Main idea

Shipping more features, more screens, more variants and more code costs a software vendor less every year, and the number of players able to attack a market rises along with that fall. A capacity everyone has distinguishes nobody: building fast and well becomes table stakes, not the advantage.

What stays scarce is what has not fallen in price: understanding the customer's business, their real users, their technical constraints, their standards, their field culture and their competition, before deciding what to build. Scarcity does not disappear, it changes place — it leaves "doing" for "understanding".

The question that separates two vendors then stops being "can we develop this feature?" and becomes "do we understand the situation well enough to build the right thing?". The first asks about a capacity the market distributes; the second asks about work nobody can buy ready-made.

Layer added by "PM, Developers, and AI: Roles Are Blurring, Responsibilities Remain" (2026-08-01). The list of what stays scarce gets sharper when you take it from the side of a product team rather than a market: the quality of the customer signal, proprietary data, business knowledge, distribution, hard-won trust, operational constraints, accumulated learning, and the ability to correctly interpret what customers are really asking for. These are exactly the things a competitor cannot generate. A team that produces a lot without reading its market better gains nothing: it merely produces mediocre things faster.

Why it matters

This changes a product team's investment criterion. As long as producing is the bottleneck, accelerating production is rational; when it stops being the bottleneck, the same effort funds an abundance that no longer differentiates, while the real bottleneck — understanding — stays underfunded.

It also gives a reading of markets that fill up fast: the arrival of credible newcomers there signals no technical superiority, only the collapse of the price of entry. Defending yourself by producing more is answering on the one ground where the gap has closed.

Nuances and limits

Abundance of production is not yet a reality everywhere: in fields with heavy reliability, certification or real-time constraints, producing stays expensive and slow, and the execution advantage still holds.

And a superior understanding does not convert into an advantage on its own: as long as it stays in a few heads, it is an individual talent rather than a company asset.

Open questions

• How does a team recognize that the bottleneck has actually left production, rather than merely assuming it?