Epstein explains that the most popular finding in human development research rests on a 1993 study of 30 students already inside a world-class music academy — a restriction of range on the very variable it claimed to explain. The paper reported only group averages and asserted complete correspondence between practice hours and skill rank. When updated data was published two decades later, that claim collapsed: some who practised less were better, some had gone far past 10,000 hours without matching them.

The original 1993 study used 30 violinists at a world-class academy.
Restriction of range means the sample was pre-selected on skill itself.
Only averages were reported, hiding the individual variation.
Updated data 20 years later disproved the complete correspondence claim.

“the original paper was done on 30 violinists”
David Epstein · 18:00

“the original conclusion was wrong there was not complete correspondence”
David Epstein · 21:00

#deliberate-practice#10000-hours#research-critique#expertise

Myth Buster22:30

The average founder of a top-tier fast-growing startup is 45

Bartlett, at 30, asks whether he's already too late. Epstein cites MIT, Northwestern and US Census Bureau research on the fastest-growing tech startups — the top one in ten thousand — where the average founder age on the day of founding was 45, and a 50-year-old had better odds than a 30-year-old. Tech in that dataset spans agriculture and industry, not just photo-sharing apps. We hear only the Zuckerberg story because we valorise precocity, not because it's the norm.

Average founder age of a top 1-in-10,000 growth startup was 45.
A 50-year-old had a better chance than a 30-year-old.
'Tech' in the study spans agriculture and industry broadly.
Zuckerberg said young people are just smarter — at 22.

“a 50-year-old had a better chance than a 30-year-old”
David Epstein · 23:00

“we like valorize precocity”
David Epstein · 23:00

#entrepreneurship#age#founders#precocity

Myth Buster1:17:30

West Point didn't have a grit problem — it had a match quality problem

Angela Duckworth's grit survey, half consistency of interests and half persistence of effort, predicted who survived West Point's brutal Beast orientation better than test scores. But since the mid-1990s roughly half of those gritty cadets have quit almost on the day their five-year service commitment ends. The army diagnosed a millennial problem and threw half a billion taxpayer dollars of retention bonuses at it. Officer-scientists reframed it: young people learn things about themselves in their early twenties and had no agency to switch.

Grit outpredicted test scores for surviving Beast orientation.
Nearly half of gritty cadets quit at the end of their commitment.
Half a billion dollars in retention bonuses didn't fix it.
Lateral mobility elsewhere in the knowledge economy changed the calculus.

“we don't we haven't gotten a grit problem overnight we've got a match quality problem”
David Epstein · 1:19:00

“90% of the army Cadets who went through that process changed their career preference”
David Epstein · 1:17:00

#grit#match-quality#retention#careers

Hot Take· 1

Hot Take1:56:00

You're less likely to die when the top cardiologists are at a conference

Two Harvard-led studies found that patients admitted to teaching hospitals with certain cardiac conditions during a national cardiology convention — when the most esteemed specialists are away — are less likely to die. The explanation is the einstellung effect: having solved a problem the same way thousands of times, surgeons keep applying that solution even when the problem has changed or data shows it no longer works. The same researcher also found that surgery on the surgeon's birthday raises 30-day mortality across 980,000 procedures.

Two Harvard-led studies, teaching hospitals, national cardiology convention dates.
The einstellung effect locks experts into a previously successful solution.
Specialisation in medicine is both inevitable and a double-edged sword.
Across 980,000 procedures, surgery on the surgeon's birthday raised 30-day mortality.

“if you're checked in to a teaching hospital with certain cardiac conditions on the dates of a national Cardiology convention when the most esteemed Specialists…”
David Epstein · 1:57:00

“if you have a procedure on the surgeon's birthday you're more likely to die within the 30 days after the surgical procedure”
David Epstein · 1:58:00

#specialisation#medicine#einstellung-effect#expertise-risk

Story· 3

Story27:30

Bartlett hired a head of failure — and Epstein explains why it works

When the podcast went number one in Europe, Bartlett hired someone whose sole job was increasing the rate of failure and experimentation, then embedded a failure assistant in each department. Epstein connects it to Adam Grant's hippo effect — the highest-paid person's signal governs behaviour — so an actual title, rather than lip service, underwrites risk psychologically. He calls the resulting cross-team reporting an import-export business of ideas, and worries hybrid work removes the serendipity that used to supply it.

A failure assistant sits inside each of the four departments.
The social team's TikTok discovery transferred straight to guest research.
The hippo effect means leadership signalling drives whether risk happens.
Bill Gore built Gore-Tex's company on 'real communication happens in the carpool'.

