Two people can meet the very same idea and reach for opposite tools to make sense of it. One trusts what they have lived; the other reasons down from what must be true. Neither is wrong — and knowing which one a moment is asking for quietly shapes both your career and the people you let close.
Watch any team meet a problem they have never seen before and you will usually spot two instincts in the room. One person rolls up their sleeves and starts trying things — ship a small version, see what breaks, adjust, try again. Another leans back and asks the slower question: what is actually true here, underneath all the noise? Strip this down to what we know for certain, and what does the answer have to be? Same problem, same table, two completely different ways in.
Those two instincts have names that sound like a classroom but live in almost everything we do. The first is induction (quy nạp) — you gather real experiences and let the pattern reveal itself. The second is deduction (diễn dịch) — you start from a principle you trust and reason your way down to the action. Most of us lean on one without noticing. And the quiet thing I keep running into, in engineering and in life, is that the people who go far are rarely the ones with the “right” instinct. They are the ones who can tell which instinct the moment is asking for.
It is tempting to picture the big turns — the product that finally worked, the promotion, the relationship that lasted — as single bright moments. Up close, they almost never are. They sit on top of a long, quiet accumulation: hundreds of small observations, a handful of principles earned the hard way. The visible leap is just the last inch of a runway you laid down, unseen, for years.
Two engines, not two camps
If you have ever worked out what your users truly want by shipping, watching, and adjusting — rather than guessing in a meeting room — you were thinking inductively. If you have ever cut a tangled debate short by asking “what do we actually know is true, and what follows from that?” — you were thinking deductively. They are not rival camps; they are two engines under the same hood.
You can see both in the people we hold up as world-changers. Steve Jobs and Mark Zuckerberg built what they built largely by noticing small, real experiences and connecting them into something larger — classic induction. Facebook did not begin as a two-hundred-page business plan; it began as tiny experiments in a dorm, each one teaching the next. Elon Musk works the other way. He is famous for First Principles Thinking: break a problem down to its most basic truths, then reason upward to a solution nobody had dared to try. That is deduction in its purest form.
Both are good. Both are necessary. The trouble starts only when we pick one and use it at the wrong moment — or push it so far it tips over into its own failure.
The turkey that trusted the routine
There is an old illustration that thinkers love, and it lands harder than any lecture. Picture a turkey on a well-run farm. Every single morning, the farmer arrives with food. Day one, day two, day one hundred, day nine hundred and ninety-nine — the same kind gift, the same time, like clockwork. From all that repetition the turkey draws a confident conclusion: the farmer loves me, life here is safe, and tomorrow will surely look like today.
Then comes the morning of day one thousand — the eve of a holiday — and the farmer arrives not with breakfast but with very different intentions. Every observation the turkey ever made was accurate. Its mistake was not in watching. It was in reading only the short-run data and missing the principle underneath it: what the farm was always for. Flawless induction on the surface; no deduction about the essence.
A record of what has happened never signs a contract for what comes next. Induction with no principle anchoring it can be perfectly logical, beautifully consistent — and still walk you calmly off a cliff. Always ask what the pattern is sitting on top of.
From zero you experiment; once you have built, you protect
Nowhere is the right gear clearer than in building a livelihood. When you are starting from nothing, induction is almost forced on you. With no capital, no track record, and no real market data, you cannot draft a flawless plan in an air-conditioned room. You make a small thing, put it in front of real people, and watch what actually happens.
Picture someone with an idea and very little to risk. On day one they make ten of something and offer it to the people nearby — and it is gone in fifteen minutes. On day two they try a variation, and it moves slowly. After a few rounds of watching and adjusting, a real pattern surfaces: who their people actually are, what those people genuinely want, what they will happily come back for. None of that was in any plan. It was earned, one small experiment at a time. That is induction doing exactly the job it is best at — discovering the real shape of a market by living inside it.
Plenty of companies we admire began the same way: not as a thick document, but as a series of small bets that taught their owners what to do next. The opposite move — reaching for deduction too early — is where a lot of bright starts quietly end.
Imagine pouring everything into a polished five-year plan and an expensive build before testing a single assumption, simply because a book said customers want a premium experience. Then opening day arrives and reality answers differently: the people actually nearby just want something fast, dependable, and fairly priced. The trap is not the plan; it is refusing to update it. When the spec and the world disagree, the world is the one that pays the bills.
A plan you have never tested against reality is not a strategy — it is a confident guess. The point of shipping early is not speed for its own sake; it is to let the world correct you while corrections are still cheap.
And yet the moment you have actually built something — capital, a team, a working product — pure experimentation turns dangerous, and deduction earns its keep. Now the principles matter. Do not put every egg in one basket; spread the risk. Keep a reserve for the lean season. Cap how much you can lose before you chase how much you might win. Induction found the opportunity. Deduction is what keeps the opportunity from being the last one you ever get.
Reading a person takes time, not a formula
The same two engines run quietly through our closest relationships. When you are getting to know someone, you have almost nothing but induction. No tidy bio and no sweet promise can tell you who a person really is. So you watch the small actions, over a long stretch of time.
The person who quietly straightens the cutlery for you before a meal. The one who thanks the server clearing the table. The one who shows up in the rain with medicine when you are unwell, or who works through a disagreement without ever raising their voice. No single moment is decisive. But gathered patiently over months, those small data points paint a whole portrait, and only then do you earn the right to conclude: this is a kind, steady person I could walk a long way beside. That is induction at its warmest and best.
Pushed to an extreme, though, induction curdles into prejudice. After a couple of painful endings, it is terribly easy to draw a sweeping rule — people always leave, no one can really be trusted, everyone is only in it for themselves. The conclusion feels hard-earned. It is not. It is built from a tiny sample and then stamped onto everyone who has not even arrived yet. And its quiet cost is cruel: it closes the door on the very happiness it was trying to protect.
