Controversy became a moat.
Why the companies that won tech’s most contested markets won because of their controversy, and why most companies only ever pay its tax.
The claim
Technology now produces things faster than society can decide whether they should exist: a car with no one at the wheel, a market on the next election, an AI that decides what it will refuse. Each one lands in an argument before it lands in a market.
At the same time, AI has made products cheap to build and cheap to copy. What a company can make matters less every year. What matters is whether it can win the argument its product starts, and act on that argument in every decision.
What the market already shows
Look at who came to dominate tech’s most contested markets: Waymo, Palantir, Kalshi, OnlyFans. The usual reading is that they won despite their controversy. I think they won because of it.
In each market, the companies best equipped to compete stayed out because the reputational price was too high. Google walked away from the Pentagon’s Project Maven after thousands of its employees protested; Palantir took the work and still runs it. GM spent $10 billion on Cruise before shutting its robotaxi program down, while Waymo went from 50,000 paid rides a week to 500,000 in under two years.
The price of entry was a belief, not a technology.
But a moat is also a tax, and it never stops being collected. Most companies only ever pay the tax. They defend their belief, survive each fight, and never let the belief decide anything.
Borrowed and earned
Borrowed controversy is provocation bolted onto a business as advertising. Cluely built its brand on “cheat on everything,” then pivoted to an ordinary note-taking app. Borrowed controversy burns off.
Earned controversy follows from a belief the business is actually built on, and the company will pay to hold it. Anthropic refused to drop two red lines with the Pentagon, on mass surveillance and autonomous weapons. It was labeled a supply-chain risk, and its app went to number one in the App Store. Earned controversy compounds.
A belief only counts when a reasonable person could disagree with it and it actually decides what the company does.
The moment it’s tested
Every belief eventually meets a divergence moment: a decision the company can’t avoid that pulls against what it says, in public, with a price on both sides. Companies hold, fold, or sell. OpenAI announced an adult mode for ChatGPT under the principle of treating adults like adults, then shelved it after its own advisors warned against it and its age checks kept misclassifying minors. The belief was announced without the product that could hold it.
Why philosophy
When a question is genuinely unsettled, argument decides it, and whoever makes the best case shapes what people come to accept. That’s why Karp, Thiel, and Amodei are taken seriously by people who disagree with them: they argue their positions in public instead of hiding them. Philosophy is the discipline of building those arguments, which is why it sits at the center of this work and not on its surface.
What I do with it
I find the belief underneath a company’s controversy and build the strongest honest case for it. Then I work out everything that belief should decide: what to build, remove, and refuse, which fights to take, and who to lose. Then I design that company, so the belief stops being something it defends and becomes something it fights with. I call it Conviction Design.
