What first principles actually mean

The Question

Elon Musk talks about it. Shane Parrish writes about it. It shows up in every serious guide to better thinking. First principles reasoning — the practice of stripping a problem down to its irreducible truths and rebuilding from there — has become one of the most widely cited frameworks in productivity, entrepreneurship, and decision-making.

And yet, when you actually try to apply “first principles thinking” to a real problem — a career decision, a business model, a daily schedule — the advice collapses into something frustratingly slippery. What are the principles, exactly? How deep do you have to go? How do you know you’ve gone deep enough, and not just deep enough to feel clever? And what do you do when the principles you find don’t point anywhere useful?

The term has taken on the flavor of a password — something people say to signal rigorous thinking rather than actually perform it. This post is an attempt to make it concrete again.

The Argument

First principles thinking is, at its core, a method for distinguishing between constraints that are actually fixed and constraints that just feel fixed.

Most of what limits us in any domain — career, business, design, daily life — is not physics. It is convention. It is inherited assumption about how things have to be done, accumulated over time into a kind of invisible scaffolding that feels load-bearing even when it isn’t. First principles thinking tears apart that scaffolding and asks a simple, ruthless question: what is actually true here, independent of how it’s always been done?

The clearest example is Musk on rockets. In 2001, when Musk looked into buying a rocket, he found the prices obscene — tens of millions of dollars for a single launch. He asked the question most people don’t ask: what does it actually cost, in raw materials, to build a rocket? The answer was roughly 2% of the market price. Everything between the materials cost and the purchase price was convention, margin, institutional overhead, and risk-aversion accumulated across decades of aerospace culture. That gap is where SpaceX was born.

The method, applied rigorously, looks like this:

1. State the problem clearly. Not “I need more time” but “I am currently spending 11 hours a week on activities that do not move my most important projects forward.” Precision matters here. Vague problems resist first principles; they just absorb your assumptions and hand them back dressed as insights.

2. List every assumption you are making. Not the obvious ones — those are easy to spot. The invisible ones. The ones you have never articulated because they feel like facts. Things like: this role requires being in the office, a business needs to reach profitability within 18 months, creative work can’t be scheduled, I’m not the kind of person who can do X. Write them down. All of them.

3. For each assumption, ask the hardest version of the question. Not “is this true?” — that’s too easy to answer yes. Ask: “What is the evidence that this is actually a physical or logical constraint, and not just a convention or a story?” A physical constraint is: the laws of thermodynamics. A logical constraint is: you can’t be in two places at once. Everything else is negotiable in principle, even if not in practice.

4. Rebuild from what survives. Most assumptions won’t survive the scrutiny. That’s the point. What remains — the genuinely fixed constraints — is the foundation. Now you can ask: given only these actual constraints, what is the best possible design for this problem?

The discipline is in step two and three. Most people skip straight to step four — they call their rationalization “first principles” because it sounds better than “I assumed my way to a conclusion I already wanted.”

The Counterpoint

Here is what the first-principles enthusiasts tend to understate: this is genuinely hard to do well, and the failure mode is pernicious.

The most common misuse is using first principles as a license to dismiss accumulated wisdom. Conventions exist because they encode hard-won knowledge. The reason aerospace rockets cost so much isn’t purely convention and institutional laziness — a large part of it is that rockets are extraordinarily dangerous, and the overhead cost reflects real safety processes paid for with real failures. Musk’s insight was correct, but it came paired with a tolerance for failure that not every application of first principles can afford. Rebuilding a manufacturing process from first principles is different from rebuilding a hospital protocol from first principles. The cost of being wrong is not symmetric.

The second failure mode is more personal and harder to admit. Most people who invoke first principles are doing the opposite of thinking from scratch. They are working backward from a conclusion they already hold — a business they want to start, a relationship they want to leave, a constraint they find convenient — and then strip-mining their assumptions until they find a chain of logic that justifies it. Feynman’s warning is precisely about this: the easiest person to fool is yourself. First principles thinking, done carelessly, gives you a sophisticated vocabulary for sophisticated self-deception.

There is also a quieter cost. The person who questions every convention from first principles has to spend enormous cognitive energy on things that most people can simply accept and move on from. Sometimes the most important move is to borrow the convention, recognize it as convention, and go do something harder. Not every constraint in your life is worth dismantling. Some are worth accepting in order to focus your attention on the one or two that actually matter.

What To Do With It

Pick one belief you hold about your work or life that you have never seriously questioned. Not an obvious one. Something you treat as settled — a limit on your career trajectory, a rule about how your industry operates, a constraint on what is achievable for someone with your background or resources. The kind of belief you would call a “fact” if someone asked you why you weren’t doing something.

Write it down.

Then ask this exact question: what would have to be true for this to be false?

Don’t answer quickly. The quick answer is always “nothing, it’s just true.” The slow answer is where the work is. What evidence would you need to see? What would someone who held the opposite belief point to? What would have to change in the world, or in your situation, for the constraint to dissolve?

Sit with it. Stay uncomfortable. Notice the difference between “I’ve thought about this carefully and it’s genuinely fixed” and “I’m uncomfortable with what happens if it isn’t fixed.”

That discomfort — the specific, directional kind that points at your own position rather than someone else’s — is first principles in practice. Not a framework. Not a brand. A way of holding your assumptions loosely enough to actually examine them.

That is harder than it sounds, which is why almost nobody does it and almost everyone talks about it.

Sources

• Richard Feynman — The Feynman Lectures
• Shane Parrish — Farnam Street