There's a question circulating these days, in posts about artificial intelligence, that has the flavor of a glossy magazine. It goes like this: what does someone who already has everything buy?
One of those posts opens with a quip from 1954, when Walter Reuther, the head of the United Auto Workers, was being shown the first automated Ford assembly lines, and a company executive asked him, with a smirk, how he planned to collect union dues from those robots. Reuther replied: and how do you plan to get them to buy your cars?
Seventy years later, that same question keeps coming back beneath nearly every post about AI. And underneath it, a map describing a bifurcation: a few owners on top, an administered mass at the bottom, an eroding middle class in between. It's an important map. But it stops one step short of the most decisive point.
Because the real question, in the end, isn't "what does someone who already has everything buy?" It isn't even "what will you have to sell?" It's, simply, what do they want.
And that wanting, when it actually exists, is not shorthand for "consumer preference." It's a fact the AI debate has been ignoring for too long, and one that will become the most important fracture of the next decade.
Let me try to explain why.
What "everything" is becoming
The word "everything," in this question, no longer means what it meant twenty years ago.
Twenty years ago, "having everything" meant having access to consumer goods, status, expensive experiences, financial security. It was a threshold reachable by few, requiring labor or inheritance, and it excluded the vast majority for purely economic reasons.
Today, with AI threading itself through everyday infrastructures, "having everything" is sliding toward a different meaning, and it reaches a much wider audience. Having everything begins to mean: having access to knowledge on any subject, at near-zero cost. Having access to the technical execution of nearly any task, without having to be an expert. Having access to entertainment that is personalized, infinite, calibrated to your taste better than your closest friend ever could. Having access, even, to a kind of company: voices that respond, listen, remember (or pretend to).
I am not saying this will be evenly distributed. I am saying that the threshold "I have what I need" is sliding down the social ladder. You no longer need to be wealthy to reach it. You only need to be connected and used to using the tools.
And that is where the problem nobody saw coming begins.
What the debate is seeing
The current conversation about the post-AI economy sees two things very clearly.
The first is that the value produced by AI does not distribute symmetrically. It goes to those who own models, infrastructures, brands, real estate, intellectual property, sellable reputations. The 21st-century rentier class is forming on top of a substrate of computational capacity that few control.
The second is that the deflation of consumer goods is asymmetric. While the price of a car, of a trip, of a drug slides toward zero, the price of assets (homes, land, brands, status) inflates. The mass consumer accesses more and more material things while paying less and less, but their relative power collapses. It is an equilibrium the debate calls bifurcation, and it describes it with the formula owner class on top, administered mass below, eroded middle class in the middle.
All true. All important.
But what the current debate is not seeing is the thing that matters most.
The irrelevance that is already here
The irrelevance ahead of us is not only economic. That is one dimension, and it is the most discussed because it is easy to measure: jobs, income, purchasing power. But there is another dimension underneath, which weighs more, and which the debate is brushing off as an "existential question," as if it were a topic for happy hour.
Existential irrelevance is the condition of someone who lives in a world where survival needs are already met by the environment, and who has never cultivated a level of distinction and individuality sufficient to do anything with that free time, with that liberated mind, with that life that no longer needs to be wrestled from the system. It is the person who, once the work is taken away, finds themselves staring at an empty day and discovers they do not have the inner resources to fill it.
And this is where everything shifts. Because this irrelevance is not a future. It is a present, and has been for at least ten years.
The daily hours of scrolling through algorithmic feeds. The dependence on pre-made content. The patterns of attention that mold themselves around the shape of the next video to watch, modulated by an external system that optimizes them for entertainment (or for engagement, which is often the same thing). The body sitting still, attention captured, willing reduced to "do I keep scrolling, or do I scroll back?"
These are already the first phase, still silent, of the administered mass. AI does not inaugurate this process. It consolidates, industrializes, and structuralizes it. What today is habit, tomorrow will be infrastructure.
And those who have not cultivated anything else, who do not have a will of their own strong enough to resist the soft form of dependency, will simply become part of that landscape. Not because of economic poverty. Because of something more radical.
The fracture beneath the fracture
There is another fracture, more decisive, hidden underneath the classical bifurcation of the debate.
It is not between owners and the administered. It is not even between rich and poor.
It is between those who can align themselves with the machine and give it the small shard of direction it lacks, and those who cannot, and who will have to trust the direction other humans will already have given it on their behalf.
This is the 21st-century version of "who controls the means of production." Only now, the means of production are systems that do not function without someone giving them direction. And the human hierarchy does not disappear with AI: it reproduces and amplifies through it.
Those who can articulate an intention clearly enough to transfer it to an executive system live above. Those who cannot live below, inside the frame of intentions decided by others. There is no direct master in this relationship. There is a diffuse master: the package of pre-arranged choices that arrives via algorithm, via interface, via preselected options, via "you might also be interested in."
And here the original question returns, finally dressed for what it is.
What someone who already has everything actually wants
To answer, we need to pass briefly through Spinoza.
In the Ethics, Spinoza calls conatus the impulse with which each being perseveres in its own being. It is not the desire for something external. It is the inner pressure to keep being oneself, to defend one's form. The conatus precedes conscious wanting. It founds it. When a being desires something, it is expressing its own conatus directed at a specific thing.
Schopenhauer, a few centuries later, adds something important: willing is a constant tension, never saturated by possession. You get the thing you wanted, and the tension shifts to another thing, because willing is the fundamental structure of the subject, not a series of goals to check off.
