If Intelligence Gets Cheap, What Becomes Scarce?

When an input to a process becomes abundant, value moves to whatever is still constrained. Asking what becomes scarce is more useful than asking what becomes cheap, because the scarce things are where advantage and attention should go.

What stops being scarce

Reasoning applied to well-posed problems. Generating options. Producing first drafts. Translating between representations. Breadth of coverage. Explaining things.

These were expensive because they required a person's trained attention. That constraint is loosening.

What becomes scarce

Knowing what's worth doing. If everyone can build anything cheaply, the differentiator is choosing correctly. Wrong choices now happen faster and at greater volume, which makes the choosing more consequential rather than less.

Verified truth. Generating a plausible claim is now nearly free, so the marginal value of plausible collapses. What's scarce is the claim that's been checked. This is why quoted-span verification, reconciliation, and oracles matter more than they did — they're the mechanism that separates verified from plausible in a world flooded with plausible.

Attention. More generated output competing for the same human hours. Whether it's review capacity in a team or reader attention in a market, the constraint is unchanged and the supply pressing against it has risen.

Trust. In systems, in information sources, in people. When anything can be produced convincingly, the value of a source that's been reliable over time rises. Reputation becomes more valuable precisely because the surface signals of quality became cheap to fake.

Context that isn't written down. The unwritten organizational knowledge — why the system is like this, what was tried, what the customer actually needs — becomes the scarce input, because it's the thing free capability doesn't come with.

Accountability. Someone who can be asked to explain and who bears the consequence. Structurally scarce, not just currently scarce.

Coordination. Getting people aligned and moving. Possibly harder as more work happens in parallel.

💡 The pattern

Everything on the scarce list has one of two properties: it requires being embedded in a specific situation over time (context, trust, reputation, knowing what's worth doing), or it requires bearing consequences (accountability, verified truth in a sense — someone stands behind it).

Neither can be produced on demand. Both accumulate slowly, in a particular place, through a history that can't be shortcut. That's a reasonable working definition of what stays scarce when generation becomes free.

What follows practically

For an individual: invest in the accumulating things. Depth in a domain, a track record people can point at, context nobody else holds, judgment about what matters. These compound and can't be bought.

For a team: the harness, the verification, the captured context. ✅ Free capability with a bad harness produces bad results — the differentiator moves entirely to what surrounds the model.

For a product: if the value proposition is generation, it's exposed. If it's verified generation, trusted distribution, or proprietary context, it's better positioned.

⚠️ The failure mode to avoid: competing on the abundant thing. Producing more, faster, is the losing move when everyone can produce more, faster. The winning position is somewhere on the scarce list.

🔍 The check to run on your own position

Ask what you're currently valued for, and which list it's on.

If it's on the abundant list — producing artifacts, applying known patterns, generating options — the position is eroding regardless of how good you are at it, and possibly because of how good you are at it.

If it's on the scarce list — judgment, context, accountability, trust — it's appreciating.

Most people are on both lists, in different proportions. Knowing the proportion is the actionable part, and it's a question you can answer honestly in five minutes.

The takeaway

Cheap capability makes verified truth, unwritten context, attention, trust, accountability, and knowing what's worth doing the scarce goods. All of them either accumulate through being somewhere over time or require bearing consequences — which is why they can't be generated on demand. Don't compete on the abundant thing. Check which list your own value sits on, and move deliberately toward the one that's appreciating.