In 1930, with much of the world sliding into the Great Depression, the economist John Maynard Keynes wrote a short, strange, hopeful essay. It was called "Economic Possibilities for our Grandchildren."
Everyone around him was worried about the crash. Keynes asked them to look a hundred years ahead instead.
He argued that if productivity kept growing and there were no great wars, the standard of living in the advanced economies would be several times higher by 2030. The problem that had dominated human life for all of history, getting enough to eat and survive, would be largely solved. People would need to work far less. He suggested that three-hour shifts, or a fifteen-hour week, might be enough to satisfy the old human urge to be useful.
We are now very close to his deadline. So it is worth asking how he did.
What Keynes got right
On the big number, Keynes was remarkably close. Output per person in rich countries has grown enormously since 1930. A typical household in those countries today has comforts that would have seemed like luxury to almost anyone in his time: reliable electricity, clean water, cheap food from every continent, medicine that turns once fatal illnesses into inconveniences, and a device in your pocket that holds most of the world's knowledge.
He also saw something many economists missed. In the same essay, he used the phrase "technological unemployment" for jobs lost to machines faster than new uses for labour could be found. He thought of it as a temporary phase of adjustment, a sign of progress rather than doom. That debate is now back at the centre of every conversation about AI.
What he got wrong
The hours never fell the way he expected.
Working weeks did shorten for decades after his essay, and weekends became normal. But in most rich countries, full time work still means something closer to forty hours than fifteen. Many of the most educated and highest paid people work longer than that, by choice.
Why? There are plenty of economic explanations. Wants expanded as fast as incomes did. Things that were once luxuries became expectations. Some costs, like housing and education, grew faster than wages.
But there is a deeper reason, and Keynes half saw it himself. He worried that people raised for generations to strive would not know what to do with leisure. He imagined a kind of collective nervous breakdown, a world of people with free time and no idea how to fill it.
What he underrated is how much people get from work beyond money.
We kept working long after we needed the money, which tells you work was never only about the money.
What work actually gives you
Think about what a job provides besides a salary.
It gives structure. A reason to get up, a shape to the week, a sense that time is moving toward something.
It gives status. When you meet someone new, one of the first questions is what you do. For better or worse, the answer tells people where you stand.
It gives community. Many of your daily conversations, friendships and arguments happen at work. For a lot of adults, colleagues are the people they see most.
And it gives something harder to name: the feeling of being needed. Someone is depending on you to do your part. If you did not show up, something would go undone. That feeling is one of the deepest sources of dignity a person can have.
The philosopher Hannah Arendt drew a distinction in The Human Condition between labour, which keeps us alive and is consumed as quickly as it is produced, and work, which builds things that last, and action, which is how we appear to one another and shape a shared world. When people fear losing their jobs to machines, they are rarely afraid of losing labour alone. They are afraid of losing the other two.
The machines arrive
For most of history, new machines replaced muscle. The tractor replaced the plough horse. The loom replaced hand weaving. People moved into work that needed their minds.
AI is different in an important way. It is starting to do things that used to require thought: writing, summarising, coding, analysing, designing. Not perfectly, and not everything, but more each year.
This time the move is less obvious. When machines took over physical work, there was a clear next rung on the ladder: clerical work, professional work, knowledge work. Millions of families climbed it within a generation or two. Grandparents who worked the land had grandchildren who worked in offices.
It is harder to see the next rung when the machine is climbing the same ladder you are. Some new kinds of work will certainly appear, as they always have, and many will be hard to imagine from here. But it would be dishonest to pretend the transition will feel easy for everyone.
That raises the question in its sharpest form. If a machine can do much of what you were paid to do, what is left that only a person should do?
It is the wrong question to answer with a list of tasks, because any list will shrink. Better to ask what we want people to do even when a machine could do it.
What humans are for
A few answers seem likely to hold.
Responsibility. Someone has to own the outcome. When a decision affects real people, you want a person who can be held to account, who will lose sleep if it goes wrong, who signs their name to it. A system can recommend. A human has to answer for it.
Care. When your grandmother is ill, you do not just want her medication dispensed correctly. You want someone to sit with her, notice that she seems quieter than yesterday, and hold her hand. Much of the most important work in any society is care: raising children, nursing, teaching, looking after the old. Its value comes partly from the fact that a person chose to give it.
Judgment. Machines can produce options faster than you can read them. Deciding which ones matter, which ones fit, which ones are right for these people at this moment, is still a deeply human act. It draws on values, context and experience that no dataset fully captures.
Meaning. People write poems that no one will pay for, cook elaborate meals for family, learn instruments they will never perform, and build things just to see if they can. We do these because the doing is the point. A machine could produce a better poem. It could not make it yours.
In the Gulf, hospitality is a good example. Serving coffee to a guest is not about efficiently delivering caffeine. It is about the gesture, the attention and the relationship it expresses. A machine could pour the coffee. Nobody would feel welcomed by it.
A machine could pour the coffee. Nobody would feel welcomed by it.
Leisure is not idleness
The ancient Greeks had a word, scholē, that meant leisure. It is the root of the word school.
For them, leisure meant time free from necessity, available for thought, conversation, learning and taking part in public life. It was what labour was for.
Keynes understood this. The closing pages of his essay are about learning the art of life itself: valuing ends above means, the good above the useful, and taking delight in the small things of each day. He thought the people who would do best in an age of abundance were those who could enjoy life for its own sake.
That vision sounds lazy until you try to live it. It asks people to build meaning on purpose instead of having it handed to them by necessity.
A hopeful, serious question
There are real risks in the years ahead. Jobs will change faster than people can retrain. Gains from AI may flow to a few unless societies decide otherwise. Many people will feel, painfully, that they are no longer needed in the way they once were. None of that should be waved away.
But the question Keynes asked his grandchildren's generation was never only economic. He was asking what a good life looks like once survival is no longer the main problem.
For most of history, humans were for survival. We worked because we had to, and meaning came along with the struggle.
If machines carry more of that load, survival stops being the answer. We will have to choose one.
That choice might be the most human thing we ever do.
Further reading
- Essays in Persuasion — John Maynard Keynes
- The Human Condition — Hannah Arendt
- Leisure, the Basis of Culture — Josef Pieper



