Companion experiences for the newsletter — don't just read about it, try it.
A condensed scatter of all 30 specialties (drag the sample-size floor and watch the fake trend dissolve), plus a live OpenEvidence run showing why the hard half of point-of-care — building the question from a 1,200-document chart — never reaches the test set.
Data explorerSame answer, different score. Showing citations boosts OpenEvidence from 71% to 83% — while hurting every other model. The finding that upends how we evaluate clinical AI.
~3 minWearables skew toward the healthy and wealthy. But CDC data shows chronic disease clusters where coverage is thinnest. Every dot is a state — drag the coverage line and explore the gap.
Data explorerContinuous monitoring works — it cut blood pressure across 106,261 patients. So predict who's actually wearing the monitor, then meet the people the data forgets.
~3 minA naive model reaches for the answer the internet gives most often. Play the model, then meet the patient the common answer would have hurt.
~3 minBuild the checker that flags a confidently-common wrong answer before a human signs it. Discover why model confidence is the wrong signal to trust.
~3 minFrontier AI models ace medical benchmarks — then break when the question is rephrased. See if you can do better.
~3 minFrontier AI models top the medical benchmark — then collapse under stress. Toggle the perturbations and watch the leaderboard reshuffle.
~3 min20 runs. 3 models. 2 temperatures. One patient on Camila. Watch LLMs struggle with the estrogen trap hidden in prior visit notes.
~5 min