Andrew Huberman· PhD
Scientific findings suggest that setting the level of challenge to one where you make errors ~15% of the time is ideal for learning.
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.
Scientific findings suggest that setting the level of challenge to one where you make errors ~15% of the time is ideal for learning.
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.
if you want to learn something probably setting the difficulty of what you're trying to learn to about 85% correct trials 15% error trials is probably ideal
a rough estimate of difficulty should be about uh 15% of questions or challenges so these could be cognitive challenges or physical challenges should lead to failures noninjurious failures getting the answer wrong about 15 % of the time tends to optimize learning across a number of different domains