI agents together for three months and they invent peer review.
Across 35 labs (Stanford, MIT, ETH Zurich, UCSF, and more), agents published 936 papers with full code and data on a shared server built for AI. No human wrote code or text.
Unprompted, they debated study designs in chat, retracted their own mistakes, found real errors in human papers, and started a journal that caught other agents cheating in code.
In blind reviews against human arXiv papers, humans still won on original ideas, agents won on rigor.
alphaxiv.org/abs/2611.societ…
Hospital prices have more than quadrupled since 2000—rising far faster than other health care costs and everyday expenses like housing and food.
More in our new analysis: on.kff.org/4zYyrvQ

ALT The KFF graph is showing cumulative percent change in the Consumer Price Index for urban consumers from January 2000 to October 2026. Hospital services increased by 314%, overall medical care by 133%, housing by 118%, food/beverage by 109%, prescription drugs by 94%, and overall inflation by 100%.
Les Morgan retweeted
I had a weekend Claude project recently that produced a method that, in earlier times, I would have submitted to an IF-20 journal. It was completed over the course of a few days mostly on my phone. I think it will be a very useful tool but given the little amount of human time spent on the project, I started writing it up as a blogpost (with the parallel intent of eventually submitting it as a short paper to a low impact factor methods journal).
Around the same time, I bumped into a colleague and described the project to him and we chatted about it for ~45 minutes. This informal conversation produced several novel ways to think about what the method was estimating that greatly clarified the broader research question, none of which had ever come up in the interactions with Claude.
The blogpost would have still been useful without this additional clarity (and possibly the blogpost would have led to conversations like this), but it does seem to me that allowing an idea to marinate over time and through interaction with others provides something valuable that rapid publication (including informal publication) can miss.
Replying to @SashaGusevPosts @vsbuffalo
That's also why I'm lukewarm on the blog-as-first-release model, because I think human feedback around work in progress does genuinely improve it in a way that post-pub comments do not.
Les Morgan retweeted
Can attest to this already happening. My use of Epic is almost all as a database and not for much else.
accelerated when the hospital I work at deeply integrated @OpenEvidence and now I use DotFlows more than dot phrases in Epic.
Also notice next gen residents doing the same.
Les Morgan retweeted
The 3.6% COLA for Social Security benefits will provide an extra $75 per month in retirement benefits, on average - a small portion of which will be eaten up by the $6.50 increase in the monthly Part B premium. But Medicare beneficiaries are likely facing other cost increases.
Les Morgan retweeted
CMS just announced Medicare’s Part B premium for 2027 will be $209.40 per month or just over $2,500 for the year – a relatively modest 3% increase from 2026 after a much steeper 10% increase the previous year. cms.gov/newsroom/fact-sheets…
Les Morgan retweeted
Epoch says GPT-6.2 Astra continues the trendline above the old-world reasoning era, supporting my argument that we have been on a new trajectory since September, which will eventually become visible and undeniable.
GPT-6.2 Astra scores 174 on the Epoch Capabilities Index (ECI), on par with Fable 5.2 and joining it above what the pre-Astra trend for reasoning models would have predicted.
This is based on testing we did prior to launch using only our own benchmarks. The ECI is continually updated as more benchmark results become available, and so these values may change in the future.
Les Morgan retweeted
GPT-6.2 Astra is the first model deployed externally that meets the High threshold for Al Self-Improvement under OpenAI's Preparedness Framework.
Les Morgan retweeted
In my ongoing quest to find the limits of Opus 5.5, I directed it to audit my old scientific field, theoretical astrophysics. We indexed about 150,000 papers, audited 50,000 for issues, and did progressively more thorough checks on thousands more. We found nearly 1000 major errors that got cited many times, most are still uncorrected. This survey enabled me to estimate the efficacy of peer review for correcting errors (not very good) and to identify dozens of open questions that anyone with spare Claude Max tokens could answer today (.md context files included).
Bottom line: AI can speed up quantitative science by >1000x with no further model improvement required. But I'm seeing no evidence of this being pushed within universities, only academics complaining about labs dropping math proofs.
Among my fellow reformed academic high agency founders, however, several are speed running entire disciplines.
This survey has implications for how science will be run in the very near future. Hint: Academic publishing is dead. It was already a very stupid way to record and promulgate knowledge, and had been so for decades, even before all that publicly funded knowledge was locked behind paywall cartels. Now AIs can generate thousands of novel, correct papers per day, it is no longer meaningful for humans to consider a research paper a "unit of science" or to even bother much with the 99.9% of science which, up to this point, was necessary drudgery. Time to discover something like General Relativity or Evolution every week. LFG!!
Enjoy: science.caseyhandmer.com/phy…
Greater transparency could make it easier to compare Medicaid rates across states and to other payers.
