Time was a filter

Perhaps we have forgotten that some things are supposed to take time. There is a line about the Ents in The Lord of the Rings that has stayed with me for years. Treebeard explains that Old Entish is a slow language, and then adds the part that actually matters, which is that "we never say anything unless it is worth taking a long time to say." There is something almost absurdly incompatible with the AI age in that idea.

Today we optimise for the opposite. Faster answers, faster code, faster documents, faster analysis, faster decisions. Most of that is genuinely useful and I use it every day, so this is not an argument against any of it. But somewhere along the way we started treating every kind of friction as a defect, and I am not convinced that all of it was.

Sometimes the time it takes to write something is part of the thinking. Struggling with a problem for an afternoon is often how you discover that you were asking the wrong question, and rewriting the same paragraph five times is not wasted effort, it is how an unclear thought slowly becomes a clear one. And when you cannot produce something immediately, you tend to end up asking the more uncomfortable question, which is whether it is worth producing at all.

That question used to answer itself. Not because we were disciplined enough to ask it, but because the cost of doing the work asked it on our behalf. An idea had to survive a few hours of effort before it became a document, and plenty of ideas did not survive. We rarely noticed, because we never saw the ones that quietly died along the way.

AI changes that equation. A vague thought can become ten pages in minutes, and an architectural assumption can turn into diagrams, an implementation plan and working code before anyone has really had time to sit with it. That is extraordinary leverage, and I would not want to give it back. But it also removes a filter that most of us never recognised as a filter, which was simply time.

What interests me is not the speed itself, but the distance that has disappeared between having a thought and turning that thought into something real. That distance used to hold things. Doubt, a second opinion, a night of sleep, the slow realisation halfway through implementation that the premise was slightly off and that everything built on top of it would have to move with it. None of that was ever written into a project plan. It happened because there was room for it.

I have written before about how AI can help build an internally consistent world around a mistake, where the documentation agrees with the code, the code agrees with the tests, and all of it is wrong together. This feels like the other half of the same problem. The whole structure now gets built before anyone has had time to doubt the first assumption. And a quickly produced document looks exactly like a slowly produced one, because nothing on the page tells you how long anyone actually spent thinking about it.

None of this makes slowness a virtue. Plenty of slow work is bad work, and anyone who has sat through a three hour meeting knows that taking a long time to say something is no guarantee that it was worth saying. The Ents were not wise because they were slow. They were disciplined because they were slow, and eventually that constraint turned into a value. We have lost the constraint, which means the value is now optional.

So perhaps the thing worth being careful about is building systems where nothing takes long enough for anyone to reconsider it. If the pause is no longer forced on us, it has to be chosen instead. Review that is allowed to reach the premise and not only the syntax. Someone who can say that the whole thing is pointed in the wrong direction. The habit of leaving something overnight, not because slow is better, but because tomorrow you might disagree with yourself.

AI does not only accelerate good thinking. It accelerates the moment between having a thought and making it real, and that moment may have been worth more than we realised. The question was never really whether we can produce it. It is whether it was worth taking a long time to say.