What physics teaches about building companies
Field notes on transferring scientific training to product work.
I spent years searching for new particles in the debris of proton collisions at CERN. Most of that time, I found nothing — and learning to find nothing precisely turned out to be the most useful skill I carried into building companies.
In particle physics, you don't get to argue a result into existence. You state what you expect to see, you measure, and the measurement decides. A beautiful hypothesis and an ugly one are treated exactly the same by the data. Building products rewards the same habit: the market is the detector, users are the measurement, and no amount of conviction substitutes for what they actually do.
The work was mostly signal against background — pulling a faint pattern out of an overwhelming amount of ordinary noise. That is also the founder's daily problem. Every day brings more requests, opinions, and possible directions than any team can act on. The discipline is the same in both places: define what would count as signal before you go looking, so you're not fooled by noise that happens to be shaped like good news.
Physics also made me comfortable with uncertainty as a working condition rather than a problem to escape. You never report a single number; you report a value and how sure you are of it. Founders live in the same grammar. The question is rarely “is this right,” but “how confident am I, what would change my mind, and what's the cheapest experiment that would tell me?” A plan you can't imagine being wrong is not a plan — it's a hope.
Then there is scale. The experiment I worked on was built and run by thousands of people across dozens of countries, and it still had to produce one coherent result. It taught me that large systems are legible if you insist on making them so: clear interfaces, honest documentation, and the assumption that the next person needs to understand what you did without asking you. Later, building safety-critical software in the automotive industry, that instinct became non-negotiable. It carries directly into a company — a product is just a system that many people, most of whom will never meet you, have to be able to trust.
The temptation, coming from research, is to over-value rigor and under-value shipping. Physics can afford a ten-year measurement; a company usually cannot. So the real transfer isn't the caution — it's knowing exactly how much rigor a given decision deserves, and spending it there and nowhere else. Research-grade thinking about the few questions that matter, production-grade discipline about the rest, and the judgment to tell which is which.