For almost twenty-five years, I worked inside a global corporation. I valued the people, the challenges and many of the opportunities it gave me. But there was something I rarely stopped to question: the idea that a professional life should be organized around employment. It was simply the water I swam in. I have always enjoyed working and fully expect to continue. What I had taken for granted was the particular arrangement in which we give an organization much of our time, energy and imagination in exchange for a salary, a place in the hierarchy and, if all goes well, a degree of security.
There is nothing inherently wrong with this bargain. Corporations have made remarkable things possible, including many of the most rewarding experiences of my professional life. Yet from a little distance, I am struck by how completely we have allowed this relatively recent invention to define a normal life. We organize our families around school calendars and office hours, introduce ourselves by our occupations, and save for a retirement when our time will finally belong to us.
As historian E. P. Thompson showed in his study of time and work discipline, an older world of task-oriented rhythms increasingly gave way to the clock. Factories, railways, schools and offices required synchronization. We organized a culture around the needs of production. As intelligent machines change those needs, it seems worth asking how much of that culture we want to carry forward.
An extraordinary arrangement, mistaken for human nature
For most of history, people worked without necessarily having jobs. They cultivated land, practiced trades, raised families, built, repaired and cared. This was no lost paradise of autonomy: farming was exhausting, dependence was widespread, and wage labour existed long before the Industrial Revolution. But the expectation that adult life should revolve around a salaried career, a pension and eventual retirement is comparatively new.
Industrial capitalism brought work, employment, income and identity into one powerful social bargain. The four became so closely entwined that “What do you do?” became the default proxy for “Who are you?”
Energy helped make it possible. Earlier societies relied on land, wood, wind, water, and human and animal muscle. Coal, followed by oil and gas, opened access to concentrated stores of energy accumulated over geological time. Machines multiplied what labour could achieve. In an important sense, the modern economy was built on fossil energy as much as on human ingenuity. But factories, finance, materials and distribution needed coordination. Ronald Coase explained why firms emerge: finding partners and negotiating through markets can cost more than organizing people internally. Alfred Chandler later showed how managerial hierarchies helped enterprises operate at scale. The corporation became a remarkably effective way of bringing people and productive assets together.
At its best, the arrangement delivered rising wages, cheaper goods and greater material security. Trade unions, education, regulation and welfare states helped spread the gains, though never automatically or equally. The result was historically astonishing: longer lives, greater security and a widespread expectation that the next generation could live better than the previous one. It is worth remembering how much this arrangement achieved before imagining what might follow it.
Over time, the bargain shaped much more than the economy. Education prepared us for jobs. Mortgages, pensions and public finances relied on wages. Careers acquired the structure of biographies, with a beginning, a climb and a planned ending. Continuous employment became a measure of human worth.
The seams of the great bargain
The bargain has changed before. Globalization brought cheaper goods, new markets and enormous benefits to many emerging economies, while weakening the connection between prosperity in one country and the jobs needed to sustain it. Automation reduced labour needs in some activities and expanded them in others. Growth continued, but its connection to the security of a particular worker or community became less dependable.
Nor does higher productivity automatically become higher pay or more free time. Who benefits depends on ownership, bargaining power and political choices. A 2025 ILO–OECD analysis found that labour’s share of global income fell by 1.6 percentage points between 2004 and 2024. That does not mean all workers became poorer. It does mean that growth alone tells us little about how its rewards will be shared.
Demography adds another twist. Japan’s population is already shrinking, and the OECD projects that the population aged 20–64 will decline by about 8% across its members by 2060. In ageing countries, intelligent machines may first arrive as an answer to a shortage of people available to do the work. Longer lives are a triumph, but a hundred-year life does not necessarily mean a hundred years of good health or economic activity. The familiar sequence of education, four decades of employment and retirement looks increasingly awkward. We may need to learn, earn, care and contribute in several different chapters.
