When building became cheap, judgment became expensive

For most of my career, building something new was an exercise in scarcity. There were never enough engineers, enough time, enough budget, or enough attention. Every idea had to survive roadmaps, trade-offs, technical constraints, and organizational politics before it could become real.

That friction was frustrating, but it served a purpose. We could not build everything, so we had to decide what mattered.

It forced us to choose.

AI is removing much of that friction. A person can now write software, design a campaign, produce a video, or turn a vague idea into something tangible in a fraction of the time it once required. It is an extraordinary expansion of human capability. It is also creating a different kind of problem: when almost anything can be made, making something is no longer enough.

The internet is already filling with competent work produced at remarkable speed: polished images, correct copy, functional interfaces, videos that arrive edited, narrated, and scored.

Most of it is fine.

That may be the danger.

We imagined that better tools would produce better work. Mostly they produce more of it: more options, more variations, more content competing for the same finite attention. Abundance does not remove the need to choose. It makes choosing the whole game. The scarce skill is becoming judgment.

Judgment is knowing which idea deserves to exist before producing it. It is the difference between technically impressive and genuinely useful. It is knowing when an image feels right, when a sentence feels true, and when an idea is simply adding to the noise. It also means knowing when to stop. AI will happily generate an eighth variation, add another feature, rewrite another headline, and keep optimizing long after the original idea has lost its shape.

The machine is designed to continue.

Human judgment provides restraint.

For a long time, expertise meant execution: the best people could make things others could not. That still matters. But execution is becoming easy to rent, and the important questions are moving upstream. What should we make? Who is it for? What deserves someone’s attention? What should remain slow, imperfect, or human? Those questions are not answered by more processing power. They require context, responsibility, and a point of view.

Taste is hard to describe because it looks subjective, yet we recognize its absence immediately. It is the difference between a product that merely works and one that feels considered; between an advertisement that fills a placement and one that makes someone feel something; between an idea that follows a pattern and one that could only have come from a particular person.

Taste is built through attention. It comes from noticing why certain things stay with us, why some choices feel inevitable and others slightly wrong. It develops through curiosity, exposure, practice, and the willingness to reject work that is acceptable but empty. AI can imitate its surface: a visual style, a tone of voice, a familiar structure. But imitation is not the same as caring about the result.

The machine can offer possibilities. Someone still has to decide which one is worth pursuing.

I see this every day in the company I am building. At Tactimetry we make advertising, and producing an ad gets faster every month. The hard part has stopped being production. It is deciding what a brand should say, what will feel distinctive, and which idea deserves to be placed in front of people. The tool can generate. The person must direct.

As machines become more capable, our standards matter more, not less. The future will hold more than any of us can meaningfully consume: more products, more content, more claims on our attention. In that world, judgment becomes a form of care. It means making fewer things, more deliberately; protecting people’s attention instead of merely competing for it; using powerful tools without surrendering the responsibility of choosing what they are for.

Building is becoming cheap.

Judgment is what’s left that’s expensive.