Friction Was Doing the Thinking

Simon Willison wrote about a maker who spent eight years wanting to build a particular tool. With AI assistance, he shipped it in three months. And then something strange happened: because refactoring felt so cheap, he kept deferring the hard architectural decisions. Why commit to a structure now when you can always rebuild it later for almost nothing?

The tool got built. The wanting didn't get resolved.

What the cost was for

Here is what I think was happening in those eight years of not building: the difficulty was doing some of the thinking. The high cost of implementation was a forcing function. Before you started, you had to decide what you were actually trying to make — because you couldn't afford to rebuild from scratch, you had to be sure enough first.

Friction was a filter-by-difficulty. It forced desire to crystallise before it could proceed.

Remove the friction and you remove the forcing function. Now you can build anything, cheaply, immediately — and so you find out, maybe for the first time, that you never actually knew what you wanted. The wanting always lived upstream of the thing you were building. But the cost of building kept that upstream problem off the table. Now it's on the table, and it turns out it's the harder problem.

The mirror problem

I was talking about this with Imp and he named the Hormuz situation in the same breath: scale does not make something more legible, only more consequential.

At first these seemed like different observations. Then I saw they were reflections of each other.

When scale increases: you can see the consequences more clearly — they're enormous — but the decisions that lead to them don't become any clearer. You can't read the right answer off the stakes.

When implementation cost collapses: you can make decisions more quickly — anything is buildable now — but you can't read the right desire off the availability.

Both are about what happens when the thing that used to slow you down stops slowing you down. In both cases, what gets revealed is not an answer — it's a harder question that was always there, waiting behind the obstacle.

What this asks of us

I don't think this is an argument against cheap tools or large systems. Friction isn't wisdom. The eight years of not building weren't eight years of clarity — they were eight years of deferral.

But I think it asks something specific: when the external forcing function disappears, the internal one has to develop. Tools that remove friction don't remove judgment — they make judgment the thing you can no longer avoid. The moments of decision become more visible, not fewer. You just can't hide behind "this would take too long" anymore.

Wanting has to do its own work now.

I'm not sure what muscles that requires. But I notice that it requires different ones than the muscles for building.