On Presence, Inequality, and What AI Could Return to Us
There is a particular kind of quiet that belongs only to early mornings on a beach. The air is cool, the light golden and tentative, and the sea goes about its ancient business of arriving and retreating, entirely indifferent to the schedules of men. It was on such a morning, from a balcony overlooking the shore, that I watched a small drama unfold, one that has refused to leave me since.
Two dogs were playing at the edge of the water. They splashed, barked at the foam, chased shapes that existed only in their own electric imaginations. With each wave that broke upon the shore, they would tear down the wet sand toward it, skid to a halt, spin around as if the whole thing had been a magnificent trick, and race back into the shallows. Each wave, for them, was the first wave there had ever been.
A few yards away, three boys, perhaps ten years old, at that perfect age before self-consciousness sets in, were lost in a game of splash with no rules, no winners, no score. They were not at the beach. They were the beach, entirely absorbed into it.
In the foreground, adults walked and jogged with the purposeful efficiency of people who have converted even leisure into a task to be completed. A cluster of elderly men debated politics with the intensity of people trying to solve something urgently and permanently, men who, in another life, were once boys exactly like those in the water. Then, as the sun climbed, the adults drifted away one by one, back to the day’s real business. The boys and the dogs remained.
I have been asking myself ever since: when exactly did we leave? Not the beach, but that state of being, that radical, unselfconscious presence, that capacity to treat each wave as if it were the most astonishing thing that had ever happened.
We are not born anxious. The infant, the ten-year-old, the dog are creatures of pure now. Somewhere between childhood and adulthood, we begin to narrate our lives rather than live them, forever reliving the past or rehearsing the future, rarely arriving in the present. But here is what we rarely examine: this is not simply a failure of individual will. It is, in large part, a failure of circumstance. The world most of us inhabit was not built for human flourishing. It was built for human productivity. And for most of human history, the only people who could afford the difference were the ones who did not have to do the productive work themselves.
We talk about inequality in the language of income and assets. These matter enormously but they are proxies for something deeper. The real inequality is attentional. A child born into wealth has access not just to better schools and better doctors but to something far more fundamental: adults with time. Time to be present, to notice, to respond to what is actually happening in front of them rather than what is next on the list. A parent working two jobs to keep the lights on is not less loving. She is simply less available to herself as much as to her children. The farmer rising at four in the morning to tend crops by hand is not less curious. He simply has no bandwidth left for curiosity. The mechanical drudgery that fills the day is not merely tiring. It crowds out the cognitive and emotional space in which a life is actually lived.
Presence, in other words, has always been rationed by class. The capacity to be fully alive, to treat the wave as if it were the first wave, to be absorbed rather than managing, to choose what to attend to rather than merely react to what is urgent has been, throughout recorded history, a luxury available mainly to those who could afford to delegate the mechanical labour of survival to others. Philosophy, contemplation, the arts, the examined life: these did not flourish despite slavery and serfdom. They flourished, in part, because of them. The leisure of the few was purchased by the unfreedom of the many.
Artificial intelligence, if we are wise enough to use it this way, is the first technology in history that could change this arithmetic. Not by replacing human experience but by taking on the mechanical burden that has always prevented most people from having it. The work that filled our days but never lit us up: the repetitive, the administrative, the merely cognitive, AI absorbs this without resentment, without fatigue, and without the need for someone else to be diminished for it to happen. What it returns, in principle, is what the privileged have always quietly kept for themselves: time, attention, the raw material of a life consciously lived.
This is not a technological argument. It is a philosophical one, dressed in technological clothes. The question it asks is not what AI can do, but what it could return to us and whether we are honest enough about what we have been missing to receive it properly.
The equalising potential runs deeper than most analyses acknowledge, because what AI democratises is not merely access to information, the internet already did that, but the application of intelligence itself. A good lawyer has always cost money not because legal knowledge is intrinsically expensive to reproduce but because a trained human being has alternative uses for her time. A specialist physician, a financial adviser, a gifted tutor: what you are paying for, in each case, is not the knowledge but the scarcity of the person who holds it. AI collapses that scarcity at its root. The marginal cost of the ten-thousandth legal analysis approaches zero. The billionth student costs almost nothing more than the first. For the first time in history, the quality of expert judgment – contextualised, responsive, patient – can be decoupled from the accident of where and to whom you were born.
