The quaternary economy: what comes after work

Eight percent. That is the share of French employees Gallup counts as engaged at work, against 12 % across Europe and 20 % worldwide¹. Nine people out of ten walk through their working days without investing much of themselves in them. And we are about to spend the coming decade looking for ways to save something almost nobody is attached to.

The public debate on AI and employment has been going in circles for three years. On one side, those who announce the collapse. On the other, those who point out that every technical revolution ended up creating more trades than it destroyed. Both camps share the same blind spot. They count the jobs. What will still give a human being value, once their productive contribution has become marginal, stays outside the frame. Another economy is opening up behind that question, and we can already see it surfacing. I call it the "quaternary economy".

The five percent threshold

Let us start with the reasonable camp. Daron Acemoglu, Nobel laureate in economics, estimates that only around 5 % of tasks will be profitably performed by AI within the decade². The figure is useful because it recalls something enthusiasm keeps trampling: between what a machine can do in a demo and what a company has any interest in handing over, there is cost, integration, legal liability, and the time an organisation takes to change its habits.

The people working with these systems daily see an entirely different slope. In the survey Anthropic published in June 2026, more than a third of respondents expect AI to handle most or nearly all of their work tasks within twelve months. 57 % believe it has made their skills more valuable, and 10 % consider it likely they will lose their job within the year³. The macroeconomic projection and the experience of those with their hands in it simply do not describe the same world.

Acemoglu’s arithmetic holds. The field, though, has a strange property that this kind of model does not capture. The moment you precisely define a task a human would perform better than a machine, you have just supplied the specification, the evaluation criterion and often the dataset needed to train for it. To define X is to arm the research that produces X. That loop existed in no earlier wave of automation: the power loom did not reprogram itself by reading a description of what it could not do.

And while we argue about the cognitive, Physical Intelligence is breaking the fourth wall and catching up faster than expected. Unitree shipped roughly 5,500 robots in 2025 and is targeting ten to twenty thousand units in 2026; Figure deployed its first Figure 03 units at an industrial customer’s plant in January 2026⁴. Chinese humanoid robotics has joined the dance and the volumes remain tiny, nobody claims otherwise. Look at the slope.

So let us grant the hypothesis, without taking it for prophecy: within a decade, the share of tasks where humans remain the best becomes residual. What follows? Human productive value turns marginal, so there is less reason to pay people, so less consumption, so less reason to produce. The engine chokes on its own. The problem reaches far beyond employment: the fuel running the whole machine is what starts to fail.

What the machine reveals nothing about

One question economists rarely handle remains, because it belongs to anthropology: where does human value retreat to once it leaves production?

Take sport. A robot runs the hundred metres in four seconds. Nobody travels to watch it twice. Watching Usain Bolt, what lifts us reaches far past the stopwatch: a body made of the same flesh as ours, bound by the same limits, has just pushed them back. The basketball player sinking a hundred shots in a row reveals invested lifetime, discipline, consented pain. The robot programmed never to miss reveals nothing at all. It is a technical performance, and it leaves us cold.

The same mechanism operates wherever the result only matters through what it reveals of the process. In go, which I play, the whole interest lies in the fact that your opponent shares your being-in-the-world. Against an AI that outclasses you, there is nothing left to play for. In art, the rise of live performance and self-staging is no accident. In part of care, too: under anaesthetic, we want the machine if it operates better, but in the path leading up to it and away from it, we want a human nurse, because a machine does not feel what we feel and cannot represent it to itself. To optimise a farming plot, on the other hand, we bring in the machine without the slightest qualm. Only the result counts.

There lies the criterion. What makes humans specific is a missing capability. Their fatigue, their fallibility, the finitude they share with whoever is watching. We pay more for the potter’s vase with its small flaw than for the perfectly calibrated industrial one, which has better use value. That small flaw works as a signature.

This is what the quaternary economy means: an economy where we are valued for who we are, and where the worth of a result lies in the authenticity of the process it reveals. We climb one rung up Maslow’s pyramid, towards self-realisation. We stop being interchangeable links, since singularity is precisely what creates the value.

The stage and its bill

The other side is harsh. To be valued for your singularity, you have to display it. The quaternary economy is therefore an economy of spectacle.

Content creators are its weak signal, and the numbers already describe the structure of the world to come. The global creator market is worth more than 250 billion dollars in 2026, across an estimated 200 to 300 million creators. Around 4 % of them earn more than 100,000 dollars a year, and more than half take home less than 15,000⁵. A very sharp pyramid, where attention concentrates at the top with the brutality of a power law.