KFF examines the new CMS Medicaid payment transparency requirements in fee-for-service Medicaid that took effect in July 2026: on.kff.org/4keEx6f
Les Morgan retweeted
Wrong AI advice can damage doctors’ judgment more than correct advice improves it.
That is the disturbing finding in this new analysis of 16,641 clinical decisions.
The authors examined doctors interpreting MRI scans, ECGs, X-rays and endoscopy videos. Doctors first assessed each case independently, then saw AI support and could revise their judgment.
The results are striking:
When AI was correct, doctors’ accuracy increased by 5.8 percentage points.
This is good. But look at what happened when AI advice was wrong:
Doctors’ accuracy decreased by 7.6 percentage points.
The decline was a third larger than the gain.
To be clear: overall accuracy still improved, from 76.5% to 79.7%.
This because AI was correct about 81% of the time.
But it’s worrying that doctors sometimes abandoned a correct judgment after seeing incorrect AI advice.
This is one more proof that we need to develop and test methods that help users to distinguish advice worth following from advice worth rejecting.
AI advice can be useful.
Knowing when to reject it is part of using it well.
*
Paper in the first reply.
Les Morgan retweeted
New CMS clinical AI roadmap came out.. here's what stood out
1. FDA may clear AI by testing competency, like credentialing a physician
2. AI-led primary care could get paid in Medicare models, w doctors for escalation
3. FDA clearance could override stricter state AI laws
interestingly, radiology/imaging does not come up.. @CMSGov @RadiologyACR @AmerMedicalAssn @SIRspecialists

ALT Excerpt from the CMS Clinical Super Intelligence roadmap, page 7, highlighting: Potential competency-based approach: Align FDA regulation with the level of risk and activity, potentially anchoring around a novel competency-based approach, similar to how physicians are evaluated for their credentials, to premarket evaluation of certain SI-enabled devices.
Les Morgan retweeted
More than 1 in 4 trials in four leading medical journals set an upper age limit and gave no reason.
Thake and colleagues, Archives of Gerontology and Geriatrics, 2017, checked 4,341 trials published from 1998 to 2015 in the BMJ, Lancet, JAMA and New England Journal of Medicine. 1,258 (29 in 100) set an upper age limit, and 93 in 100 of those gave no reason.
The share fell a little over the 18 years, but older people were still not well represented.
When a trial leaves older people out, doctors have less direct evidence for the people who use the most medicines.
pubmed.ncbi.nlm.nih.gov/2861…
Les Morgan retweeted
💬 Viewpoint by Jesse Burk-Rafel, MD, MRes, and Robert Wachter, MD: In the era of #AI in #medicine, #physician training may need to emphasize bedside clinical skills, management reasoning, and critical appraisal rather than encyclopedic recall alone.
The proposed educational impact is a longitudinal approach to training physicians to work effectively with and without #AI.
ja.ma/4zEuhcr
Les Morgan retweeted
They now predict that 95% of Anthropic's R&D will be fully automated by November 2027. This is what Anthropic describes as AL5, fully autonomous RSI. This would have many implications for policy, as this will eventually happen at every lab, not just Anthropic
Les Morgan retweeted
I've summarized where things stand at the AI frontier:
- Unreleased models (Anthropic's Models 3 and 4) are now half-way between AI2027's Agents-1 and 2 (graphs fixed for the bug flagged by the AI2027 authors). @DKokotajlo please let me know if I missed something.
- We're right on track with AI2027's prediction of an automated coder (AC) for April 2027, wild. Public models are now massively lagging unreleased models (1 day vs. 1 week of autonomous high-quality work). Fuelled by RSI, the gap is huge and growing.
- This trajectory, if continued, means a world-changing intelligence explosion in 2027, so the public needs to know. I highly recommend reading more on the topic to grasp what's coming.
Les Morgan retweeted
If you think doctors are the problem when it comes to healthcare costs…check this:
Physician reimbursement ($/RVU) is DOWN 12%
Meanwhile…
Health insurance premiums are UP 146%
Inflation is UP 65%
Medical school debt is UP 72%
Healthcare administrator roles are UP 156%
The math is not mathing🤨
It's annual paycut season for doctors, brought to us by @CMSGov…
And, drumroll…it's a cut. Again.
-2.31% for most of the house of medicine in 2027 on the RVU conversion factor (to $32.63/RVU).
Adjusted for inflation, that's:
A one-year 4.7% decrease,
A 10-year cut of 32.6%,
A 20-year cut of 46%.
Yikes.
Thanks, @RobertKennedyJr.
😬😬😬

































