If AI Removes One Kind of Scarcity, What Becomes Scarce Instead?

If AI Removes One Kind of Scarcity, What Becomes Scarce Instead?

For most of the history of making things, the ability to make them has been constrained by some form of scarcity: skill, labour, time, specialized knowledge, access to expensive tools, or people capable of translating an idea into something that actually works.

Technology has always rearranged these constraints. But generative AI is doing something particularly interesting: it is making certain capabilities abundant at remarkable speed.

A person who could not write software can now prototype an application. A small company can produce research, analysis, presentations and communications at a volume that once required a much larger team. A founder can move from idea to something demonstrable astonishingly quickly.

We tend to describe this primarily in terms of what AI eliminates: time, labour, cost, friction.

But after listening recently to a room of people building and funding companies, I found myself wondering about the other side of the equation: If AI removes one kind of scarcity, what becomes scarce instead?

When building becomes easier, building stops being the proof

For an early-stage company, building a working product once demonstrated something significant. It showed technical capability. It required time, expertise and often capital.

Now, the product may already be expected before the first serious investor conversation. The difficult question becomes less Can you build this? and more:

• What do you know, possess or understand that makes what you’ve built difficult to replace?

The scarcity has moved.

When information becomes abundant, direction becomes scarce

Generative AI can produce enormous amounts of competent information very cheaply: research summaries, strategies, recommendations, reports, plans, options and ideas. More is available to us than we can reasonably evaluate, let alone act upon.

So the scarce resource may not merely be attention. It may also be direction. Critical thinking.

Which information matters? Which interpretation deserves confidence? Which of several plausible paths should we pursue? What are we trying to accomplish in the first place?

Producing more possibilities does not remove the need to establish what those possibilities are for. The ability to establish direction among them may become considerably more important.

When answers become easier, questions become more consequential

There is a strange inversion happening in knowledge work. Answers that once required expertise, research, access or time can increasingly be generated almost instantly.

But an answer is useful only in relation to the question being asked. And the quality of that question depends partly upon what someone has noticed.

What problem are we actually solving? What assumption are we making? What evidence would change our minds? Are we solving the symptom because the underlying problem is harder to name?

• A beautifully generated answer to the wrong question is still the wrong answer.

This makes me wonder whether problem-framing becomes more valuable as problem-solving becomes easier.

When production accelerates, judgment becomes the bottleneck

AI can produce, compare, synthesize, simulate and recommend. Eventually, somebody has to choose. Who decides whether the evidence is sufficient? Which trade-off is acceptable? Whether an exception matters? Whether a technically workable answer is actually the right thing to do?

And when the decision has consequences, who is accountable for it?

• As AI takes on more production, human work may migrate toward places where ambiguity, trade-offs and responsibility concentrate.

The quality of judgment becomes more—not less—important.

When competence becomes abundant, trust becomes harder to infer

We have historically used visible artifacts as proxies for capability. A polished presentation suggests effort. A sophisticated website suggests resources. A working prototype suggests technical competence. A thoughtful report suggests someone spent time thinking.

Those signals were never perfect. Now they are becoming noisier. AI can help almost anyone produce work that looks polished, articulate and competent.

• What does polish prove now?

Perhaps we begin looking elsewhere: track record, specificity, evidence, reputation, relationships, the quality of someone’s questions, the accumulated experience behind a recommendation.

Trust may become more valuable precisely because some of the signals we once used to infer it become easier to manufacture.

When everyone can access intelligence, context becomes an advantage

The same AI model can be available to millions of people. That doesn’t mean millions of people get the same value from it.

One person gives it a generic prompt. Another gives it years of structured knowledge: decisions, customer conversations, experiments, failed attempts, cases, research and observations.

The underlying intelligence may be similar. The context is not.

So the question becomes not merely Which model are you using? It becomes:

• What do you know that the model doesn’t?

When the average becomes excellent, difference becomes expensive

Generative AI is exceptionally good at producing plausible work from patterns that already exist. And the quality of that work continues to improve.

That means the average may become remarkably good.

• If competent execution becomes inexpensive and widely available, competence itself may stop being sufficient differentiation. What becomes valuable then?

Taste, originality or unusually deep expertise may matter more. So might strange combinations of experience, or the ability to recognize when a conventional answer no longer explains enough.

The point isn't that any of these things are magically immune to AI. It's that abundance changes their relative value.

Scarcity doesn’t disappear. It moves.

This is why I’ve become less interested in asking only: What will AI replace?

Replacement tells us what may be disappearing. It doesn’t necessarily tell us where value is moving.

If software becomes easier to build, building stops being sufficient proof of advantage. If information becomes easier to produce, direction becomes more important. If answers become easier to generate, problem-framing becomes more consequential. If competent outputs become abundant, trust becomes harder to infer.

• If everyone can access powerful intelligence, distinctive context may become more valuable. If execution accelerates, judgment carries more weight. If established patterns become easier to reproduce, difference becomes harder to sustain.

AI isn’t simply eliminating scarcity. It is relocating it. And that makes me wonder whether the more useful question for organizations isn’t only what AI can now do for us. It may be: What becomes more valuable because it can?

::::::::::::::::::::::::::::::::::::::::

Contemplations & Provocations | A field essay by
Marie Fe Isla Rae

© 2026 Marie Fe del Rosario. All rights reserved.
Terms of Use · Privacy Policy · Permissions & Collaborations

All content on this site—including text, frameworks, methodologies, visuals, and strategic concepts—is protected by copyright and may not be reproduced, adapted, or used without prior written permission.

Authorship and web design by Marie Fe.

Field notes on consequential moments, creative strategy, experience, narrative, design—and the work of making possible futures tangible.

For people interested in what happens between imagining and making.