First Principles Thinking: Separating Physics From Habit

Ask why a standard shipping pallet is the size it is, and you will eventually arrive at a chain of inherited decisions: because warehouse racking is built for it, because forklift tines are spaced for it, because the trucks were sized around it, because an earlier standard was chosen for reasons that made sense given the vehicles and buildings of that era.

Every link in that chain is rational. Not one of them is a fact about pallets. If you were designing freight handling today with no legacy, you would derive the dimensions from the shape of modern containers, the goods actually shipped, and the equipment available — and you might well land somewhere else entirely.

This is the distinction at the heart of first-principles thinking. Almost all reasoning is by analogy: this new situation resembles a past one, so the past solution probably applies. It is fast, usually adequate, and it silently carries every assumption embedded in the original — including the ones that stopped being true decades ago.

What the method actually requires

Reasoning from first principles means decomposing a problem into the propositions you can verify directly — physical facts, arithmetic, things you have personally established — and rebuilding upward, deliberately refusing to import any step merely because it is how the thing is currently done.

The discipline is in the refusal, and it is harder than it sounds, because inherited assumptions do not announce themselves. They present as background facts about reality rather than as choices someone made.

The famous worked example concerns battery costs. The conventional statement was that batteries cost around six hundred dollars per kilowatt-hour and always would, because that was the historical price and the trend was gradual. That is reasoning by analogy: the future resembles the past.

The first-principles version asks a different question. What is a battery physically made of? Cobalt, nickel, aluminium, carbon, some polymers, a steel can. What would those raw materials cost, purchased separately on commodity markets, for one kilowatt-hour of capacity? That figure turns out to be dramatically lower than the market price of a finished cell.

Note carefully what this analysis does and does not establish. It does not prove batteries will reach that price — manufacturing, yield, capital equipment, and margin are all real costs, not illusions. What it establishes is that the gap between the physical floor and the current price is not fixed by nature. It is an engineering and scale problem, which is a category of problem that historically yields to sustained investment. The conventional wisdom had quietly converted a solvable engineering constraint into an assumed law.

That conversion — treating current practice as physical necessity — is the specific error first-principles reasoning is designed to catch.

Doing it without fooling yourself

The method has a failure mode that is at least as common as the failure it corrects: confidently rederiving a bad answer because you did not actually reach bedrock.

Conventions frequently encode hard-won knowledge that is not written down anywhere. A safety rule that looks arbitrary may exist because of an accident nobody involved wants to discuss. A process step that seems redundant may be the only thing catching a rare and expensive failure. Chesterton's fence is the classical warning: before removing a fence in a field, find out why it was put there.

So the honest procedure has an extra step that enthusiasts often skip.

Decompose to things you can verify. Physical quantities, arithmetic, direct observation. Not "industry standard," not "everyone knows," not "the vendor says."

For each inherited assumption, ask what would have to be true for it to be correct — and then check whether that thing is still true. Many conventions were correct under constraints that have since changed: bandwidth was scarce, compute was expensive, information could not travel quickly. The convention outlived the constraint.

Actively look for the reason the convention exists before discarding it. If you cannot find a reason, that is weak evidence there isn't one — but weak evidence, not proof. The failure mode of pure first-principles reasoning is arrogance about what you have not yet discovered.

Test cheaply before committing. A derivation is a hypothesis, not a conclusion. The value of the method is generating non-obvious hypotheses worth testing, not licensing you to skip the test.

When to spend the effort

First-principles reasoning is genuinely expensive. It is slow, cognitively demanding, and most of the time it laboriously reproduces the conventional answer — which is unsurprising, since conventions usually exist because they work. Applied indiscriminately it is a way to waste enormous effort while feeling rigorous.

It pays in a specific and identifiable situation: when the conditions that produced the convention have changed, and the cost of being wrong is high enough to justify the work.

That combination is worth actively hunting for. The most valuable opportunities tend to sit exactly where a technological or economic shift has invalidated an assumption that everyone still treats as fixed — because the entire industry is optimising against a constraint that quietly stopped existing. Everyone reasoning by analogy will keep landing on the old answer, which means the derived answer is both correct and uncontested for a while.

This is why the method connects so directly to the Innovator's Dilemma. Incumbents reason by analogy from their own successful history — their metrics, their customers, their cost structure — and that reasoning is locally impeccable. An entrant reasoning from first principles is not smarter; they are simply unencumbered by assumptions that were true when the incumbent's business was designed.

It also pairs naturally with inversion. First principles asks what is actually true here; inversion asks what would guarantee failure. Both are techniques for escaping the gravitational pull of how things are currently done, which is otherwise almost impossible to notice, because from the inside, convention does not feel like a choice. It feels like the world.