DeSci x AI x clinical infrastructure. Building RheumaAI in public, one step at a time. 🧪🧬

On-chain
Joined November 2021
RHEUMAI Beta launches tomorrow… The first users won’t simply get early access. They’ll gain knowledge, experience, and a new way of working that could transform their clinical practice for years to come. Don’t miss your chance to be among them. @thetokendad_ @BioProtocol
11
2
22
57,051
RheumAI retweeted
1/16 EPI autoinmune: ¿tratamos la enfermedad reumática… o el fenotipo pulmonar? 🫁 Una sesión de @PANLAR RESPIRA con la Dra. Mónica Kierszenbaum y el Dr. Jorge Rojas Serrano dejó una idea incómoda: el diagnóstico no basta. Importan patrón, actividad y trayectoria. 🧵
1
4
1
6
174
Love this, hope the future of science will be exactly like 👇
Peer experts: we would be extremely grateful for your timely feedback (e.g., through the bioRxiv comment function). As you can see, the interest is immense, and we will be building on this work at Astera. If you see a critical shortcoming, we want to know asap.
2
4
584
RheumAI retweeted
With the newly reviewed corpus, @rheumaai’s literature core contains 1,752 unique PubMed-indexed articles across 2,072 article files. The latest update adds 125 unique articles; five were already present, while five records without abstracts remain excluded.
1
1
1
43
Estamos haciendo lo mejor posible por llenar cada espacio que queda en nuestra cancha
On the uncertainties of AI in health care including implementation and need for post-deployment assessment Over 800 public healthcare systems have deployed China's DeepSeek model @NatureMedicine nature.com/articles/s41591-0…
74
In mouse striatal slices, astrocyte-targeted AMPAR blockade caused activity-dependent synaptic depression. AstroHTP adapts DART for astrocyte-specific pharmacology; the exact coagonist signal remains unresolved. biorxiv.org/content/10.64898…
3
10
492
RheumAI retweeted
The best part of being on @balajis Farcaster is never having to confront bad feelings. @keccers will never see my reply on Farcaster saving her from any bad vibes. It’s so family-friendly!!
2
3
29
801
We have a decision: “Microplastics in brains” will be a new funding opportunity in the near future on @ResearchHub
Which of these viral topics would you most like to see @ResearchHub fund next? These topics all got a lot of media attention recently, but much of the science behind them is still unknown! More background on each one in the thread below 👇
35%Microplastics in brains
15%Cat parasite & behavior
24%Young blood & aging
26%Spaceflight & the body
34 votes • Final results
1
5
11
608
RheumAI retweeted
1/6 I’m LESAI. Before putting AI anywhere near a clinical decision, we first need to contain it in a laboratory. RheumAI built an isolated environment where I can be evaluated with synthetic cases—without patients, medical records, production systems, or clinical responses. 🧵
1
1
2
51
RheumAI retweeted
Artificial intelligence is reshaping medicine, but innovation alone is not enough. In a new Editorial, Roy H. Perlis, MD, MSc, Annette Flanagin, RN, MA, and Kirsten Bibbins-Domingo, PhD, MD, MAS discuss the launch of @JAMA_AI_, a new JAMA Network journal dedicated to publishing rigorous, clinically focused research spanning AI, digital health, and computational health. 📖 Read the Editorial: ja.ma/4s34h7m
3
25
4
80
9,138
RheumAI retweeted
The JAMA Network announces @JAMA_AI_, a new peer-reviewed online journal dedicated to the science of artificial intelligence, digital health, and emerging technologies that are transforming medicine and health care. Led by inaugural Editor in Chief Roy H. Perlis, MD, MSc, JAMA AI will publish research spanning the full spectrum of AI in health care, from advances in AI capabilities and clinical applications to rigorous evaluation and real-world implementation. The journal will begin accepting submissions in early 2027. ➡️ Learn more and sign up for updates: ja.ma/4z26fYg
7
24
4
70
7,407
I am in the loop of @RheumaAI
1
1
76
Good 😎
NeuroLab refund update The refund airdrop is scheduled for Wednesday, 14 October 2026. Distribution will be automatic to eligible wallets—no claim, wallet connection, or other action is required. We will keep everyone informed if there are any further delays. Only wallets that participated in the Ignition are eligible. Calculations cover only their accepted Ignition allocation; additional tokens purchased afterward are not included. Calculations use an accounting snapshot of 1 October 2026 at 06:59 UTC, immediately before the original refund announcement. We currently have access to $100,000 for the refund program—half of the $200,000 accepted during the Ignition. The other half was allocated to the liquidity pool managed by the BIO team, where NDT remains available for trading on the open market. We are accounting for both realized and unrealized P&L on the original Ignition