BRAINBITES
WEEK OF JUL 17โ23, 2026 ยท ISSUE 21
BRAINBITES ย โ Your weekly health, fitness, and longevity digest. Five minutes that cover the highest-signal podcasts and science before they hit the mainstream.
This week: neuroscientist Tommy Wood walked Huberman through the science of optimal learning โ the 85% rule, error-driven plasticity, and why the sweet spot for neural adaptation feels like controlled difficulty. And Sam Harris appeared in Essentials with meditation as attention training.
From: Huberman Lab ยท Huberman Lab Essentials
๐ข 15% โ the optimal error rate for maximal learning speed
๐๏ธ Tommy Wood on error signals, cognitive performance, and the 85% rule
๐ก Deep Dive: errors aren't the cost of learning โ they're the currency
๐ฐ News Bits: the 85% rule from Nature Communications ยท meditation as attention training ยท sleep replays errors, not successes
GM.
The most actionable episode this week is Tommy Wood on Huberman. Wood studies cognitive performance, and the finding he led with reframes practice: the optimal error rate for learning is approximately 15%. Below that, you're coasting. Above that, you're drowning. The sweet spot feels uncomfortable โ which is why most people avoid it.
โฑ ~5 min read
๐ง THE ERROR RATE
Why the optimal learning zone feels like failure โ and why that's the signal, not the noise.
Two findings from the same week. One: the brain learns fastest at roughly 15% error rate โ a prediction from machine learning confirmed in human data. Two: meditation is attention training, and the attention it builds is the same faculty that lets you stay in the error zone without flinching.
The thread this week: the discomfort you feel when learning is not interference. It's the signal that adaptation is happening.
๐ข THE VITAL SIGN ยท NEUROSCIENCE / LEARNING
15%
The optimal error rate for maximal learning speed โ the '85% rule' from computational learning theory.
The 85% rule comes from a 2019 Nature Communications paper: for binary classification tasks, ~15% error rate maximises information gain. Wood extended this to human domains โ motor skills, language learning, academic study โ all show fastest gains at difficulty levels producing about 1-in-7 failures.
Errors trigger acetylcholine and norepinephrine release โ the neurochemical cocktail that opens plasticity windows. Correct responses don't trigger the same release. Practice without errors reinforces existing circuits. Practice with errors builds new ones.
Sleep consolidation adds the final piece: the brain replays error trials during sleep more than success trials. Errors + focused attention + good sleep = the complete learning chain.
If you're not failing 15% of the time, you're not learning. You're performing.
The 85% Rule, Nature Communications ยท July 2026
๐๏ธ THE EPISODE ยท NEUROSCIENCE / PERFORMANCE
The 85% rule โ why the optimal learning zone feels uncomfortable
Tommy Wood on error rates, cognitive performance, and why making mistakes faster is the strategy.
Dr. Tommy Wood ยท Huberman Lab ยท July 2026
Wood bridges cognitive science and performance. The conversation is crisp and directly applicable โ whether you're learning a language, training a skill, or structuring study.
Errors generate neuroplasticity โ comfort generates maintenance.
Errors trigger acetylcholine and norepinephrine release, opening plasticity windows. Correct responses don't. If your practice doesn't produce errors, it's reinforcing existing circuits, not building new ones.
๐๏ธ Skip to ~25:00 for the neurochemistry of error-driven learning.
Calibrate difficulty, not duration.
A 30-minute session at the right difficulty produces more learning than a 2-hour session that's too easy. Progressive overload isn't just about adding weight โ it's about staying where some reps fail. For studying: use spaced-repetition software that adjusts difficulty based on your error rate.
Sleep consolidates errors, not successes.
Hippocampal replay during sleep disproportionately features error trials. Sleeping after a practice session with errors consolidates the learning. Sleeping after an easy session consolidates nothing new. Errors + sleep = adaptation.
๐๏ธ Listen โ Huberman Lab, Dr. Tommy Wood, July 2026
๐ก DEEP DIVE ยท THE ERROR SIGNAL
Wood gives you the learning science. Harris gives you the attention infrastructure. Together: errors + focus = growth.
Sam Harris's Essentials episode connects through attention. The 15% error rate only produces learning if the brain attends to the errors. Most people disengage when they fail โ emotionally or behaviourally. The meditator's skill is observing discomfort without reacting to it, which means staying in the error zone longer.
Wood's sleep finding adds another layer: the brain replays errors during sleep, but only errors that were attentionally encoded during waking. Disengaged errors don't get replayed. Focused attention during errors โ deeper encoding โ better sleep consolidation โ faster learning.
The practical synthesis: learning speed depends on error rate (15%), attention quality (staying engaged through errors), and sleep quality (consolidation). All three are trainable.
Harris strips meditation to its cognitive core: training the capacity to notice the contents of consciousness without being captured by them. This attentional stability is the meta-skill supporting every other cognitive task.
The error is the data. The attention is the processor. The sleep is the commit. Get all three right and learning takes care of itself.
๐ฐ NEWS BITS
The 85% rule: 15% error rate maximises learning speed.
A 2019 Nature Communications paper proved that ~15.87% error rate maximises learning speed for classification tasks. The model was validated in machine-learning systems and animal conditioning studies. Wood extended it to human domains: fastest gains occur at difficulty levels where errors happen about 1 in 7 attempts.
โ Nature Communications, Wilson et al., 2019
Meditation as attention training โ Sam Harris's framework.
Harris's Essentials episode strips meditation to its cognitive core: training the capacity to notice when attention has wandered and return it. No spirituality, no relaxation goals. This attentional stability supports every other cognitive task, from learning to emotional regulation.
โ Huberman Lab Essentials, Sam Harris, July 2026
Sleep replays errors more than successes โ the consolidation bias.
Hippocampal replay during sleep disproportionately features error trials. The brain uses sleep to extract corrective information from failures. A training session with errors followed by good sleep produces more learning than a perfect session or a session followed by poor sleep.
โ Huberman Lab, Dr. Tommy Wood, July 2026
P.S. Were you failing 15% of the time in your last week of training or study โ or was it too easy? Hit reply and be honest.
What did you think of today's BrainBite?
๐ง ๐ง ๐ง ๐ง ๐ง ย The 85% rule just restructured how I practice guitar, study, and program workouts
๐ง ๐ง ๐ง ย Sleep-consolidation-of-errors is wild. The Harris connection is the sleeper hit.
๐ง ย Good data. Now give me the supplement stack.
We listen to every health podcast so you don't have to. Weekly. No noise.
If it doesn't feel a little wrong, it's not teaching you anything. See you next week.
