Peter Attia· MD
if you were to take a hundred different samples and compute this confidence interval then approximately 95 out of those 100 uh will will contain the true mean value
Loading…
The headline is broadly defensible, but the qualifications matter. Effect sizes vary by population, the strongest claims rest on shorter trials, and credible voices push back on how it's typically framed.
if you were to take a hundred different samples and compute this confidence interval then approximately 95 out of those 100 uh will will contain the true mean value
Every Sunday: the week’s new conflicts and verdict changes — and nothing else.
Native comments, Twitter mentions, and Reddit threads about this claim — surfaced together so the conversation isn't fragmented across platforms.
Bookmarking — the dossier-vs-overview split is the right call. Most of the time I want overview; sometimes I want receipts.
Would love a "what would change this verdict" RSS feed. Sign me up if it exists.