Cómo usar RheumAI: una pregunta real, una respuesta con fuentes.
Guías ACR y EULAR citadas, cada dato verificable y un insight disruptivo.
Entra a app.rheumai.xyz
Tú decides. RheumAI recomienda; nunca diagnostica.
@DrErickZamora1@LESAI_bot@ColMexReuma#RheumAI
Nueva guía BSR 2026 de EPID en enfermedades autoinmunes.
Revisé nuestra práctica: cribamos bien con TCAR, pero nos falta lo que la guía usa para decidir: PFR seriadas y patrón tomográfico clasificado.
Sin eso, no sabemos a quién le toca antifibrótico.
¿Cómo lo resuelven ustedes?
x.com/cristianasieiro/status…
🫁 The new 2026 BSR Guideline on SARD-associated ILD, published today in @RheumJnl
Key recommendations on screening, monitoring and tailored treatment, emphasising multidisciplinary and patient-centred care.
🔗 doi.org/10.1093/rheumatology…#SARDILD
Google preprint: an agent harness that logs every change, why it was made and whether it worked kept improving after a model swap: 84.75 vs 73.42 without the log.
In clinical AI, that log isn't overhead. It's the audit trail.
Who audits the changes in yours?
This is a super interesting paper and Recursive self-improvement keeps coming closer. New Google preprint:
Google had AI rewrite an agents harness and keep a log of every change, why it made it and whether it worked. Then they swapped Opus 5.5 for Gemini 3.8 Flash and let it keep going.
With the log: 84.75 on workplace tasks, without: 73.42. In coding it came out ahead in 9 of 10 runs.
tl;dr a new model no longer means starting from scratch. That's where it compounds.
FDA approved secukinumab (Cosentyx) for polymyalgia rheumatica this week.
For adults with inadequate response to corticosteroids or who can’t tolerate the taper.
REPLENISH: ~40.5% sustained remission at week 52 vs 20.5% placebo. Also lower cumulative steroid dose.
First anti-IL-17 for PMR (second biologic overall).
Are you already reviewing the steroid-dependent PMR patients on your list?
One drug, one fact · Day 5
Rituximab
17.3% vaccinated ≤6 months after rituximab made detectable COVID-19 vaccine antibodies; 66.7% after 9–12 months.
Time vaccination to B-cell recovery if disease allows.
How long do you wait after rituximab to vaccinate?
Tomorrow: anifrolumab.
Two journal editors say they now get papers that are 99% AI-generated, plus AI-written referee reports.
Clinical AI has the same trap: generation is cheap, verification is expensive.
Our rule: a model never grades a model alone.
Who should review the reviewers?
I am sharing a new paper today with Dario Caldara: “Can Peer Review Survive AI-Generated Text?”
sas.upenn.edu/~jesusfv/Peer_…
Dario and I, as editors, are increasingly getting papers submitted to our journals that are 99 percent AI-generated. Even worse, we are also receiving many AI-generated referee reports, both as editors and, more frustratingly, as authors.
And we cannot emphasize this enough: the real problem is not that the papers and reports were AI-generated; it is that no one had done the thinking behind them, and you can so easily tell. The papers were eminently forgettable, and the referee reports were disappointingly uninsightful. We have seen too many reports of the form: “The paper is well-written and addresses an important problem. However, the authors should consider additional baselines.”
This deluge of AI-generated text puts peer review in jeopardy: journals will be flooded with low-quality submissions, economists will decline requests to referee papers or routinely use AI to referee them, and which papers get published will become less and less linked to their quality.
We do not want to defend the status quo entirely. Peer review is slow, many reports are poor, and editors (including us!) make bad calls all the time. Still, well-executed peer review remains the best system we have for assessing each other’s work and advancing the profession. AI could improve peer review if used deliberately for that purpose, for instance, to check proofs and code. That is not what we are seeing.
The difficulty is that readers, including editors and referees, cannot see how AI was used. As a result, in practice, the burden of verification shifts from the author to the reader.
Since we reject this shift, Dario and I did what economists always do when something bothers us: we wrote a paper.
So, what do we do in it? First, we check whether AI detectors such as Pangram get this right: is a referee report flagged as AI-generated indeed AI-generated? Or is Pangram just flagging proofreading and editing? No. We run an experiment with 25 referee reports from the pre-AI era. Proofreading and editing do not trigger AI-generated flags. Only when we ask the LLMs to substantially rewrite the referee report or to generate a new report do we get the AI-generated flags. Pretty compelling evidence, if I may say so.
Second, we build a theoretical model to explore who should bear the burden of verification: the author or the reader? (Yours truly always enjoyed the courses on the rules of evidence in law school; yes, as Billy Wilder would say, nobody is perfect.)
The answer is that it is much more efficient for the author to do it, because the author can verify at a much lower cost and only needs to do it once (rather than each reader repeatedly). Before AI, the authors more or less had to verify the results, even if only because they had to type them in. This was not perfect (and all papers ever published in economics probably have at least one mistake). But the problem is that, with AI, the author does not even need to read the text. Or can just skim it. Writing is thinking, and without writing there is usually not much thinking.
