First principles thinking is usually described as breaking a problem down to its basic truths. In practice it means checking your assumptions against reality. You strip away borrowed language and habit until only the facts that carry weight are left.
Most thinking works by analogy or habit. That is fine most of the time. You shouldn't rediscover plumbing or food safety every morning. Inherited answers become a problem when conditions change and nobody notices that the old answer no longer fits.
There is an old story about a family recipe that says to cut the ends off a roast before it goes in the oven. Everyone does it because their parent did. Years later someone asks why, and it turns out the original pan was too small. The habit outlived its reason by decades. Nobody was foolish. Nobody had a reason to check. Working life is full of roasts like this.
We borrow answers for good reasons. Thinking everything through from scratch is slow and tiring, and borrowed answers let us get on with the day. The skill is knowing when to stop borrowing. Usually a feeling signals it: something is harder or more expensive than it seems it should be, and nobody can explain why.
The method starts by asking what must be true here. Name the actual constraint and the outcome you want. Then look at everything else and find the parts that are traditions pretending to be laws.
The core of it is telling a constraint from a convention. Reality enforces constraints. A bridge must carry its load, and a budget can't spend money it doesn't have. People enforce conventions. Reports are due on Fridays, and meetings last an hour. Many conventions are sensible. The difference is that people can change a convention by agreement, and nobody can change a constraint.
A useful exercise is to write down every rule that seems to govern a problem, and mark each one by who would stop you if you broke it: physics or people. The list usually shrinks fast. What remains is the real problem, often smaller than it looked.
Take the commute. The convention says everyone travels to the office every weekday at the same hour. The constraints are narrower. Some work needs a shared room, some needs silence, some needs a particular machine, and some needs none of these. Once you separate the two, many arrangements become possible that the convention hid. You might still keep the office, but now you are choosing it.
Many problems are protected by consensus more than by reality. A product is built a certain way because the industry builds it that way. A school day is organized as it is because schedules always looked like that. First principles thinking asks whether the routine serves its purpose or only keeps itself going.
This is uncomfortable because it removes excuses. If the old way isn't necessary, the current setup is a choice someone is making. You see this in companies adopting AI. The tools get blamed for weak results when the real problems are unclear goals and tangled processes. Asking what the work requires, before asking what the tool can do, shows where the drag comes from.
Secondhand thinking at scale
AI makes secondhand thinking effortless. A model can summarize the consensus, copy an expert's tone and suggest strategies based on the average of what has been written. That helps you get oriented. It is risky as a replacement for checking. If everyone asks machines to remix old assumptions, the future becomes a sharper copy of the past.
Fluent answers have a trap. A well-organized, confident response feels true, and the mind relaxes. That is when old assumptions slip through. A rough note written after visiting a customer can be worth more than a polished report, because the note came from contact with the world and the report came from consensus.
Two habits help. Ask the model to argue against its own first answer, or to list the assumptions inside it. Then check at least one claim directly: call the customer, visit the site, measure the thing. The model is good at mapping what has been said. Only checking tells you whether it still holds.
First principles thinking also needs humility. Many inherited practices exist because reality punished the alternatives. The old advice about a fence in the middle of a field applies: before tearing it down, find out why it was built. It may once have kept cattle away from a cliff. It may serve no purpose now. Finding out is the method in action. People who skip this step tend to rebuild the old solution from scratch, at great cost.
In practice, write down plain answers before anyone discusses them: what you are trying to improve, for whom, what would get worse if you succeeded, and what evidence would change your mind. Spoken answers drift and merge. Written ones can be compared. A team that spends twenty minutes writing separately, then reads each other's answers, often finds it was solving three different problems under one name. That discovery alone can save months of wasted work.
The same questions work outside business. Someone stuck in a career can ask what they actually need from work, as distinct from what their family or their younger self expected. The answer often opens options they couldn't see before, and it can make starting over feel like a sensible response to new information about themselves.
Larger plans need the same treatment. Any serious attempt to build a better society has to start from what people need and how they actually behave. A theory of what they should want is a poor substitute. That is part of why being a utopist takes patient, practical work: a good future has to be built from real materials.
The payoff is less unnecessary complexity. When a problem is cut down to its real parts, costs often drop and whole steps disappear. Next time a process at work feels heavier than it should, write down its rules, mark each one physics or people, and start by questioning the ones marked people.