Choosing what to build

A few years ago, building a working product took weeks. Often months.

That constraint shaped the work. You had to choose carefully because execution was expensive. The question of what to build was tied to the question of what was possible to build at all.

That is no longer true.

A landing page, a tool, a workflow, a storefront, a small app — a working version can now exist in an afternoon. The gap between idea and product has collapsed.

But that did not remove the hard part.

It moved it.

The question is no longer only: can this be built?

The question is: should this exist at all?

Not every idea earns a product

There is a difference between an interesting idea and a product worth owning.

An interesting idea is easy to talk about. It has a hook. It sounds useful. It makes sense in a sentence.

A product worth owning does something specific for someone specific. It solves a real enough problem that a person would choose it over starting from zero, doing without it, or building their own version.

That gap is not technical.

It is judgment.

Most weak products do not fail because the build was impossible. They fail because the idea was too broad, too vague, or too in love with itself. They describe a category of possibility instead of a particular thing a particular person would actually use.

The more precise the user, the sharper the product becomes.

The more general the idea, the easier it is to build something that feels finished but belongs to no one.

The finishing test

One honest test for whether an idea deserves to become a product:

Are you willing to finish it?

Not prototype it. Not generate the first version. Not see how it looks when the model scaffolds the page.

Finish it.

Write the empty states. Handle the awkward edge cases. Make the copy clear. Decide what happens when something breaks. Remove the parts that make the product heavier than it needs to be.

Finishing is where most ideas lose momentum.

Because finishing is no longer about the excitement of the concept. It is about the person on the other end. The person who opens it, tries to understand it, and decides whether it is worth keeping.

If an idea cannot survive that test, it may still be a good demo.

But a demo is not a product.

Curation is now a building skill

When building gets faster, curation becomes more valuable.

The most important skill is not producing more. It is knowing what not to produce. Which idea to leave alone. Which feature to cut. Which product is worth finishing, and which one is only interesting because it was easy to make.

This is not hesitation.

It is taste under pressure.

AI makes addition cheap. It can generate more screens, more flows, more options, more versions. It can make almost any idea look real before anyone has decided whether it is good.

That makes choosing harder, not easier.

What survives

The products that last are the ones someone took responsibility for.

You can feel it in the decisions. What is included. What is missing. Who it is clearly for. What it refuses to become.

A model can help you build the thing.

It cannot decide that the thing is worth building. That commitment is still human.

And in a world where almost anything can be made to look finished, the real question is not how fast you can build.

It is whether the product deserves to be carried forward.

Not every idea needs to become a product.

But the ones that do should be finished well enough for someone else to run with.