All Posts

Doximity surveyed 3,151 physicians. Adoption is widespread, not generational. TL;DR: 61% of physicians 30 and under use AI, 53% of physicians 60+ use it - only 8 percentage point difference Neurology leads specialty adoption at 64%, but most specialties are above 50% Age isn't the primary barrier… policy clarity, tool reliability, and workflow integration matter more Specialty-agnostic tools with workflow integration reach more physicians than narrow specialty tools The age a

Benchmark studies use complete data and clear criteria. Medicine doesn't work that way. TL;DR: "AI outperforms doctors" headlines come from benchmarks with complete information and structured formats Real clinical practice involves incomplete data, ambiguous presentations, and contextual factors not in electronic records Both positive and negative synthetic research creates misleading narratives about AI deployment The gap between benchmark performance and clinical utility ru

AI-Rx - Your weekly dose of healthcare innovation Estimated reading time: 3 minutes TL;DR 300 million people with rare diseases wait 5+ years for diagnosis DeepRare (multi-agent AI, Nature) beat experienced physicians: 64.4% vs 54.6% Tested across 2,919 diseases, 14 specialties, 9 datasets from Asia, North America, Europe Architecture matters: self-reflective loops, specialized agents, LLM-agnostic design Multi-agent orchestration yielded 20-30 point gains over standalone LLM

AI-Rx - Your weekly dose of healthcare innovation Estimated reading time: 3 minutes TL;DR 40% of US physicians use OpenEvidence daily - that's infrastructure, not pilot 20 million clinical consultations in January 2026 alone Sutter Health integrated it directly into Epic workflows system-wide Three insights: integration matters more than performance, natural language is table stakes, scale validates utility differently than accuracy Lesson: Evaluate AI on workflow integration

AI-Rx — Your weekly dose of healthcare innovation Estimated reading time: 3 minutes TL;DR UpToDate launched Expert AI after testing 1,000+ versions (most clinical AI tools test fewer than 100) Trained exclusively on curated, peer-reviewed UpToDate content, not raw medical literature OpenEvidence scored 41% on specialty-specific board questions Led by practicing clinicians with extensive red teaming and validation Accuracy before launch matters more than speed when decisions a

AI-Rx Your Weekly Dose of Healthcare Innovation Estimated Reading Time: 4 minutes TL;DR: → Physicians aren’t resisting AI… they’re resisting unsafe, poorly integrated AI. → New survey data shows they want feedback loops, privacy transparency, EHR integration, training, liability clarity, and safety validation. → These aren’t obstacles; they are design requirements for clinical adoption. → AI tools gain physician trust when they are validated, integrated into workflows, and su

Recently, OpenAI updated its usage policy to ban “tailored advice that requires a license, such as medical or legal guidance, without a licensed professional involved.” That might sound like legal fine print, but it’s a major shift for healthcare. AI shouldn’t replace licensed expertise. It should reinforce it. This change matters because in healthcare, being almost right can still be dangerously wrong. A chatbot recommending a wrong dosage or missing a critical diagnosis isn

Hey Reader, Here’s what most people miss: →They treat AI scribes as “just another tech tool.” → They underestimate how quickly clinicians adapt once they feel real value. → And they forget that medicine changes fast when something gives time back to clinicians. The reason this matters is because how we adopt AI scribes will define how physicians spend their time for the next decade. Once you understand what’s actually happening, you’ll see that AI scribes aren’t a convenience

AI-Rx - Your weekly dose of healthcare innovation Estimated reading time: 5 minutes TL;DR LLMs trained on low-quality social media content show systematic cognitive decline Reasoning drops 24%, long-context understanding drops 38% "Thought-skipping" is the primary failure mode… models truncate reasoning chains The damage persists even after extensive fine-tuning on clean data Data quality isn't just a performance issue, it's a training-time safety problem Healthcare AI deploy