When AI Commoditizes Everything, What Happens to Taste?

A short essay posted to a personal blog this week triggered one of the more substantive Hacker News discussions of the past month: 222 upvotes and 184 comments arguing about whether AI is eroding, preserving, or irrelevant to human taste. The essay's central argument is worth taking seriously even if the headline oversimplifies it. When the marginal cost of producing a competent piece of writing, a functional codebase, or a passable design approaches zero, the valuable thing is no longer the production — it is the judgment that precedes it.

The Commoditization Argument

The essay frames AI-generated output as a new kind of abundance problem. Photography didn't destroy painting — it destroyed portrait photography as a living for craftspeople, but it also forced painting to evolve into forms that cameras could not replicate. The essay's author argues AI is doing the same to a much wider range of knowledge work: the functional version of almost any content artifact can now be produced by a sufficiently capable model with a sufficiently specific prompt. What remains as a genuine constraint is the person who can tell a good output from a merely acceptable one — and who cares enough to iterate until the distinction is visible.

Taste as Competitive Moat

The practical implication for practitioners is uncomfortable: the skills that were legible as competence — writing clearly, coding fluently, designing cleanly — are no longer sufficient differentiators. They are table stakes, producible on demand. The skills that become differentiating are harder to demonstrate and harder to hire for: the ability to recognize what is worth making, to have a coherent point of view about quality, and to maintain standards when the path of least resistance is to ship the first good-enough output the model returns. In a world of AI-generated abundance, taste becomes leverage. It is also something that cannot be prompted.