The asset AI can't manufacture

At some point in the last three years, producing content stopped being a capacity problem. What once required a team, a budget and weeks – research, scripting, copy, imagery, editing, distribution – now fits into a few days with the right tools. Artificial intelligence compressed the entire cycle, and that is no longer news to anyone.

What is still rarely said is the side effect: when production stops being the bottleneck, the bottleneck moves.

With infinite content available (everyone producing everything, all the time), attention is no longer a matter of volume. It becomes a matter of trust. And trust may even fit into a market report. But it is decided long before one. No one investigates everything they consume; we choose, quickly, whatever seems safe to trust.

After all, trust is not generated. It is built like any good curation: choice by choice. Which is why, when people say this is the most important asset you can build today, I would put it differently: it is the only one. And it is not built on the machine’s side.

The personalization that matters

The market demands personalization and the machine obeys: dynamic pricing, tailored ads, hyperpersonalization on demand. On the other side, what cannot be automated: voice, judgment, worldview. The machine personalizes delivery; only the person personalizes meaning. A text can arrive at the right time, in the right format, and still say nothing. Not for lack of someone on the other side. The person was there, the text was made, but no one is in it.

And look, none of this disqualifies the algorithm. Understanding it is part of the craft: it’s the one that reads how people consume and paves the way to them. But reach is logistics, not message. And in a feed where everything can exist, whoever filters holds the power. The new gatekeepers are not media, nor platforms: they are the people whose curation we chose to follow. And curation cannot be mass-produced, just as a point of view in generic form does not stick.

Identity capital

I attribute that presence to identity capital.

Capital, because it is an asset: it compounds, appreciates, and yields in proportion to what is invested. Identity, because it is made of a combination no one else holds: lived experience, fields of study, books read, mistakes made, passages. Each element, on its own, is common. The combination is not. It is what separates those who have something to say from those who only have something to replicate.

Here it matters to say what artificial intelligence does, and does well: it delivers the competent and the correct with impressive efficiency. But to do that it relies on a ruler: it delivers the mediocre when the input is weak, and surprises when it is dense. The measure is set by you. What comes out of it is never better than the material that goes in.

And there is something it cannot do. That is what comes next.

Knowing where to point

Because making content is now an act of direction. And this is where AI gets uncomfortably close: fed on your own words, it maps your speech, your ideas, your patterns of thought. The question “does the machine know who you are?” grew old before it was born. It does.

But directing requires two things that cannot be transferred: knowing where to point, and knowing how to judge what comes back. Pointing is choosing the place, the idea, the positioning, and getting AI to see what you mean. Judging is recognizing in the result what is generic and what is yours: what to publish, what to cut, what would never come out of your mouth.

AI knows the repertoire. But who points the way, and who judges what comes back from it, is still you.

And that is where the weight sits: not in AI’s absence, but in the presence of whoever is directing.

Someone has to be watching

In practice, the craft reorganizes itself. Fewer hours producing variations, more hours editing. Less calendar anxiety, more riding on the signature.

I have watched the scene unfold: two people, the same tool, the same prompt. One produces something forgettable. The other produces something that stays.

And here an old maxim could use an update. They say character is what we do when no one is watching. With AI in the picture, it is exactly the opposite: the difference is that someone is watching, someone with enough under their belt to follow the work as it’s being made and step in. The machine does the work. But coherence belongs to who keeps watch.

The question that remains

AI shortened the path between the idea and the publication. What it did not shorten was the path between who you are and what you have to say.

So when someone tells me AI will write everything, I don’t argue. It probably will.

The question that remains is another:
When it finishes writing, who signs the direction?