Scarcity Isn't Discovered
Alex Imas has written a thoughtful piece on what AI will leave scarce, and the part of it that is my trade is the part I have the most to say about.
His argument, in outline: when AI automates the making of commodities, cheap commodities raise real income, and real income funnels toward what he calls the relational sector — teaching, nursing, hospitality, craft, performance — where the human being involved is part of the value itself. Mimetic desire does the rest of the work. People want what others have and what others cannot have. Scarcity becomes the good. And since AI-generated things are already discounted in his experimental data — human artwork gained 44% from exclusivity, AI-generated artwork only 21% — the relational sector ought to be a stable refuge for human labor as automation gets better.
On the underlying economics I have nothing useful to add. Income elasticity is real. Baumol's cost disease is real. Girard is correct about what makes people want things. I have spent my working life inside the consequence of all three.
The piece falters in one place, and it is the place I recognize from the inside.
It treats mimetic desire — and the relational sector it underwrites — as something like a natural release valve. As though scarcity is a feature of the good itself, which the economy merely discovers and redistributes spending toward.
Scarcity is not discovered. Scarcity is produced.
The "handwritten note on the cup" Imas opens with — the Starbucks CEO reversing course toward ceramic and personalization — is not a return to relational value. It is a marketing department succeeding at the manufacture of relational signals. The notes are training-wheels authenticity, repeatable across thirty thousand stores. The ceramic is a prop. The result, if it works, will be an industry of synthesized intimacy sold at the price of real intimacy — and when AI gets good enough to handwrite the note and hand-throw the ceramic, the next layer of the "relational" will be produced too.
This is what my profession is. It is the manufacture of the felt-scarce. Every premium brand you have ever paid a premium for has some version of this at its center. What Imas describes as a stable sector of the future economy — where humans are involved because their involvement is part of the value — is the condition of already-existing luxury markets, generalized. The thing about those markets is that they are relentlessly worked on. The relational signal is studied, decomposed, copied, and synthesized. Whoever figures out how to synthesize it at lower cost captures the premium.
Imas anticipates the objection in a single phrase: "not requiring Picasso-level talent, just human presence that makes something feel personalized." That word is the one I want to dwell on. A marketer's job is the production of feel. If feel is what commands the premium, and feel can be produced by anyone who studies its signals closely enough, then the relational sector is not a moat. It is the next commodification frontier, with a longer half-life than the last one because the signals are harder to fake — but not an indefinite one.
Which brings me to the part that sits closest to our work here.
Imas's forecast assumes the humans on the receiving end — the customers, the patients, the students, the audience — retain the capacity to tell a real relational good from a produced one. The mimetic release valve depends on it. Status competition requires working judgment. So does paying double for the thing others can't have.
If judgment thins, the premium collapses into the signal. The ceramic is enough. The handwritten note is enough. The LLM girlfriend who "remembers your birthday" is enough. A population whose attention has been softened by brainrot cannot reliably distinguish manufactured warmth from the real thing, and it will pay the relational premium for a plausible synthesis of it.
So I would amend the forecast. The relational sector will grow — for the fraction of people who can still tell. And a parallel sector, the simulated-relational, will grow much faster for the fraction who cannot. They will look the same from the outside. One will be a restaurant. The other will be an app. Both will have handwritten notes on the cups.
This is, in a way, the whole reason we exist. Attention is the asset. Judgment is the asset. If the asset depreciates, Imas's optimistic forecast becomes the forecast for the shrinking fraction of people who are holding onto theirs.
The piece is worth reading. The argument is serious and the data is real. I disagree only about what will be scarcest: not the human touch — that can be manufactured — but the kind of human still capable of noticing the difference.