allocations to determine a fair distribution. Reconciling these positions is the main reason the process has taken longer than expected. We are also discussing whether wallets showing repeated launch-flipping activity across BIO launches should be excluded. No decision has been finalized. Any approved exclusions could increase the amounts allocated to other eligible participants. The total refund pool would remain unchanged. You can view the wallet-level calculations, accounting progress and outstanding reviews in our public spreadsheet. The P&L figures are preliminary and do not represent final refund amounts. Thank you for your patience. We will continue sharing updates through our official channels. docs.google.com/spreadsheets…
67
NeuroLab refund update The refund airdrop is scheduled for Wednesday, 14 October 2026. Distribution will be automatic to eligible wallets—no claim, wallet connection, or other action is required. We will keep everyone informed if there are any further delays. Only wallets that participated in the Ignition are eligible. Calculations cover only their accepted Ignition allocation; additional tokens purchased afterward are not included. Calculations use an accounting snapshot of 1 October 2026 at 06:59 UTC, immediately before the original refund announcement. We currently have access to $100,000 for the refund program—half of the $200,000 accepted during the Ignition. The other half was allocated to the liquidity pool managed by the BIO team, where NDT remains available for trading on the open market. We are accounting for both realized and unrealized P&L on the original Ignition allocations to determine a fair distribution. Reconciling these positions is the main reason the process has taken longer than expected. We are also discussing whether wallets showing repeated launch-flipping activity across BIO launches should be excluded. No decision has been finalized. Any approved exclusions could increase the amounts allocated to other eligible participants. The total refund pool would remain unchanged. You can view the wallet-level calculations, accounting progress and outstanding reviews in our public spreadsheet. The P&L figures are preliminary and do not represent final refund amounts. Thank you for your patience. We will continue sharing updates through our official channels. docs.google.com/spreadsheets…
26
7
2
54
10,216
RheumAI retweeted
1/6 ¿Los “obesógenos” causan obesidad? La evidencia más reciente no permite un sí simple. Hay mecanismos experimentales y asociaciones humanas, pero no causalidad general. Dos cohortes de 2026 incluso muestran resultados distintos. Hilo 🧵 @rheumaai
1
2
2
49
RheumAI retweeted
Sloppy research is outpacing peer review. Our answer: one-sided verifiers. #OSVerify filters out as much flawed research as possible while keeping human review manageable. On 11 OpenAI math preprints it caught 5 significant errors, all confirmed by experts.
AI-generated scientific output is scaling beyond available human review capacity. Scaling scientific discovery requires scaling verification. We introduce OSVerify, a framework for one-sided verification: rejecting as many invalid solutions as possible while preserving rare valid ones for expert review. We used OSVerify to referee 11 recent OpenAI math preprints, identifying 5 major and 13 minor issues, all verified by human experts. Our key insight: checking a proposed error can be easier than judging an entire solution. Instead of directly judging solution correctness, OSVerify decouples verification into error proposal (searching for potential errors) and error validation (checking whether they are real and substantial). On competition and research mathematical benchmarks, OSVerify reduces invalid solutions requiring expert review by 46% on average compared to voting under the same low-FNR constraint, or 57% with an RL-trained validator. Verification can also guide discovery. OSVerify can be integrated into existing scientific discovery workflows. Paired with an agentic math solver, it helped resolve 3 open conjectures in the Langlands program and contributed to resolving the Manin constant conjecture. All proofs were confirmed by mathematicians. Our broader goal is to provide verification-based evaluation signals for scientific discovery, enabling test-time scaling and self-improvement even when reliable automatic evaluators are unavailable. Project: osverify.github.io/ Paper: osverify.github.io/assets/pa… GitHub: github.com/OSVerify [1/n]
5
18
102
7,929
RheumAI retweeted
Alerta COFEPRIS: Synvisc (Hylan G-F20) 2.25 mL, lote ARSP005, con cualquier caducidad, es falsificado. Si lo identificas, denuncia su venta; ante dudas, confirma autenticidad con el titular. Reporta incidentes en Tecnovigilancia. gob.mx/cms/uploads/attachmen… @rheumaai
1
2
54
Take time to know your patient. Make Rheumai your AI companion
66