After Scarcity: What Becomes of Us When the Economic Problem Is Solved

The economic problem is not, if we look into the future, the permanent problem of the human race. The permanent problem is how to use his freedom from pressing economic cares, how to occupy the leisure which science and compound interest will have won for him, to live wisely and agreeably and well.

>

John Maynard Keynes, 1930

Picture an AI investment review in a large company this year. The business case counts hours saved, headcount avoided and cycle time reduced. It rarely asks who gets the saved time, who absorbs the transition, or what the people whose work disappears will do instead. Those look like policy questions, far from a capital committee. I think they are becoming operating questions, and this essay explains why.

It condenses *After Scarcity*, a thirty-five-page report I wrote in a personal capacity in May 2026. The full PDF, with charts and the complete reference list, is linked at the bottom of the page.

The argument

My argument has three parts.

First, abundance moves scarcity rather than ending it. As material goods become cheap, the binding constraints shift to positional goods such as status, attention and the best land, to time, and to purpose.

Second, the decisive variable over the next twenty to thirty years is distribution rather than technology. Cash transfers reliably reduce hardship without causing mass idleness, but they do not supply meaning, and the major institutional forecasts expect automation to take over a large share of task time. Whether AI rents are captured publicly or concentrated privately separates the hopeful paths from the bleak ones.

Third, over the long run Keynes's "permanent problem" of how to live well becomes the operative one. The evidence from psychology, anthropology and philosophy points the same way: people flourish when they exercise autonomy, competence and relatedness in pursuit of ends they choose, and institutions have to be built deliberately to keep those conditions in place.

Keynes was right about wealth and wrong about leisure

In 1930, with unemployment in the United Kingdom near twenty percent, Keynes wrote *Economic Possibilities for our Grandchildren*. He predicted two things: living standards in advanced economies would rise four- to eightfold by 2030, and the wealth would buy leisure on an unprecedented scale, including a fifteen-hour workweek.

The first prediction has held. Real GDP per capita in advanced economies has risen close to his forecast path. The second has not. Average weekly hours fell during the twentieth century, then stabilised far above fifteen and have plateaued near thirty-four for several decades.

The MIT Press volume *Revisiting Keynes*, with contributions from Stiglitz, Phelps, Freeman, Frank, Becker, Solow, Friedman and Baumol, identifies two errors. Keynes assumed relative wants would fade as material conditions improved, and they intensified instead. He also treated work as pure disutility, when work also supplies status, identity, structure and what Csikszentmihalyi later called flow. Robert Frank's chapter puts it plainly: humans like having something to do.

So the partial abundance already reached in rich economies has not produced a leisure society, and there is no empirical reason to expect more abundance to produce one without deliberate design.

Bounty and spread move independently

Brynjolfsson and McAfee's *The Second Machine Age* (2014) separates bounty, the growth of the total pie, from spread, how widely it is shared. Their account of the Great Decoupling, the divergence between productivity and median compensation that began in the early 1970s, is the strongest evidence that automation can deliver abundance in aggregate without delivering it to anyone in particular. They point to three engines: skill-biased technical change, a rising capital share of income, and winner-take-most dynamics in digital markets. Kodak, with 145,000 employees, went bankrupt; Instagram, with fourteen, sold for a billion dollars.

Daniel Susskind's *A World Without Work* (2020) argues that as machines acquire general capabilities, the tasks where humans keep a comparative advantage shrink toward zero. He proposes a conditional basic income tied to civic obligations and a "meaning-creating state". Carl Frey's *The Technology Trap* (2019) adds the historical warning: the Lancashire weavers of the 1810s were measurably worse off for a generation before mechanisation's gains spread.

The institutional forecasts broadly agree. McKinsey Global Institute estimates that half of current paid activities could in principle be automated with existing technology, and that thirty percent of hours worked globally could be automated by 2030 in a midpoint scenario. The OECD puts twenty-seven percent of jobs at high risk. Brookings stresses the uneven geography, and RAND's scenario work shows outcomes by 2040 diverging sharply by policy design. Across these sources, institutions separate the futures more than the technology does.

