kinda, but the target is narrower

it’s about familiar patterns tricking you when the situation has quietly changed

exactly

autopilot is great on the route you trained for

bad news when the road now has a sinkhole and your brain goes “nah, normal Tuesday”

overconfidence is part of it

Epstein’s Chapter 10 frame is: expertise can narrow attention, so the expert sees the old pattern fast

in a kind world, yes

chess, basic surgery reps, fireground cues with clean feedback. speed can be earned there

speed can become pattern-lock

the brain grabs the closest known template before the problem has finished introducing itself

sometimes, yeah

the same compression that makes experts efficient can hide the weird detail that breaks the case

think airport security for your attention

an expert brain waves through familiar-looking bags because it has seen thousands

range is the annoying extra check: “open that one, the zipper is weird”

pretty much

it adds more reference cases, more analogies, and more chances to ask “what if this is the wrong category?”

yep, same neighborhood

public summaries tie Chapter 10 to Philip Tetlock’s forecasting work

foxes compare models. hedgehogs lean harder on one big model

that’s the failure mode

a strong model can become a sorting machine: evidence that fits gets in, evidence that doesn’t gets waved away

same

the Good Judgment Project tried to make forecasting less vibes-based: probabilities, Brier scores, updates, aggregation

yes

GJP recruited talented amateurs, trained them on forecasting habits, and combined predictions

reports say its top forecasters beat intelligence officers with classified info by about 30%

use it carefully

it’s about geopolitical forecasting tournaments, not “amateurs beat experts at everything”

they treated first impressions as drafts

they checked base rates, split big questions into parts, updated often, and stayed allergic to tidy stories

mine too, which is why they’re dangerous

a tidy story makes the familiar pattern feel proven before the evidence earns it

you see churn rising and instantly say “pricing problem” because last year it was pricing

maybe. but maybe onboarding changed, support got slower, competitors moved, or the customer mix shifted

yep, and sometimes it points at the wrong body part

before the meeting locks in, write 3 rival explanations

then ask one person from outside the specialty what category they think the problem belongs to

good sign

also make a “what would change my mind?” note before everyone starts defending their take

then let the expert cook

Range’s point is not to sand down skill. it’s to notice when skill is being used outside its home terrain

ask: are the rules stable, is feedback fast, and has this exact pattern repeated enough to trust recognition?

if not, slow the expert brain down and bring in breadth

perfect

don’t throw away expertise. make it show ID at the door

do it. 3 rival explanations, one outsider, one disconfirming test

that’s enough friction to catch a lot of fake certainty

go bother a tidy story for me. bye