The world that's coming

Tuesday, five years from now

It's eight in the morning, a Tuesday in 2031, and the night has been at work. The file the board will open at nine has been investigated while everyone slept: every document read — all of them —, every figure cross-checked, every precedent called up, every contradiction sought and found. No one wrote this brief; everyone has read it. The evening before, a question posed in a single sentence mobilised, for a few minutes, more analytical power than an entire team could muster in a month, ten years earlier. No one found that remarkable. No one is surprised by the light when they flip the switch.

And at nine o'clock, this room will still hold exactly what it held a century ago: a few people around a table, a decision that commits more than themselves, and the question of who signs it.

Only one thing has truly changed, and it doesn't show: ten years earlier, no one would have investigated that file at all. Not out of negligence — out of arithmetic. Intelligence was rationed.

This text describes that world. Not a prophecy — a trajectory already under way, already written into the prices. Five years remain to learn to live in it.

I · Ubiquity

In 1900, electricity was a product: people went to see it at exhibitions, sold it to the daring, debated it in drawing rooms. By 1930, it was a condition: it had vanished into the walls. No one, since, "uses" electricity. We inhabit it. It is visible nowhere and necessary everywhere — to the point that its only perceptible event is the outage.

Artificial intelligence is following the same curve, faster. A general revolution is under way — closer in nature to electricity than to a tool. Frontier models already produce, in minutes and at near-zero marginal cost, cognitive work of senior-analyst calibre; each generation widens the scope — the systems become agents, chain tasks together, check their own work. You have the signs of it on your desk already: the memo that used to take weeks, the document review that used to mobilise a team, the watch that used to require a subscription and two interns — the order of magnitude of time and cost has changed, and it keeps changing.

Extend the curve five years, asking it for no miracle: intelligence ceases to be a tool you pick up and becomes a medium you inhabit. Present in every document, every exchange, every decision as it is prepared — and as unremarkable as the current in the walls.

Does believing this require believing in the singularity — the point where the machine improves the machine? That horizon remains disputed, and honesty demands saying so. But ubiquity needs no marvel: it needs only a cost curve — and the curve is already there. Whether or not the marvel arrives, the tide, for its part, is already rising.

II · What scarcity forbade

This world is readily told in terms of replacement: what the machine will take, what will be left to us. That's the wrong question. Replacement is the short story; the long story lies elsewhere — in everything that never took place.

Because intelligence has always been the most rationed resource in economic life. Expensive, slow, housed in a few heads and a few counted hours. So we sorted. Every organisation is, in its own way, a machine for economising on intelligence: analysis for the decisions that justify it, habit for the rest — and for entire territories, nothing at all. The questions no one asked, the documents no one read, the contradictions no one had the means to look for. Not out of negligence: out of arithmetic. And scarcity ended up disguising itself as the natural order of things — one doesn't willingly look at what one can't afford to read.

That is what abundance truly changes. It doesn't replace what exists: it investigates what never was. Data overflows; analysis becomes commonplace; judgment alone remains rare — a world where analysis is free is a world where judgment is the only differentiator left. But the real frontier doesn't run where the machine does better: it runs where, yesterday, no one did anything at all.

And since this abundance will be the same for everyone — vast, available, better next month — the difference between two organisations, two funds, two leadership teams will not be decided by access. It will be decided by what they choose to open, and by the shape they give it. The current is identical in every city; the cities themselves look nothing alike.

III · The lesson of the dynamo

History has already given the measure of what's coming, and it is counter-intuitive. The economist Paul David established it for electricity: four decades separate the first power plant from the productivity gains expected of it. Forty years — and not chiefly because the machine lacked power, but because it had been wired into the old form: one large central engine set where the steam engine used to stand, the same transmission shafts, the same floors stacked around the axis. The gains came the day people stopped installing electricity and started designing the factory around it — one motor per station, the workshop laid flat, the flow rethought. And what the redesigned factory produced was not the old factory made cheaper: it was what steam had never allowed — the assembly line, the precision, the series.

The lesson holds in one sentence: a new capability doesn't deliver on its promise as long as it is asked to redo the old world faster; it delivers when something that never existed is designed around it. Handing a frontier model to an organisation that isn't expecting it doesn't make that organisation more intelligent: it produces the plausible, faster.

IV · The discipline that's coming — AI design

This design is a labour, and that labour is becoming a profession.

What has to be designed is not the intelligence. It is already there, identical for everyone, better next month — no one holds a lasting advantage on that ground. What has to be designed is its interface: where it enters the organisation, what it opens that wasn't being done before, what it has the right to touch, what it must report back, what it never concludes, and who answers for what it produces.

This profession has a name: AI design — the design of intelligence itself. Design in its strict sense, the architect's sense rather than the decorator's: designing the relationship between a capability and the humans who must live alongside it. The discipline demands two halves that almost no one holds at once: knowing what a chain of intelligence can do, and knowing what a human collective can carry. The first half expires every six months; the second is as old as organisations themselves.

The world that's coming will offer three ways to get it wrong. The craftsman without a machine — sincere, conscientious, and exposed to being outpaced in depth of investigation by a better-equipped rival. Software without judgment — the score no one signs, the decision no one owns. And the most common, because it looks like modernity: deployment without design — abundance wired into the old form, ordered to redo yesterday faster: the dynamo set on the steam engine's plinth. The pincers close from both sides — the frontier beats the craftsman; governed judgment beats the software. Between the two, design decides.

V · What design refuses

The more important half of a design is negative: an interface is judged by what it forbids.

In decisions that touch people — assessing, promoting, entrusting, backing a team — the prohibitions aren't caution: they are the source of the value. The machine investigates everything; it never assesses a person. No decision is automated. The process is consented to, and its findings are returned to those who took part. What these refusals produce appears in no comparison chart: material gathered honestly — because no one opens up to a system that grades them —, legal protection for everyone involved, and the exact definition of what is being bought: a human conclusion, committed, signed, and answerable before those who decide.

And a design is never finished. The frontier moves every month; what it allows, what it costs, what must be refused to it move with it. The state of the art is not a state. It is upkeep.

VI · The question of Tuesday morning

Return to that eight-o'clock room. What makes it possible is not the machine: in five years, it will be everyone's, like the current. What makes it possible is a design — choices made months earlier, in cold blood: what the night has the right to investigate, what only the morning can decide, what the chain reports back and to whom, what it will never conclude, and the name of whoever signs. That design cannot be improvised on closing day, nor on the night of a crisis. It is conceived beforehand — like the factory built around the dynamo.

Because what this room holds most precious is not what now happens faster: it is what finally happens at all — the file that was never investigated, the question one didn't dare pay for, the contradiction one never had the means to look for.

In five years, your organisation will be immersed in the same intelligence as every other — and the scope of the possible will have changed size for everyone. One question alone will set it apart: who is its designer?