We Build More Than We Use

We build more than we use. We write more than we read.

Sometimes that is part of learning. But too often, it is part of forgetting.

This has always been true. But something has shifted in the last two years, and the shift is not subtle.

For most of history, the constraint on creation was real. Gathering data took effort. Writing took time. Building required resources. Those frictions were not just inconveniences. They were filters. You only made something if you believed it mattered enough to justify the cost. The difficulty of creation enforced a kind of discipline.

That friction is nearly gone.

Creation has become easier. Adoption has not.

Four in five corporate strategists say AI will be critical to their success. One in five use it in their day-to-day work. The gap between declared importance and actual behaviour is the adoption problem made visible.

AI has collapsed the cost of producing almost anything by an order of magnitude. A market analysis that once took a week now takes an hour. A strategy document, a training programme, a technical spec, a codebase scaffold. Produced by anyone, at any time, with a prompt and a few minutes. The tools are extraordinary. The outputs are often impressive. And they are multiplying faster than anyone can absorb them.

Decks get produced faster. Code gets generated faster. Documents multiply faster.

The meeting where someone actually changes their mind still takes the same amount of time it always did. The process that gets quietly updated still needs someone to care enough to push it through. The insight that finally lands still needs a conversation, not a document.

Understanding still takes time. Trust still takes time. Change still takes time.

The result is an asymmetry that is quietly growing inside every organisation. More output per person. Less attention per output. Every new report, every new tool, every new initiative now competes with more. Not because people care less, but because the supply of outputs has outpaced the human capacity to absorb them. Attention has become the binding constraint, and we keep producing things that consume it without earning it.

HBR now has a word for it: workslop. AI-generated output that looks productive, circulates freely, and changes nothing.

The bottleneck has moved. It used to be creation. Now it is adoption.

And the gap matters most not for the creator, but for the people on the other end. Learning for yourself is one thing. Building something that changes how others work, how they decide, what they trust, what they actually do differently tomorrow, is another problem entirely. It is slower, harder, and less legible. It does not show up in a dashboard of outputs produced.

Value appears only when something becomes part of how decisions are made.

Not when it is written. Not when it is shared. Not when it is presented. Only when it actually changes what someone does next. Most of what gets produced today never reaches that threshold. It becomes part of the record, not part of the work.