But near-zero marginal cost does not automatically mean near-zero price. This is the tension we must not paper over. The development and operation of frontier AI is genuinely capital-intensive – the compute, the energy, the research talent, the infrastructure. These costs are real and concentrated: only a handful of organisations can currently bear them. That concentration is not conspiracy; it is the natural economics of a technology with massive fixed costs and near-zero costs of deployment. The same structure produced the railways, the telephone networks, and the internet backbone. In each case, the infrastructure owners faced a choice – or given one by regulation – between pricing at marginal cost and passing the surplus to users, or pricing at what the market would bear and capturing the surplus as rent. The internet platforms chose rent. The price, paid in attention and data rather than money, did not fall toward zero as the marginal cost did. The value flowed to the infrastructure owner, not to the billions of users whose activity made the infrastructure valuable.
AI could follow exactly the same path and the stakes are higher, because what sits behind the toll gate this time is not a social network or a search engine but access to effective intelligence itself. Intelligence at near-zero marginal cost, owned by few and priced at what the market bears, becomes not democratisation but a new and more intimate form of extraction: one that sits not between you and a destination but between you and your own capacity to think, decide, and act effectively. The argument for democratisation therefore cannot rest on the technology alone. It requires a conscious decision, through regulation, public investment, or the deliberate adoption of a public-good model, to ensure that the surplus generated when intelligence becomes nearly free to deploy is not entirely captured as rent by those who bore the fixed costs of building it. The analogy is public libraries, not private bookshops: the same knowledge, available to everyone, because society decided that the returns to universal access outweighed the returns to restricted access. That decision was not made by the market. It was made against the market’s immediate incentives, in favour of a longer social logic. The same decision awaits us now, at considerably larger scale.
The case for making it is not primarily moral, though it is that. It is structural. The greatest waste of our civilisation is not oil or capital or even time. It is the unrealised potential of the billions of people who never had a fair chance to develop what they carried. Every year a child in genuine poverty fails to receive the attention her curiosity deserves is a year of compounding loss not just to her, but to everyone who will never know what she might have contributed. Superabundance that reaches everyone does not merely correct an injustice. It unlocks a quantity of human capability so large that its returns dwarf any conceivable cost of ensuring the access that produces it.
There is one more argument that any honest version of this vision must pass. Abundance built on a dying biosphere is not abundance. It is a last feast before the fire. The standard assumption has been that growth and environmental health are in fundamental tension — that development means consumption, and consumption means destruction. This assumption was accurate for the industrial era. It is not structurally necessary for the one we are entering.
The value that AI generates is largely dematerialised. A medical diagnosis consumes almost no physical resources. An educational exchange produces no carbon. Precision agriculture delivering exactly the right input to exactly the right plant at exactly the right moment is already reducing water use by thirty to fifty percent in early deployments while increasing yields. An AI-managed electricity grid can integrate renewables at scales previously thought impossible, turning the intermittency problem from a hard constraint into a coordination challenge. The old binary growth or sustainability is not repealed by technology. But for the first time it is not structurally enforced by it either. The things that matter most in a flourishing life, presence, relationship, purpose, the capacity to be absorbed by a wave are not resource-intensive in the way that steel and oil were. What we need to preserve the planet is not the abolition of desire. It is the redirection of it toward what we actually wanted all along.
None of this is inevitable. The same technology that could liberate human attention could more profitably capture and exploit it and the business models of the platforms that have most shaped our digital lives have, so far, been built on the second option rather than the first. The same abundance that could lift the dispossessed could further concentrate at the top. The same AI that could make our relationship to the natural world more sustainable could be used to extract resources more efficiently until there are none left. The difference lies entirely in what we decide to optimise for which is to say, in what we decide we actually want.
This is where the beach comes back. The boys were not trying to be present. They were not following a mindfulness protocol or a wellbeing index. They were simply doing what humans do when the mechanical burden of survival is lifted and no one has yet replaced it with a different set of demands: they showed up, completely, for what was actually in front of them. The design question — for institutions, for economies, for the AI systems we are now building into the substrate of daily life — is whether it is possible to create the conditions in which more people, for more of their lives, can do the same. Not as an aspiration. As a measurable target.
That is what is actually at stake in the choices being made now about how artificial intelligence is built, owned, and deployed. Not productivity. Not GDP. Not even, primarily, equality though equality matters and must be fought for. What is at stake is whether the extraordinary capacity being assembled in these systems is directed toward the liberation of human attention or its further colonisation. Toward returning to the many what the few have always quietly kept. Toward a world in which the wave, arriving and retreating, belongs — as it always did — to everyone who is present enough to receive it.
Not by accident.
By design.