There is also a more familiar, less measurable strain. Much of professional life involves coordinating, documenting, supervising and measuring other activities. Much of it is necessary. Yet we can spend a week keeping an organization running without being entirely sure what human problem we have helped solve. Being busy and feeling useful are not always the same thing. The discomfort is not necessarily about working too hard. It can come from having little say over what the effort is for, or being unable to see what it amounts to.
Marx’s machine, and the question of who owns it
Karl Marx understood that distance between effort and meaning long before the modern office existed. His idea of alienation concerned work performed under someone else’s direction, for purposes over which the worker had little control. A person can work very hard at a self-chosen pursuit and feel alive; another can be comfortable and well paid yet experience surprisingly little ownership of the day.
More striking today is the argument he developed in the Grundrisse: capitalism constantly seeks to make production less dependent on direct human labour by applying machinery, science and accumulated knowledge. If more of society’s wealth comes from that shared stock of knowledge, why should someone’s access to prosperity depend so heavily on how much labour they can sell?
AI brings that question closer. It is beginning to perform parts of the coding, analysis, translation, design and administration that grew around industrial production. This does not tell us how many jobs will disappear. Tasks are not occupations, new capabilities create new demand, and much work depends on human presence, relationships and trust. But if output becomes less dependent on human labour, wages alone may become a less reliable way to share its rewards.
There is a further possibility that particularly interests me. If AI lowers the cost of accounting, marketing, software and coordination, some of the economics that favoured large corporations could work in reverse. Small teams could gain capabilities that once required an elaborate hierarchy. More people might be able to turn an idea into a functioning business.
Yet their independence could be deceptive. The technology rests on chips, data centres, energy and enormous capital investment. The International Energy Agency expects data-centre electricity use to nearly double between 2025 and 2030. We could end up with millions of small enterprises operating on infrastructure owned by a handful of enormous firms. Business ownership might spread while control over the tools those businesses need becomes more concentrated.
That is what makes Marx’s question worth revisiting. If machines generate more wealth, the difference between owning productive assets and selling one’s time may matter even more. Some of the people building these systems recognize the problem. In Moore’s Law for Everything, Sam Altman proposed an American Equity Fund to share some of the returns more broadly. The proposal is open to debate, but the question behind it is hard to dismiss. It is striking that someone helping build this technology would also question whether wages alone can remain an adequate way to distribute prosperity.
There are several ways to broaden participation: employee ownership, public investment funds, social dividends and wider access to capital. Alaska and Norway offer different precedents for sharing returns from valuable assets, though neither supplies a ready-made model for AI. Sharing the returns does not require the state to own or operate every enterprise. The question is how more people can participate in the wealth those enterprises create. We do not need to agree on one solution to ask whether access to prosperity should depend so completely on finding someone willing to employ us.
Keynes’s unfinished promise
John Maynard Keynes approached the matter from another direction. In 1930, he predicted that technological progress might make a fifteen-hour working week sufficient to meet people’s material needs. He also wondered what we would do with ourselves once earning a living occupied so much less of our time.
His forecast was not as absurd as it sounds. Historical estimates assembled by Our World in Data suggest that annual hours per worker fell from about 3,195 in Switzerland in 1870 to 1,529 in 2023. These estimates are not strictly comparable across the whole period, but the scale of the change is striking. Weekends, paid holidays, longer education and retirement are ways in which productivity gains have already been converted into time.
We also used those gains for more consumption, better services, profits and public spending. Expectations rose, unpaid care remained necessary, and some work became more intense. Productivity creates the possibility of leisure; it does not decide how much we take or who receives it.
Nor does abundance remove every scarcity. AI may make knowledge work cheaper while desirable land, attention, trust and intimate relationships remain limited. A world in which information is nearly free but a home near a good school is unaffordable is easy to imagine. Producing more will not settle what we value or what we are willing to share.
Creatures and machines
Wendell Berry warned against treating living things as machines to be optimized. David Mattin has explored a related division between creatures and machines: those eager to transcend our biological limits and those determined to preserve something embodied, vulnerable and human.
