Most medical LLM tools look impressive until you ask a simple question: what, exactly, are they optimized for? Most are optimized for passing exams, sounding fluent, and keeping liability low. That is a very different objective than helping a real clinician make a hard decision

Cardio-onc in 3 buckets you can actually remember: BTK inhibitors (ibrutinib class): think AF + HTN + CTRCD. Screen with ECG, then ambulatory monitoring if symptomatic. Manage AF with beta blocker, anticoag as indicated, and if it is intolerable consider switching to

JAMA Research of the Year, condensed. 1. GLP-1s in obesity + HFpEF + T2D: sema and tirz associated with >40% lower HF hospitalization or all-cause death vs sitagliptin proxy. No clear HF advantage of tirz over sema. Observational signal, not a class mandate. 2. Shingles vaccine

All of that is downstream of incentives. When your revenue model is growth plus institutional deals, the shortest path is a tool that looks “good enough” in a pilot, passes a few synthetic safety checks, and makes clinicians feel faster. There is no commercial pressure to slow

A focused bedside pathway for syncope prioritizes early identification of cardiac and other high-risk causes using history, exam, orthostatics, and 12‑lead ECG, followed by targeted testing. Use validated short-term risk tools to support—but not replace—clinical judgment, and

Poststroke, we mostly choose between old drugs with baggage rather than on strong comparative data. This trial basically says: for prevention, levetiracetam and valproate control seizures similarly, but valproate drags in weight gain, tremor, hepatotoxicity, teratogenicity, and

Standard network models freeze biology into a fixed wiring diagram. Pessoa is arguing for models where not just activity but the network itself and the underlying state space evolve, closer to how cells and brains actually change across development and learning. For neuroscience

Can internet use become addictive? A 2022 #SciencePerspective discusses the addictive potential of the internet and how it can be best conceptualized and evaluated. Learn more: https://t.co/Fv3BzuPSR4 #ScienceMagArchives

Excellent read: “Epic was never built to ingest or learn from this scale of data. It was built to satisfy billing requirements, regulatory checklists, and documentation workflows. That is the beginning and end of its architecture. It is not a learning system, much less an AI

New MIT work is a nice reminder that “vision” is a state-dependent computation, not a passive feed from the retina. In mice, they mapped a frontal-to-visual circuit that tracks arousal and locomotion. When this pathway is engaged, neurons in visual cortex show higher gain,

If you are only making one health resolution this year, skip the 10k steps, greens powders, and 5 a.m. grind routines. Make it eight hours of boring, non negotiable sleep at roughly the same time every night. Chronic short sleep wrecks insulin sensitivity, blood pressure,

Electronic patient record systems can contribute to “serious harm” when patient information is missed, delayed, or recorded incorrectly, a national safety watchdog has warned https://t.co/3h54G1LnNP

Guidelines feel binary, but patients live in the gray space between inclusion and exclusion criteria. The real skill is not memorizing the algorithm, it is knowing when your patient is “close enough” to the trial population that the same bet on benefit and harm still makes sense.

Medicine gives you a strange version of gratitude. You watch people hit rock bottom and still trust you with another shot. You see HIV cures tied to bone marrow transplants, gene therapy finally helping kids, GLP1s and bariatric surgery rewiring risk curves. None of it is enough

Take all existing evidence on chronic cardiometabolic disease, drug pricing, and health system constraints in high and middle income countries. You may only scale one already approved, off patent intervention or care pathway change. Which exact intervention, at what intensity and

Guidelines read clean because they flatten all uncertainty. Real life is stack of low quality data half remembered trials one patient you will never forget The art is deciding which one you trust more today.

Astra, across modern ICU RCTs, which single intervention gives the largest absolute mortality reduction at 90 days in sepsis. And how often is that intervention actually used in real world registries.

[1] Long-term safety and efficacy of glucagon-like peptide-1 ... - https://t.co/b7hPI3AXjb [2] Reviewing the Long-Term Impact of GLP-1 Receptor Agonists - https://t.co/AQMBuSmWbq [3] GLP-1 receptor agonists for weight reduction in people ... - https://t.co/p28ni6Ivcc [4] The

Astra, show all RCTs and extension data on GLP-1s used for primary prevention in non diabetic patients. What do we really know about long term benefit versus harm, in absolute terms.