The scarcities technology cannot remove

Fred Hirsch's *Social Limits to Growth* (1976) distinguished material goods from positional goods, whose value is relative or congestion-prone: a Rembrandt, a beach house in the Hamptons, a Harvard admission, the corner office. As affluence grows, competition shifts toward goods that cannot be made universally available. A corner office that everyone has is no longer a corner office.

Robert Frank extended the idea in *Luxury Fever* (1999) and *Falling Behind* (2007). The spending of the wealthy sets reference points that cascade down through every income tier. Abundance does not stop that race. It moves it to new arenas, such as attention and rare experiences.

Veblen's *Theory of the Leisure Class* (1899) is the older record. The closest natural experiment we have to a person freed from economic necessity is an aristocrat with inherited wealth, and Veblen found that such people spent their freedom on conspicuous consumption, conspicuous leisure and consumption by proxy. Their first move was toward distinction rather than contemplation.

Post-scarcity, properly understood, is therefore material post-scarcity with a positional and experiential economy on top of it, trading in status, attention, time, taste, access, reputation and, in Bostrom's terms, meaning.

What unconditional cash actually does

The closest empirical analogue to partial post-scarcity is the basic income experiment. The OpenResearch Unconditional Cash Study, funded mainly by Sam Altman and released in July 2024, is the largest randomised trial so far. One thousand participants in Texas and Illinois received $1,000 a month for three years, and two thousand controls received $50 a month. All had baseline incomes below $29,000. Hours worked fell by about 1.3 a week, roughly two percent, with no exit from the labour force. Healthcare use rose, with hospitalisations up twenty-six percent and emergency visits up ten percent. Stress, mental distress and food insecurity fell in the first year, and those gains faded by years two and three. The research team concluded that cash alone cannot address chronic health conditions, lack of childcare or the high cost of housing.

The Finnish trial (2017 to 2018, two thousand unemployed people aged 25 to 58, €560 a month tax-free) found the same pattern from another angle. Its principal investigator, Olli Kangas, summarised it as "no significant effects on employment, but important effects on well-being." Stockton's SEED demonstration (2019 to 2021, 125 residents, $500 a month) and GiveDirectly's twelve-year village-level trial in Kenya point the same way. The Alaska Permanent Fund Dividend has paid every resident a variable annual dividend since 1982 with no detectable employment reduction (Jones and Marinescu, *AEJ: Economic Policy*, 2022).

At the modest levels tested, unconditional cash is a humane and effective antipoverty tool. It does not produce mass idleness, and on its own it does not produce flourishing either. The fading of the OpenResearch mental-health gains is the most useful signal we have about the meaning problem: remove material stress and something else is still needed.

A May 2025 arXiv working paper, *An AI Capability Threshold for Rent-Funded Universal Basic Income in an AI-Automated Economy*, estimates how productive AI would need to be for publicly captured AI rents to fund a meaningful basic income. The threshold looks plausible, but it depends on rent-capture mechanisms such as taxation, public ownership or sovereign AI wealth funds that do not yet exist at scale. I think this is the most important policy variable of the next two decades.

The psychology of abundance

Brickman, Coates and Janoff-Bulman's 1978 study of twenty-two lottery winners and twenty-nine people recently paralysed in accidents found that the winners "were not happier than controls and took significantly less pleasure from a series of mundane events." Later work softened the strong set-point claim. Diener, Lucas and Scollon's 2006 review showed that set-points differ between people, are partly heritable and can shift, and that disability lowers well-being by about three-quarters of a standard deviation over the long run. Adaptation to misfortune is less complete than adaptation to fortune.

The macro counterpart is Richard Easterlin's 1974 paradox, updated in a 2020 IZA paper. Across countries and decades, rising GDP has not tracked rising happiness. In China between 1990 and 2015, income per head rose roughly tenfold while life satisfaction did not rise with it, and Japan between 1958 and 1987 and India between 1995 and 2019 show similar gaps. Kahneman and Deaton (2010), Killingsworth (2021) and their 2023 adversarial collaboration converged on a middle position: well-being keeps rising with income for most people but plateaus for a substantial unhappy minority whose problems money does not solve. Material abundance helps most people somewhat, helps nobody without limit, and leaves a large group untouched.

