Taste Is the Last Moat

For most of modern work, the bottleneck was production. You needed someone to write the thing, build the thing, draft the thing, code the thing. Talent was the capacity to produce acceptable output. That is no longer a bottleneck. We have crossed, in the last two years, into a world of infinite first drafts.

What remains scarce is the ability to tell which draft is good.

This is not a romantic claim. It is an economic one. When production becomes cheap, the value in the chain moves to the next scarce input. In the 2010s, the scarce input was distribution - which is why platforms captured most of the margin. In 2026, the scarce input is taste. The ability to look at twenty adequate options and know, quickly, which one is the right one to ship.

Taste is a strange thing to have as a primary work product. Most of us were trained to think of it as a secondary trait, something nice to have on top of the real skill. You were a good designer, and you had good taste. You were a good writer, and you had a good ear. The taste was the finish. The skill was the substance.

In a world of infinite first drafts, the relationship flips. The substance is commodity. The finish is the moat.

You can see this already in practice. The companies doing interesting work right now are not the ones with the biggest model budgets. They are the ones with a clear house style, a recognizable voice, a view on what good looks like. When any tool can produce any output, the company's edge is the editor who says: not this, this. That editor used to be a cost center. Now she is the only reason anyone can tell the brand apart.

Taste has three properties that make it genuinely hard to replace.

The first is that it is tacit. It lives in the body of the person who has it. You cannot write it down in a style guide, and if you try, the style guide will always miss the case you did not anticipate. Tacit knowledge survives automation because the input needed to train on it is the work itself - and the work itself, for a specific person's taste, is a tiny dataset.

The second is that it is cumulative. A taste-maker gets better with every decision she makes. The feedback loop runs through her career, not through a training run. You cannot short-circuit it. You cannot buy it. You cannot hire someone with five years of experience and expect them to have the taste of someone with twenty, even if they are smarter, because the asset being built is not raw skill, it is pattern density.

The third is that it is contextual. Good taste for a magazine is not good taste for a brand. Good taste for a brand is not good taste for a product. The taste-maker who can hold the full context - the audience, the history, the constraints, the goals - and make decisions inside that context is doing something that does not generalize. Which means it does not centralize. Which means no platform can fully absorb it.

If this is correct, the working life of a lot of knowledge workers is about to bifurcate. On one side, the jobs that were about producing acceptable drafts - copywriting, junior analysis, entry-level design, basic research, first-pass coordination. Those are under pressure and will keep being under pressure. On the other side, the jobs that are about choosing, editing, deciding, shaping. Those are about to be more valuable than at any point in living memory.

The uncomfortable implication is that the second category is a smaller category. You do not need five junior copywriters if one taste-maker can review and approve a hundred drafts a week. The pyramid flattens. The top gets more valuable. The middle disappears.

A lot of the advice being given right now misses this. The advice is usually some version of "learn to use AI tools." That advice is fine as a floor, but it does not help you answer the real question, which is what you are going to bring to a market that has more production capacity than it knows what to do with. The answer is not production capacity. The answer is judgment.

Judgment is a word we have been careless with. For the last twenty years, "good judgment" meant something like "makes reasonable decisions in meetings." That is not what it means now. Now it means the capacity to look at a wall of outputs and say, quickly and correctly, which one should leave the building.

If you are in a role that used to reward production volume - more reports, more drafts, more decks - the move is to spend less time producing and more time selecting. The output that matters is not the thirty drafts you generated. It is the one you signed off on.

The moat is taste. It was always taste. For a while, production was hard enough to hide that fact. It is not hard anymore.