We call these trades "bullshit jobs" today, after the anthropologist David Graeber. Three hundred years ago, a farmer would have said exactly the same about a marketing manager. These people are paid for who they are, and what they produce is spiritual in nature. Tomorrow it will probably be the core of the human economy. I may as well say it plainly: I am one of them. Content, stages, an income that rests on my signature. And I love it, including on the days when that dependency on other people’s attention makes me deeply uncomfortable.

The phenomenon has two sides. An attention overload of little interest, clips cut to capture three seconds of brain. And people with magnificent things to tell who would never have broken through without that channel.

Then comes the ritual objection: without work, humanity goes soft. WALL-E gets invoked, the plump bodies in their floating armchairs. Except the film tells the opposite of what it is made to say. The moment the small plant appears, the whole ship wakes up, fights to go home, and the humans take their first steps smiling on a planet of waste. Because there is novelty. Curiosity and desire are the most deeply human things there are, and Lacan gave us the formula: our first desire is the desire of the other’s desire. We want to be seen and wanted. An idle society looks for an arena. Better that this arena be a stadium, a stage or a go board, because humanity has historically built far bloodier ones.

Who holds the valve

The question of sharing remains, and this is where things get stuck.

Universal basic income has established itself as the default answer to automation. The data, incidentally, is not bad: the study run by OpenResearch paid 1,000 dollars a month for three years to a thousand people in Illinois and Texas, against a control group receiving 50 dollars. Recipients did not leave the labour market, they slightly reduced their hours, increased their savings and their spending on others, and gained in their capacity to decide about their own lives⁶. The experiment is rather reassuring.

It measures what money does to people. The position of whoever holds the valve falls outside the protocol. Paying an income to everyone flattens the base of a society of deeply social, hierarchical beings, while installing above it an absolute vertical inequality: that of the actor deciding when the tap opens and when it closes. If that actor is a government, we have just invented a lever of docility of unprecedented power, and history says clearly enough what gets done with those. If it is private companies redistributing a share of their profits, the dependency changes hands without changing nature. Like every system promising equality from the top down, this one carries an authoritarian seed. Just ask yourself who, in your country, would be holding that tap.

Another path exists, and it comes down to one simple shift: distribute ownership rather than income. If the AI systems and robot fleets doing the producing largely belong to those who live off what they produce, then everyone holds their own productive capacity. Material scarcity collapses, value leaves goods and services of its own accord, and lodges itself where it cannot be copied: in singularity, presence, the gesture that reveals a person. The quaternary economy then becomes the mechanical consequence of a choice about ownership.

This road has its chasms, obviously. The main one carries a name science fiction has explored to the bone: neofeudalism. A handful of owners holding the economy the whole of humanity lives inside, and the rest of the world renting. Nothing forces us to treat that as fate: our societies have shown real antibodies, and negativity bias always makes us overestimate the downward slope. But the gap between the two trajectories will be decided by choices in law, taxation and governance, made over the next five to ten years by people who, for the most part, have never yet opened a report on the question.

I am not an economist, and several economists I respect do not share this reading at all. But when 92 % of French employees walk through their days without engagement, it becomes hard to defend the current economy on the grounds of what makes it desirable. What comes after it is enough to make your head spin, and frankly, enough to make you want it.

We have spent two centuries asking machines to produce in our place. The quaternary economy will begin the day we decide who owns them.

Footnotes

1. Gallup, “State of the Global Workplace 2026”, France data, collected January to December 2025. https://www.gallup.com/workplace/705338/state-global-workplace-france-country-level-data.aspx ↩

2. Daron Acemoglu, “The Simple Macroeconomics of AI”, NBER Working Paper 32487. https://www.nber.org/papers/w32487 ↩

3. Anthropic, “Anthropic Economic Index report: Cadences”, June 2026. https://www.anthropic.com/research/economic-index-june-2026-report ↩

4. Production volumes and industrial deployments reported in 2026 for Unitree (around 5,500 units shipped in 2025, targeting 10,000 to 20,000 in 2026) and for the first Figure 03 units installed at an industrial customer in January 2026. ↩

5. Grand View Research, “Creator Economy Market Report”, 2026, together with concurring sector surveys on the distribution of creator income. https://www.grandviewresearch.com/industry-analysis/creator-economy-market-report ↩

6. OpenResearch, “Unconditional Cash Study”, results published from July 2024. https://www.openresearchlab.org/unconditional-cash-study ↩