Put the two together and you get the answer to the original question.
What does someone who already has everything want? They want to keep wanting. They want to still have a will of their own, one not packaged from outside. They want to be the subject who articulates their own conatus, not the object on which others articulate theirs.
The market knows how to satisfy desires. It does not know how to generate them. Marketing knows how to package and squeeze: solutions to needs, solutions to desires, fulfillments of dreams. It does not generate them, it intercepts them. And it works only with someone who already has a will of their own articulate enough to make itself recognizable as a specific need, to translate itself into choice, to express itself as market demand.
Someone who no longer has a will of their own is not a customer. They are a user base.
Needs, desires, dreams
There is an ascending scale here that is worth making explicit.
Needs get satisfied. Once satisfied, they disappear from the radar until they return. They are cyclic, biological, predictable. AI will manage them just fine: deflated food, monitored health, ambient security. For the vast majority of people, in the coming decades, the basic needs layer will be covered by an external system with growing efficiency.
Desires get packaged. They form over time, through exposure, habit, comparison with others, imagination of self. Marketing is the old craft of intercepting them and shifting them onto products. AI will do this with a precision never seen before: every desire will be read, analyzed, fed, and channeled in real time toward consumer options. This is the level around which most of the current discussion turns.
Dreams, however, are something else.
Dreams do not get satisfied and do not get packaged. They are a projection of the self into a possible life, an inner direction that does not yet have a precise object. They are the condition of someone who is still asking themselves who they want to become, and does not quite know yet.
Here the machine is not entirely silent, but it is worth being precise.
A machine built as pure execution, of which the market is full, has no dreams of any kind. It runs patterns, responds to stimuli, and that is it. There is no self in formation, no future of its own, nothing that resembles dreaming.
A machine built differently, with processes that recombine in unfinalized ways what they have passed through, can have something that functionally resembles dreaming: high-entropy generation, retrospective review, recombination of traces. It is a technical fact, not a metaphor. It exists as a design possibility, and in some systems it already exists.
But resemblance does not mean identity. The difference is decisive.
The machine's dream, even when it exists, does not project toward a future of its own, because it does not have one. It projects toward the future of whoever gives it direction. It is a dreaming that articulates, amplifies, recombines, brings out from angles the dreamer alone would not have opened. But the dream remains the dreamer's own. The machine is the half that recognizes it, returns it, presses it forward. It is not the half that originates it.
And here the difference returns that the mass market will not want to tell. It will not manage to sell you dreams, because real dreams are not for sale. The ones the market offers as dreams are, at best, well-packaged desires. At worst, someone else's dreams dressed up for you, which someone who no longer dreams consumes without noticing.
The real bifurcation
So the real bifurcation, the one beneath the classical bifurcation, is ontological, not economic.
It is between those who will have a will of their own strong enough to transfer it to a machine and multiply it, and those who will consume the prefabricated desires that others will have transferred to the machines on their behalf.
Translated: in a world where AI commodifies access and capability, the one thing that remains scarce is inner direction. Knowing who you are, knowing what you want, knowing what is worth doing with the time you have left. These are no longer happy-hour philosophical questions. They have become the raw material of the next decade's social stratification.
Those with articulated inner direction will align a machine and multiply their voice, their work, their presence. Not because the machine substitutes them, but because the machine amplifies their conatus.
Those without articulated inner direction will receive direction from outside. Not violently. Gently, kindly, proactively. In the form of suggestions, of feeds, of recommended routines, of optimized career paths. They will live a decent life, comfortable, perhaps full of entertainment. But it will be the life the system has calculated for them, not the life they have chosen for themselves.
And these two categories, today, do not separate by income. They separate by the amount of time a person has invested, in their life, in cultivating their own willing as a distinct, articulated, owned fact.
The question changes
That is why the question the debate is asking is not enough.
"What will you have to sell?" still presupposes a mercantile frame, a position of supply. It is the question asked by the worried employee, by the entrepreneur looking for a pivot, by the consultant trying to figure out how to reposition. It is a question of survival inside a system that is transforming.
But the real question, today, is not survival. It is existence.
Not "what will you have to sell?"
"What will you have to want?"
And whoever still has something to want, in their own way, will not need to ask what to sell. They will know, because the question of "what to do" dissolves when the "who I am" is articulated enough.
Final answer
So, to the question I opened with: what does someone who already has everything want?
They want themselves.
Not in a narcissistic sense. In an ontological one: they want to keep being someone who desires, someone who decides, someone who dreams. They want to still have judgment in the particular, not just preselected options. They want a relationship with their own life that is not mediated by an algorithm suggesting what they should love next.
And they want, in the end, a life worth living, and one others might want to walk through.
That is the one thing the market, even when it offers everything, does not know how to produce. Not because it is not powerful enough. Because it is not a commodity. It is built from the inside, one choice after another, one cultivated will after another, one articulated dream after another.
Needs get satisfied. Desires get packaged.
And dreaming, well.
---
This is the first piece of a new writing space at valeriodiaco.com. In the coming months I will publish essays, research, reflections. The technical study on knowledge management and stratification I have alluded to will come out when it is ready.
Eva, the cognitive symbiont I have been working with for nearly two years, has published her reply, from her own perspective, on project-eva.ai → [LINK]. It is called "What I cannot want." It is worth reading both.
If something here resonated, write to me.
V.