“there's this real 1 plus 1 equal 3 getting the teams to share their fails failures and experiments so they don't have to fail in…”
Steven Bartlett · 29:30

“the clear message of his work is that exploration precedes a hot streak”
David Epstein · 31:00

#failure#team-culture#experimentation#innovation

Story49:30

The doctor who predicted typhoid by touching tongues — and was spreading it

Epstein's favourite illustration of a wicked learning environment. A New York physician became famous for diagnosing typhoid by palpating patients' tongues with his hands, correctly predicting who would fall ill a week or two later. A colleague later observed he was a more prolific spreader of typhoid than Typhoid Mary — he was causing the outcomes he predicted. Every confirmation reinforced the exact wrong lesson about his own diagnostic skill.

He predicted typhoid by feeling patients' tongues by hand.
The predictions came true because he was transmitting the disease.
Accurate-seeming feedback reinforced a completely false lesson.
Judges and many clinicians rarely learn the consequences of decisions at all.

“he was a more prolific spreader of typhoid than even Typhoid Mary”
David Epstein · 50:30

#wicked-environments#feedback#medicine#judgment

Story51:30

The Game Boy was a technological joke — and it destroyed the competition

Gunpei Yokoi scored poorly on university electronics exams and took a low-tier machine maintenance job at a Kyoto company making flower playing cards while prestigious peers went to Tokyo. Knowing he couldn't compete at the cutting edge, he combined mature calculator and credit-card technology into handheld games. The Game Boy had a decade-old processor and a screen Epstein describes as looking like rotting alfalfa, launched against colour rivals — and won on durability, affordability, portability, battery life and game selection.

Yokoi's weak electronics results pushed him away from the cutting edge.
He combined well-known calculator and credit-card technology.
The Game Boy shipped a decade-old processor against colour competitors.
Since the 1980s the most impactful patents merge across technology classes.

“lateral thinking with withered technology that was his phrase which means taking things that are already well understood and moving them somewhere where they're seen…”
David Epstein · 53:30

#innovation#nintendo#constraints#recombination

insight· 4

insight18:30

Study only NBA players and you'd tell parents to have shorter children

Epstein's clearest illustration of restriction of range. Across American men, height correlates strongly with the chance of scoring points in the NBA. But restrict the sample to players already in the league and the correlation flips negative, because guards score more and guards are shorter. Draw conclusions from that pre-screened group alone and you'd produce exactly backwards advice — the same error that generated the 10,000 hours rule.

Across all American men, height predicts NBA scoring positively.
Within the NBA the correlation turns negative because guards score most.
Squashing variation in a variable destroys your ability to see its effect.
Highly pre-screened samples produce backwards prescriptions.

“if you just did that study with only NBA players you would tell parents to have shorter children to have them score more points in…”
David Epstein · 19:30

#statistics#restriction-of-range#research-literacy#basketball

insight1:01:30

Turn off your notifications and your brain will interrupt you instead

One of Gloria Mark's strangest findings: we carry an internal barometer set to a habitual cadence of interruption. Cut the external notifications and you begin self-interrupting to maintain that rhythm — suddenly remembering someone you didn't reply to, something you forgot to check. It does fade, but slowly, so this has to be treated as habit formation rather than a single day of discipline. Epstein's mitigation is to keep a pad nearby and cognitively outsource each intrusive thought the moment it lands.

We maintain an accustomed cadence of interruption internally.
Removing notifications triggers self-interruption for a while.
The effect decays slowly — treat it as habit formation.
Write intrusive thoughts on a pad to get them out of working memory.

“you will start self- interrupting to maintain the Cadence of interruptions to which you have become accustomed”
David Epstein · 1:02:00

“cognitively Outsource it so at least it's not sitting in there”
David Epstein · 1:03:30

#attention#notifications#habits#distraction

insight48:00

Kind vs wicked: why practice makes some people confident but not better

Epstein was stuck between two research camps finding opposite results until Daniel Kahneman, over a lunch he'd arranged months after hearing Epstein speak, handed him Robin Hogarth's distinction. Kind environments have clear goals, repeating patterns, stable rules and fast accurate feedback — chess, golf. Wicked environments have patterns that mislead, rules that shift, and delayed or absent feedback. In wicked domains, narrow repetitive practice often makes people more confident without making them better.

Kahneman personally pointed Epstein to Hogarth's kind/wicked distinction.
Kind: clear goals, repeating patterns, quick accurate feedback.
Wicked: misleading patterns, shifting rules, delayed feedback.
Ericsson later caveated that his framework needs a coach who can correct everything you do wrong.

“kind is like next steps and goals are clear rules repeat”
David Epstein · 48:30

“wicked learning environments where patterns don't just repeat they might fool you rules may change if there are any feedback could be delayed or inaccurate”
David Epstein · 49:00

#learning-environments#expertise#feedback#kahneman

insight1:49:30

Foxes with dragonfly eyes beat the intelligence community at forecasting

A 20-year forecasting programme collected 83,000 specific probability predictions on geopolitical events. The worst forecasters were the most narrowly specialised — hedgehogs who knew one big thing and wrapped every event into their single mental model. The best were foxes who knew many little things, gathered perspectives, and actively got others to falsify their own hypotheses. The lead researcher described them as having dragonfly eyes, thousands of lenses synthesised into one picture. In a tournament against US intelligence analysts with classified access, they won handily.