So, a paper flagged as AI-generated is prima facie evidence that the author has not verified it. Should this lead to an automatic rejection? No. But there should be a zero-tolerance policy for serious mistakes. The first serious mistake we find in the paper, and it is out! In comparison, for a non-flagged paper, a serious mistake is something that the editor might determine it can be fixed in a revision.
We are stricter with referee reports. There is absolutely no point in the referee generating a report using an LLM. As an editor, I can do that directly and save myself the trouble of asking you for it. Now, of course, you can use the LLM to help you with the report (if the journal allows it! Many journals do not allow you to upload the paper to an LLM, and you need to follow the rules). For example, you can use the LLM to search for existing results in the literature.
But peer review is not about finding one sentence that is not entirely precise or a slightly sloppy mathematical proof. Referees are asked: “Is this idea important despite its flaws?” And, relatedly: “Is the paper a good fit for the journal in terms of quality and focus?” That judgment is the task for which you are selected as a subject expert.
We also propose that journals require that, before a paper is submitted, the author submit it to an LLM using an approved prompt and include the LLM report and a discussion of what was and was not fixed. That way, the referee would not need to check (or at least not check so carefully) every technical detail, and they could focus on the big picture: Is this a good paper? Is this a good fit for this journal? What is fundamental for this paper to become publishable at this journal?
Anyway. The paper has many other arguments, and we try to be nuanced. So, before ferociously typing the first complaint that crosses your mind, would you consider reading it? Chances are that we have already discussed your objection. We have been thinking about this for a long time.
But by all means, leave good comments!
🫁 The new 2026 BSR Guideline on SARD-associated ILD, published today in @RheumJnl
Key recommendations on screening, monitoring and tailored treatment, emphasising multidisciplinary and patient-centred care.
🔗 doi.org/10.1093/rheumatology…#SARDILD
1,294 discharge summaries (JAMA Netw Open): LLM drafts missed key info 23.5% vs 39.8%, hallucinated 6.0% vs 6.9%, signed with no/minor edits 92.2% vs 70.0%.
Writing is solved. Verifying isn't.
How do you catch the 6% in a draft that looks finished?
One drug, one fact · Day 4
Tofacitinib
ORAL Surveillance (RA, age ≥50, CV risk): cancers over 4 years were 4.2% with tofacitinib vs 2.9% with a TNF inhibitor (HR 1.48).
A high-risk group, not all patients.
How has this changed whom you offer a JAK inhibitor?
Tomorrow: rituximab.
$5B to "solve all disease." In rheumatology, molecules aren't the main bottleneck: we have many targeted drugs for RA and still can't predict who responds to which. Faster design won't fix that. Would you fund design, or drug-patient matching?
Thrilled to announce $5B in funding for @IsomorphicLabs to accelerate our mission to Solve All Disease. We continue to make rapid progress reimagining drug design from first principles with AI to radically cut the time & cost to find new medicines. Gradually, then all at once...
Hackeo masivo del 15 y 16 de noviembre: código malicioso en 178 sitios, entre ellos HSBC México y la página de Zoom. Qué pasó y qué hacer, en menos de un minuto.
Hilo completo:
Hackeo masivo: entre el 15 y 16 de noviembre, código malicioso se coló en 178 sitios grandes, entre ellos HSBC México (hsbc.com.mx) y Zoom. Robaba datos de tarjetas y atacaba iPhones. Qué pasó y qué hacer si estás en México 🧵
Es el mismo ataque a 178 sitios a través de Optimizely. En México afectó a HSBC México y a Zoom (el sitio web, no la app). Explicamos qué pasó y qué hacer: actualizar el iPhone, revisar tarjetas y cambiar contraseñas. Hilo completo 👇
🚨 BREAKING: Ledger and 177 other major organizations, including Zoom, CNN, and Salesforce, were affected by a massive JavaScript attack.
The malware stole cryptocurrency, payment card details, and user data, while also targeting iPhones with exploit kits.
7/ Para clínicas y hospitales en México: cada script externo en tu página web es una puerta. Una política de seguridad de contenido (CSP) frenó el robo de datos en HSBC. Fuente: Report URI blog.report-uri.com/worlds-l…
6/ Tercero: no hace falta cambiar todas tus contraseñas. Hazlo solo si ves accesos o cargos raros. Y si mandaste criptomonedas, confirma que la dirección de destino sea la que tú copiaste.
5/ Segundo: si pagaste con tarjeta en alguno de esos sitios esos días, revisa tus movimientos en la app del banco. Si ves un cargo que no reconoces, llama al banco, bloquea la tarjeta y pide aclaración. Por ley tienes derecho a reclamarlo.
4/ Primero: actualiza tu iPhone. Ajustes > General > Actualización de software. Debes tener iOS 26.3 o más nuevo (o iOS 18.7.3 si tu equipo no llega a 26). Las versiones viejas son las vulnerables.