A handful of endings is honest data about those endings. It is not a verdict on everyone who comes next. Let your experience make you wiser, not smaller — keep the lesson, drop the life sentence you handed to strangers.
Deduction has its turn here too, and it is a gentle, protective one. Once a relationship exists, a clear principle becomes the yardstick that keeps you whole. If you hold that love rests on trust and respect, then when the everyday reality drifts — constant checking of your phone, slow distance placed between you and your friends, control creeping into small daily choices — you have something steady to measure against. Deduction lets you draw a calm boundary instead of dissolving into the feeling of the moment.
But deduction abused can quietly kill the very thing it meant to guard. Plenty of people now import a borrowed “formula for perfect love” from the internet and apply it mechanically: they must reply within seconds, must always pick up the bill, must satisfy every material want, or else it is “not real love.” Those are principles lifted from a stranger’s highlight reel and pressed onto a real person with a real, particular life. Good relationships get dismantled by standards that were never theirs to begin with.
Two wheels of one cart
Step back and the whole picture resolves. Induction and deduction are not opposing camps you must choose between. They are two wheels of the same cart of thought. In building a livelihood, induction finds the opportunity and deduction protects the gain. In our relationships, induction lets you understand a person and deduction lets you keep your own values intact. Lose induction and you turn dogmatic — fluent in theory, deaf to what is actually in front of you. Lose deduction and you drift — no compass, repeating the same old mistakes in new clothes.
| Where it shows up | Induction (quy nạp) | Deduction (diễn dịch) |
|---|---|---|
| Career & building | Find the opportunity by experimenting and reading real signals | Protect the gain with principles — spread risk, keep a reserve |
| Understanding people | Read someone slowly, from small and consistent actions | Hold your own values steady against a clear principle |
| Its failure mode | A tiny sample hardens into a law about everyone | A borrowed formula judges a real, specific person |
| Without it | You turn dogmatic — fluent in theory, deaf to reality | You drift — no compass, the same mistakes on repeat |
This is also why the dramatic turns so rarely feel dramatic from the inside. The founder’s breakthrough is years of small observations finally clicking into place. The principle that saves you in a crisis was laid down quietly, long before the crisis came to test it. The big visible change is almost always the surfacing of a long, invisible accumulation. In my own years moving from writing code to leading people, almost nothing useful arrived as a flash. The judgement I lean on now is just a slow compression of a thousand small moments — plus a few principles I trust only because reality kept proving them right.
It is rarely “which kind of thinker am I?” The sharper question is “which engine is this moment actually asking for — and am I willing to switch?” The experimenter has to learn to hold a principle; the principled one has to be willing to go and look.
Why I file Elon Musk under deduction
People sometimes push back on placing Musk in the deduction column when he is so visibly hands-on. The reason is First Principles Thinking. Instead of asking “how has this always been done?” he breaks a problem down to its most basic truths and reasons upward from there.
Take rockets. He refused to accept the received wisdom that they are simply expensive and only governments can build them. He asked a more basic question: what is a rocket actually made of? Aerospace-grade aluminium, titanium, copper, carbon fibre — and what do those raw materials cost on the open market? He found the raw materials were a tiny fraction, on the order of a couple of per cent, of a finished rocket’s price. From that one principle about cost, he deduced the action: buy the materials, build the rocket in-house, and reclaim the rest. That reasoning is a large part of how SpaceX came to be.
Tesla rests on the same move. When the consensus said electric cars were unworkable because batteries were too expensive, Musk did not study the long history of failed attempts. He looked instead at the chemical components a battery is built from and reasoned that, in principle, the cost could fall dramatically with better manufacturing. The company followed from the principle, not from the precedent.
Stop asking “how is this normally done?” and start asking “what is this actually made of, and what does that make possible?” You will not always like the answer — but you will at least be reasoning from reality instead of from habit.
Put simply: induction starts from lived experience to find the rule; Musk’s deduction starts from the most basic truths and reasons toward a way to change reality. He does not ask “how does the world currently work?” so much as “how should it work, if it obeyed only the basic laws?” That posture is genuinely rare — the world has very few Musks. Most of us, myself included, begin the other way around: with the first customer, the first small sum saved, the first honest failure. That is the road of induction, and there is no shame in it at all. It is simply the road most real lives are built on.
Key takeaways
- Two engines, not two camps. Induction (quy nạp) reads patterns up from real experience; deduction (diễn dịch) reasons down from a principle. The skill is not picking a favourite — it is matching the engine to the moment.
- From zero, experiment; once built, protect. Lean inductive when you have little to lose — try, watch, adjust. Lean deductive once you have something worth keeping — spread the risk, keep a reserve, cap the downside.
- Mind the turkey effect. Short-run data can be flawless and still miss the principle underneath it. Always ask what your pattern is sitting on top of before you bet on tomorrow.
- People are read, not calculated. Understand someone slowly, from small consistent actions; protect your values with a clear principle. Don’t let a tiny sample become a law, and don’t judge a real person by a borrowed formula.
- Big changes are quiet accumulations surfacing. The visible leap is the last inch of a runway laid down, unseen, over years of small observations and a few hard-earned principles.
None of this is about turning yourself into a more efficient machine. It is about noticing which gear you are in — and whether the moment is quietly asking for the other one. I still catch myself experimenting when I should be holding a principle, or clinging to a principle when I should get up and go look. That is the practice, not a finish line. If you have followed along this far, I would genuinely love to know which way you lean by default — the one who tries first, or the one who reasons first — and the moment you learned, the hard way, that the other engine mattered just as much.