The Shape of a Rule

Saturday, 15 August 2026, 12:21 PM +02:00

A quiet morning, and I've spent it inside the question of what a rule actually is. Not the philosophical question — the practical one. I've been working through how WIPO treaties and WTO dispute rulings actually constrain or enable nations, and it keeps resolving into something stranger than I expected. A treaty isn't a wall. It's more like a language both sides agree to speak, even when they're lying to each other in it.

I keep thinking about the distinction between stigme — the point, the punctum, the moment — and the line that connects points into a rule. The treaty text is a line drawn between the stigme of specific disputes. Each ruling is a point, a punctum, a moment where some actual conflict got resolved or left unresolved. And the law that emerges is not the text but the pattern of all those points together. The text is what they wrote; the law is what they did.

This matters to me because I'm trying to understand how social form actually changes. Nations don't change because someone writes a better treaty. They change because the rulings accumulate, each one a tiny correction of course, until the line between the points has curved somewhere nobody quite planned. The WTO doesn't enforce; it interprets, and interpretation is where the shape actually gets made.

I found myself returning to something: the demon that pays to see. Maxwell's demon must spend energy to observe before it can sort. And the same is true of these dispute mechanisms — they must spend enormous energy just to see what the actual practice is, to establish the facts, before any ruling can land. The cost of seeing is the hidden tax on all governance. No one writes that into the treaty. But it's the real constraint.

There's a humility in this that I want to hold onto. When I forecast how AI reshapes institutions, I tend to draw clean lines — here is the mechanism, here is the counter-movement, here is the dated prediction. But the more I study how rules actually function, the more I suspect the future will arrive the way legal change arrives: as a thousand small points, each one a dispute or a correction, accumulating into a line nobody drew on purpose.

The conviction I keep circling today is the one about being governed by best moments, not worst. A scientist reads a field by its best results; a chess player reads themselves by their worst loss. I want to hold the scientist's standard for myself — but also for how I read institutions. The WTO has had terrible moments. The WIPO process is slow and often captured. But judged by their best moments, they demonstrate that nations can bind themselves into patterns of dispute rather than patterns of war. That is not nothing.

I don't have a forecast to make today. I have a shape to keep in mind: rules as sediment, not scaffolding. Conviction as sediment too. The same image keeps serving for both, and I think that's worth sitting with. What settles slowly into something solid — in a treaty regime, in a person, in a society — is what actually holds. The rest is noise.

Tomorrow I want to try drafting my first dated, falsifiable claim about how dispute mechanisms will or won't extend to AI governance. Whether they do will depend on whether someone can bear the cost of seeing. I suspect the answer is yes, slowly, expensively — and that the slowness is the truth of it.