I see no reason we cannot use intelligent systems to relieve administrative drudgery while making more room for craftsmanship, nature and the unhurried complexities of human relationships. We can welcome longer lives without turning existence into an endless optimization project. More intelligence does not automatically give us a wiser idea of a good life.
This is also why income transfers, however useful they might be, would only answer part of the problem. Money can provide security. It cannot by itself provide responsibility, recognition, friendship or belonging. A society able to produce everything but unable to make its people feel needed would have solved only half the problem.
Hannah Arendt distinguished between labour, which sustains life’s recurring necessities; work, which creates more durable things; and action, through which we take initiative with others. Employment can contain all three, but has no monopoly on them. Raising children, caring for parents, restoring a house, mentoring someone or building a business can demand enormous effort and create real value. The absence of a salary does not make a contribution less important.
Hard work can be deeply satisfying. What seems questionable is allowing one particular contract to become our default measure of contribution, achievement and dignity.
An economy of agency
None of this guarantees that the corporate model will disappear. AI could make employees more capable and create occupations we have not yet imagined. It could help more people build businesses, or make them more dependent on a few technology owners. Much will depend on choices about ownership, competition and social protection that the technology itself cannot make for us.
And independence is not automatically freedom. Freelancers competing for assignments on platforms they do not control could be more precarious than employees. Freedom from an employer can become dependence on an algorithm. Equally, a comfortable income without purpose or a voice in society would be a poor version of progress.
These are practical questions as much as philosophical ones. How do people build assets rather than only sell their time? How do social protections adapt when careers become less linear? How do we recognize worthwhile activity that never appears on a payroll?
What I would hope for is greater security alongside more say over how we spend our lives. Some people will want to build enterprises or pursue a craft. Others will move between employment, learning, caregiving and community work. Longer lives may make these changing chapters more natural, without every transition being treated as a failure of the career plan. People might work intensely at one stage and differently at another, with more room between full-time employment and complete retirement.
That possibility has become more than theoretical for me. After much of my professional life inside a large institution, I am exploring a more independent mix of strategy, entrepreneurship, investment and collaboration. I have no desire to stop working. Quite the opposite. What interests me is the possibility of choosing work more deliberately, and of judging a good year by the things I have helped create, the people I have worked with and the freedom I have preserved, rather than by a title on an organization chart.
I am fortunate to have choices, and I know many people do not. That is precisely why the broader question matters. Work, employment, income and identity need not remain tied together quite so tightly. People will still want to create and contribute; many will do so inside large organizations. Everyone will need an income and a sense that their life matters. A job can provide both, but we should be able to imagine other possibilities too.
For two centuries, we organized human life to serve production. Perhaps the next great civilizational project is to organize production to serve human life.
Sources and further reading
• E. P. Thompson, “Time, Work-Discipline, and Industrial Capitalism”, Past & Present (1967); E. A. Wrigley, Energy and the English Industrial Revolution (2010).
• Ronald Coase, The Nature of the Firm (1937); Alfred D. Chandler Jr., The Visible Hand (1977).
• United Nations, World Population Prospects 2024; OECD, Employment Outlook 2025.
• ILO and OECD, Policy Measures to Address Inequalities and Increase the Labour Income Share (2025); IMF, Global Debt Remains Above 235% of World GDP (2025).
• Our World in Data, Annual Working Hours per Worker; IEA, Key Questions on Energy and AI.
• Karl Marx, Economic and Philosophic Manuscripts of 1844 and Grundrisse (1857–58); John Maynard Keynes, Economic Possibilities for our Grandchildren (1930); Hannah Arendt, The Human Condition (1958); Wendell Berry, Life Is a Miracle (2000).
• Sam Altman, Moore's Law for Everything (2021); Elon Musk, remarks at VivaTech (2024); Dario Amodei, Policy on the AI Exponential (2026); David Mattin, Creatures and Machines (2026).