LLM Book Knowledgebase

First-Principles Thinking

7 min read

First-Principles Thinking

First-principles thinking is a method of reasoning that begins from foundational truths — the most fundamental, well-established facts about a domain — and constructs conclusions from that base rather than from analogy with existing solutions. It is contrasted with reasoning by analogy, which takes existing solutions as templates and iterates on them.

The method is associated with Aristotle (who identified “first principles” as the irreducible starting points of any domain of knowledge), but in contemporary usage it is most closely associated with Elon Musk, who made it the explicit foundation of his engineering and business approach.

The Core Distinction

Musk’s articulation is the clearest available:

“The normal way we conduct our lives is reasoning by analogy. That means we do something because it’s similar to something else, or what other people are doing. When you think this way, you only get slight iterations. It’s easier to reason by analogy rather than from first principles, so that’s what we do most of the time. And in most of life, we should reason by analogy. Otherwise, mentally, you wouldn’t be able to get through the day. It would be too much thinking. But for important things, that kind of thinking is too bound by convention or prior experiences. You will hear, ‘It’s always been done this way,’ or ‘Nobody’s ever done it.’ That is a ridiculous way to think.” — The Book of Elon, Eric Jorgenson

The key insight is that analogical reasoning is cognitively efficient but produces incremental improvements only. When a problem space has been misconfigured by historical accident — when costs are high because they were always high, not because they must be — analogical reasoning will perpetuate the misconfiguration indefinitely.

The Battery Example: First Principles Applied to Cost

Musk’s most frequently cited application of first-principles thinking concerns Tesla battery costs:

“People assumed batteries for electric vehicles would always cost $600 per kilowatt hour. The first-principles approach to battery costs is this: What are the batteries made of? What are the materials that make up the batteries? What is the market value of those material constituents? It’s got cobalt, nickel, aluminum, carbon, some polymers for separation, and a steel can. Okay, what if we bought that amount of material at the London Metal Exchange? What would each of those materials cost?” — The Book of Elon, Eric Jorgenson

The result of this analysis was that the raw materials in a battery pack cost far less than 600 figure as given.

The Rocket Example: The Idiot Index

The same analysis applied to SpaceX’s founding:

“Most people think, ‘Historically, all rockets have been expensive. Therefore, in the future, all rockets will be expensive.’ But that’s not true. A rocket is made from aluminum, titanium, copper, and carbon fiber. Break it down further and ask, ‘How much of each material is used? Now, what is the cost of all these raw components?’ If you have them stacked on the floor and could wave a magic wand to create the rocket, what would the cost of the rocket be?… For rockets, that turned out to be a relatively small number, well under 5 percent of the current cost.” — The Book of Elon, Eric Jorgenson

From this analysis came what Musk calls “The Idiot Index”:

“How much more does a finished product cost than the cost of its materials? If a part or product had a high Idiot Index, we could cut the cost with more efficient manufacturing techniques.”

A high Idiot Index does not mean the product is impossible to make cheaply. It means the current manufacturing process is inefficient relative to what physics allows.

The Physics Test as Floor and Ceiling

Musk’s consistent application of physics as the ultimate arbiter:

“Physics is law. Everything else is a recommendation. I’ve met many people who can break the laws of man, but I have never met anyone who could break the laws of physics.” — The Book of Elon, Eric Jorgenson

“Physics does not care about hurt feelings. It cares about whether you got the rocket right.” — Elon Musk, Walter Isaacson

This framing has two implications. First, it sets a ceiling: if something would violate conservation of energy, it’s not going to work. Second, it sets a floor: if physics allows something, there is no fundamental reason it cannot be achieved — only engineering challenges, which are soluble in principle.

How to Apply It: The Axiomatic Base Method

Musk describes the process:

“Break something down to the most fundamental principles. Start by asking: What am I most confident is true at a foundational level? That sets your axiomatic base. Then you reason up from there. Then you check your conclusions against the axiomatic truths.” — The Book of Elon, Eric Jorgenson

The practical steps:

  1. Identify the domain’s fundamental constraints (physics, biology, economics)
  2. Ask what the theoretical minimum is given those constraints
  3. Calculate the “Idiot Index” — the gap between current performance and the theoretical minimum
  4. Design a path toward the theoretical minimum, deleting the sources of inefficiency

Historical Antecedents

First-principles thinking was not invented by Musk. Leonardo da Vinci practiced a version of it:

“A commitment to test knowledge through experience, persistence, and a willingness to learn from mistakes. By rejecting dogma and superstition, Leonardo took responsibility for his own search.” — Da Vinci Decoded, Michael J. Gelb

“‘No one should imitate the manner of another, for he would then deserve to be called a grandson of nature, not her son.‘” — Da Vinci Decoded, Michael J. Gelb

Leonardo’s dimostrazione principle — testing knowledge through experiment rather than accepting authority — is first-principles thinking applied to natural philosophy. He refused to accept received anatomical knowledge when it conflicted with his direct observations of dissected bodies.

Darwin’s method was analogous: he refused to accept the fixity of species as a given and built his theory from the direct observation of variation in nature, catalyzed by Malthus’s model of competitive pressure.

Limitations and Costs

First-principles thinking is expensive. Musk himself acknowledges:

“I try to be hyperrational. If the reasoning fits, and you’re not violating the laws of physics, that’s the thing you should try to do.”

But this hyperrationality has costs:

  • It is time-intensive: most problems cannot receive full first-principles analysis
  • It can produce overconfidence in the face of unknowns that don’t appear in the foundational model
  • Musk’s own timing errors (off by 200% in early SpaceX schedules) suggest that first-principles analysis, while useful for costs, may be less useful for schedule estimation, where human and organizational factors dominate

First-principles thinking is most powerful for physical/material cost problems. It is less powerful for problems involving human behavior, culture, politics, and complex systems where the "fundamental principles" are harder to identify and the Idiot Index has less meaning.

Connections to Other Concepts

First-principles thinking connects directly to the-algorithm-engineering-process — Musk’s five-step manufacturing method begins with questioning requirements (step 1), which is itself a first-principles exercise. It also underlies reality-distortion-field: Jobs’s RDF was partly a social technology for forcing his teams to recalculate what was truly necessary versus what was merely conventional.

Darwin’s scientific method — building from observation to general laws — is a biological science version of the same epistemological commitment: don’t trust received wisdom; go to the phenomena and reason up.

Backlinks