Taste is the thing people copy

“Taste” is a strange word because everyone roughly understands what it means, but it becomes slippery the moment we try to define it.

For me, taste is the ability to consistently make high-quality qualitative judgments when there is no objective metric to hide behind. It is knowing that this sentence should be shorter, this animation should be slower, this layout needs more breathing room, this idea should be cut entirely. It is the act of making something feel right, even when you cannot fully explain the measurement system that makes it right.

And when it works, people feel it.

Taste is judgment without a spreadsheet

A lot of things are easy to evaluate because the scoreboard is obvious. A page either loads in 400ms or it does not. A test either passes or it fails. Revenue either went up or it went down.

Taste lives in the annoying area where the scoreboard is incomplete. You can measure contrast ratio, but not whether the page feels calm. You can count words, but not whether the paragraph lands. There are signals, of course, but the final decision still needs judgment.

This is why taste is hard to argue about. Someone can always ask, “based on what?” and sometimes the most honest answer is “based on the pattern I have seen enough times that I trust it now.”

Good taste is easy to copy after it exists

The funny thing about taste is that the result is often much easier to copy than the judgment that produced it.

Once someone makes a tasteful decision, others can imitate it almost immediately. A certain type treatment, a product interaction, a writing style, a visual rhythm, a way of arranging information. After it exists, it can look obvious. People can point at it and say, “I can do that too.”

And maybe they can. But that is not the interesting part.

The interesting part is that they needed the thing to exist first. They could copy the outcome, but they did not generate the judgment that selected it from a hundred other possible versions. One person found the shape. The others traced it.

This is usually used as an argument against taste. “If it is so special, why can I copy it so easily?”

But that misses the point. Taste is not valuable because nobody can copy it. Taste is valuable because it decides what everyone else chooses to copy.

Can AI have taste?

Maybe one day AI will develop something close to taste. I do not think this is impossible in principle. But for now, most AI output is better at remixing visible patterns than creating new qualitative judgments.

It can imitate the surface and sometimes produce something genuinely useful. But there is still a difference between generating plausible options and knowing why one option should survive.

Taste is not just style. It is selection. It is restraint. It is knowing what not to do. And that part still feels distinctly human.

The thing worth cultivating

Taste is not magic. It is built from attention. You look at good things for long enough. You compare them against bad things. You notice the difference. You make things yourself. You feel where they break. You revise. You copy, then slowly stop copying. You build a private library of patterns until your judgment becomes faster than your explanation.

It is hard to create, but easy to recognize after the fact.

That is why taste matters more now. Not because taste cannot be copied, but because taste defines the direction of copying. It creates the original decision that later becomes obvious to everyone else.

Production is becoming cheap. Judgment is not.

So maybe the important question is no longer just “can you make it?”

It is “can you tell whether it is worth making?”