Speed Is Not Creativity

A few weeks ago, I asked an AI to generate twenty variations of a book cover. In four minutes I had twenty designs. All technically clean. All somehow usable. None of them were good.

AI gets called a “co-creator.” It generates content, designs campaigns, develops visual concepts. Everyone is excited about the speed. Where a designer used to take days, AI delivers in minutes. Where a copywriter used to write three drafts, the machine produces thirty. More output in less time. That gets called creative.

It’s not creative. It’s fast.

Speed and creativity are not proportional. Most of the time they work in opposite directions. The best ideas I’ve seen in thirty years didn’t come quickly. They came after a phase where nothing worked. After the third discarded approach. After the conversation where someone said: That’s not it. What do you actually mean?

The moment an idea is born feels fast. But before that moment lie hours, days, sometimes weeks where nothing happens. Where you don’t know what you’re looking for. Where you have to understand the problem before you can have a solution. That lead time is not waste. It is the process.

An algorithm doesn’t have that lead time. It doesn’t have a moment where it doesn’t know what it’s looking for. It doesn’t search. It computes. It takes what exists and rearranges it. Fast, yes. Varied, yes. But the difference between variation and creativity is the difference between a permutation and a thought.

A permutation arranges existing elements in a new order. A thought brings forth something that wasn’t there before. That sounds abstract, but it’s very concrete. When a musician writes a melody that hits you, it’s not because the notes are new. The notes all exist already. It’s because he heard something he couldn’t articulate before, and now he can. That hearing doesn’t come from speed. It comes from attention. From time spent with the material. From failure.

The twenty covers the AI delivered had a shared problem. They all looked like covers that already exist. That’s logical. The AI was trained on existing covers. It knows what a cover looks like. It doesn’t know what a cover could look like that nobody has made yet. For that, it would need to want something. To miss something. To find something inadequate.

When AI gets celebrated as a “co-creator,” something similar happens to what happens with the word augmentation. A word that means something specific gets stretched until it fits everything. If every output is creativity, the word becomes meaningless. Then the spam filter is creative. Then autocorrect is creative. Then every computation that produces a new output is a creative act.

What gets lost is not the word. What gets lost is the understanding of what the process costs. Creativity costs time, frustration, uncertainty. It costs drafts that end up in the trash. It costs the courage to throw away something that works because it’s not good enough.

If speed is creativity, then slowness is a defect. Then the person who spends three weeks on an idea is inefficient. Then the process with no guaranteed outcome is a risk that should be replaced by an algorithm.

I use AI for things where speed matters. First drafts. Research. Structures that I then rework. That’s useful. But I don’t confuse it with the moment when something emerges that wasn’t there before. That moment needs something no algorithm has: the willingness to be slow.