The Why School (Training Second-Order Thinking)
One 'why' question per meal, answered the next day — how second-order thinking gets built.
- Difficulty
- Easy
- Time to result
- ~ongoing to results
- Steps
- 5
- Confidence
- 91%
Shah argues second-order thinking is the single most powerful predictor of success, that brains hate it because it's taxing, and that whether you enjoy it or hate it is set early. His answer is a deliberate training regime he calls the 'Why School': pose one deep origin/why question per meal, let the person go find the full history, and repeat. He identifies two childhood correlates — strategy games (builds second-order thinking) and physical games (builds rigour) — and says having both is the winning combination.
Origin
Kunal Shah's own coinage, offered as advice to parents but explicitly framed as a general training method. He notes the question of what actually produces great second-order thinkers in childhood 'is not fully discovered or discussed yet' — these are his loose empirical correlations.
Core principles
- 01Second-order thinking is known as the most powerful trait to predict someone's success.
- 02The brain hates second-order thinking unless trained into it — then it becomes a reward cycle rather than a hate cycle.
- 03'Why' is the deep question; it breaks a lot of things in your head.
- 04Strategy games build second-order thinking; physical games build rigour; both together compound.
- 05Origin stories force the second-order chain naturally, because history forces you to trace causes.
- 06AI is a co-pilot here, not a cheat — but the answer must be pushed to the second-order insight.
How to run it
- 1
Pose exactly one why question per meal
One question, not many. Shah's examples: why do humans wear jewellery? Why is it so expensive to advertise on the Super Bowl? Why is a microphone called a microphone, and why 'phone'?
Pro tip Origin-story questions ('what made earphones happen?') are the highest-yield type — they force a causal chain through history.
- 2
Give them until the next day
Do not accept an immediate answer. The delay forces real depth-seeking rather than a first-order guess.
- 3
Demand the full depth of history
The answer must go into the full historical depth of why the thing is so, not a surface definition. Going into that depth is where the second-order chain gets built automatically.
- 4
Allow AI, but force the second-order derivation
Shah explicitly permits ChatGPT — 'we have to assume that we are going to be all co-pilots of AI now.' The rule is that you must then derive the second-order insight: if this is true, where is the similarity of this found somewhere else? Give me an example of this pattern elsewhere.
Pro tip The AI supplies the first-order answer; the human's job is the transfer to a different domain.
Watch out If the exercise stops at the AI's answer, it trains lookup, not second-order thinking.
- 5
Pair with strategy games and physical games
Add strategy games (which correlate with second-order thinking) and physical games (which correlate with rigour and discipline). Shah's observation: having both is what lets people do really well.
Watch out This is Shah's loose correlation, not established science — he says so explicitly.
In the wild
A single why question, dispatched at a meal, answered the next day with the full historical depth — status signalling, portable wealth, mating display. The chain from 'why' to first cause is the exercise.
→ The learner builds the causal-chain habit that second-order thinking depends on, without ever being taught 'second-order thinking' as a term.
Shah's own adult version of the same exercise: asking what lab-grown diamonds do to diamonds, then finding the answer in what cultured pearls did to pearls (they destroyed pearls' royal status once everyone could have them).
→ A second-order investment thesis — the diamond profit pool gets destroyed after a temporary lab-grown profit window.
Common mistakes
Asking many questions instead of one
The method is one why per meal. Volume converts the exercise into trivia; depth on a single question is what breaks things open in your head.
Banning AI from the exercise
Shah explicitly disagrees with the 'you can't just ask ChatGPT' rule. The failure isn't using AI; it's stopping at AI's first-order answer instead of deriving where the pattern appears elsewhere.
Starting too late
Shah's claim is that whether the brain finds second-order thinking rewarding or hateful is set by early training. Later on, people simply hate doing it — the taxing effort never converts to a reward cycle.
Is it for you?
Best for
Parents, managers and self-directed learners deliberately building the causal-reasoning muscle that predicts long-run judgment.
Not ideal for
Situations demanding immediate decisions — this is a slow-compounding training regime, not a decision procedure.
From the transcript
“one of the things I recommend to uh parents to do with their kids, I call it the Wi-Fi school, is uh ask them a…”
“You just take one question and if you train them on why, they evolve at a very different rate”
“second-order thinking came from playing strategy games as a kid.”
“Brain Brain absolutely hates second-order thinking unless you've trained your brain with second-order thinking all the time that it looks forward to it.”
From the episode
Kunal Shah on winning in India, second-order thinking, the philosophy of startups, and more