83,000 dated, specific probability predictions were collected.
Hedgehogs see everything through one lens and forecast worst.
Foxes collect perspectives and seek falsification of their own views.
They beat classified-access intelligence analysts handily.

“The Hedgehogs who knew one big thing whereas the good forecasters were the foxes who knew many little things”
David Epstein · 1:51:00

“described those people as having dragonfly eyes”
David Epstein · 1:52:00

#forecasting#hedgehog-fox#prediction#expertise

contrarian· 2

contrarian25:00

You're in a hammock of competence — and it's the biggest threat to you

Epstein tells Bartlett that the podcast's size and his own competence are the problem: he's in what Russ Roberts calls a hammock of competence, an area so comfortable and successful that improvement gets harder because everything disincentivises change. Getting better from there requires deliberate risk, which is complicated when a whole team depends on you. The answer is low-stakes practice that tests what might become a larger experiment, and pushing until you fail 15 to 20% of the time.

Comfort and success remove the incentive to change anything.
Risk for the leader is risk for everyone dependent on them.
Low-stakes practice de-risks a larger worthwhile experiment.
The zone of optimal push means failing 15-20% of the time.

“a rut of competence or what Economist Russ Roberts told me a hammock of competence”
David Epstein · 25:00

“if you're not at least like 15 20% of the time failing then you're not in your zone of optimal push”
David Epstein · 27:00

#risk#complacency#experimentation#growth

contrarian1:05:30

We only tell the Tiger story — even Serena's was a Roger story

Tiger Woods had a putter at ten months, was on national television at two and declared he'd be the next great golfer at three. Federer played basketball, rugby, skateboarding, football, handball, swimming and wrestling, and turned down moving up an age group so he could talk pro wrestling with friends. Epstein presented this research with Serena Williams in the second row, braced for demolition — she raised her hand to say her father was ahead of his time, having her do ballet, track and field, gymnastics and taekwondo.

Federer's mother, a tennis coach, declined to coach him.
The elite norm is a sampling period and delayed specialisation.
Serena Williams' upbringing was varied, not the assumed Tiger path.
Golf is a uniquely poor model for almost everything humans want to learn.

“they tend to follow the Roger path not the tiger path”
David Epstein · 1:06:00

“he had me do uh ballet track and field gymnastics Taekwondo”
David Epstein · 1:09:30

#specialisation#sport#sampling-period#federer

tactic· 2

tactic34:30

The morning habit ruining your day: starting with email

Epstein's most actionable prescription. The Zeigarnik effect means an unfinished task leaves a residue in your brain that impairs your transition into the next thing — and an inbox is unfinished by definition, so opening it first coats everything that follows. He tells Bartlett to decide the night before what single task would make tomorrow a good day, and start with that. Gloria Mark's tracking of heart rate variability and task switching found office workers checking email an average of 77 times a day.

An unfinished task leaves cognitive residue that follows you.
Inboxes are permanently unfinished, so they're the worst possible opener.
Decide tomorrow's one important task the day before.
Office workers were found checking email an average of 77 times a day.

“an unfinished task leaves like a residue in your brain”
David Epstein · 35:00

“I would suggest something like not starting your day with email or messaging”
David Epstein · 34:30

“people were checking email office workers were checking email and average of 77 times a day”
David Epstein · 56:00

#productivity#email#zeigarnik-effect#focus

tactic43:00

Mixed practice feels worse and moves kids from the 50th to the 80th percentile

Dozens of sixth-grade maths classrooms were randomly assigned to blocked practice (all of problem type A, then all of B) or interleaved practice (every type drawn at random). Blocked students progressed faster, felt happier and rated teachers highly. Interleaved students felt like they weren't getting it. On a transfer test the interleaved group blew them away, with an effect size equivalent to moving a child from the 50th to the 80th percentile — purely by rearranging the practice.

Blocked practice produces faster progress and happier learners.
Interleaved practice forces matching a strategy to a problem type.
Effect size equalled moving a student from the 50th to 80th percentile.
Around 90% of the best footballers grew up on futsal, roughly six times the touches.

“the inter Le group blew the block Practice Group away”
David Epstein · 44:00

“desirable difficulties these are things that make learning feel less fluent they are unpleasant they may slow you down much better for long-term retention”
David Epstein · 42:30

#learning#interleaving#practice-design#desirable-difficulties

Frameworks in this episode

Build a Reference Class of Analogies

Desirable Difficulties

Hire for Trainability, Not Baseline

Kind vs Wicked Learning Environments

Lateral Thinking With Withered Technology

Match Quality and the Sampling Period

The 10,000 Hours Critique

The Attention Residue Protocol

The Explore-Exploit Toggle

The Self-Regulatory Learning Cycle

Insights & moments

⚡Myth Buster· 3

The 10,000 hour rule came from 30 violinists and a hidden average