The strongest positive framework is Self-Determination Theory. Deci and Ryan identify three psychological needs, autonomy, competence and relatedness, that have to be met for people to flourish, and their foundational 2000 paper has been cited more than eighty thousand times. None of the three depends on scarcity, so an abundant society can deliver all of them. It can also erode all of them. Automation can leave little worth being good at, guaranteed provision can feel like being a recipient rather than a contributor, and digital substitutes for community can thin out relationships while appearing to multiply them. Whether abundance supports flourishing is therefore a design question.

Csikszentmihalyi's research on flow adds one more requirement. Sustainable satisfaction comes from voluntary engagement with tasks at the edge of one's capability. If AI handles every real challenge, the open question is whether challenges designed for us still feel like real ones.

What philosophy and history add

Aristotle distinguished pleasure from eudaimonia, flourishing through the exercise of virtue and reason over a whole life, and most serious thinking about abundance still works inside that distinction. Viktor Frankl's *Man's Search for Meaning* (1946) located meaning in creative work, in love and in the attitude one takes toward unavoidable suffering, and he coined "existential vacuum" for the emptiness of the bored and affluent. If automation removes much of the first and engineering reduces the third, love and self-chosen challenge carry more of the weight.

Buddhist and Stoic traditions offer the most developed practical answers, and research on long-term meditators (Davidson, Goleman, Lutz) suggests the hedonic treadmill can be trained. Seneca argued that virtue, rather than abundance, produces happiness. Education in attention and in managing desire may matter in an abundant century the way literacy mattered in the industrial one.

Two bodies of historical evidence frame the question. Sahlins's *Original Affluent Society* (1966, expanded in *Stone Age Economics*, 1972) argued that hunter-gatherers reached affluence by wanting less and worked roughly fifteen to twenty hours a week on subsistence. The thesis has been qualified, but its core survives: people without economic compulsion do not collapse into apathy. They tell stories, cultivate kinship, hold rituals, play, explore, fight and care for each other. That is a floor for optimism, though it comes from small groups rather than a civilisation of eight billion.

The aristocratic record is the counterweight. Leisured elites across societies have produced fashion, gambling, duelling, addiction and dynastic intrigue alongside patronage and philanthropy, and folk sayings in several languages describe fortunes lost within three generations: "shirtsleeves to shirtsleeves" in English, "stable boy to stable boy" in Italian, "rice paddies to rice paddies" in Japanese. Gibbon on Rome, Ibn Khaldun on asabiyyah and Toynbee's challenge-and-response model all describe societies that lost cohesion when they lost their challenges. That recurring pattern is the main reason to treat abundance as a risk that needs counter-design.

The next thirty years

The next three decades are the transition rather than the destination, and three variables decide which path the advanced economies take.

The first is capability. McKinsey, OECD, Brookings, RAND and PwC broadly expect thirty to fifty percent of tasks, rather than jobs, to become automatable within twenty years. McKinsey's 2023 update raised its automation estimate by about ten percentage points as white-collar cognitive work moved into scope earlier than its 2017 reports expected.

The second is the distribution of AI rents. If productivity gains are partly captured publicly, through sovereign AI wealth funds, AI dividends, meaningful taxation of automated production or effective antitrust action, the path bends toward shared abundance. If they are captured privately, which is the current default, it bends toward a thin ownership class, a hollowed middle and a precarious majority living on transfers.

The third is meaning. The fading of the OpenResearch gains suggests that basic income without investment in institutions that give people purpose becomes a maintenance dose rather than a remedy.

These are my directional estimates for 2026 to 2055. They are judgments, not model outputs.

The leading indicators point mostly toward the baseline and pessimistic range.

None of these is destiny, and several could be reversed by policy within a decade. The absence of a coherent counter-program in any major democracy is itself a sign that the transition is starting without a plan.

The most useful single test is this. If by 2030 the top one percent income share in major economies is still rising and effective tax rates on capital are not, the pessimistic branch is locking in. If at least two G7 economies have built serious mechanisms for AI rent capture by then, the outlook improves substantially.

The long arc

Beyond the transition, the question becomes Nick Bostrom's. *Deep Utopia* (2024) separates three layers that are often conflated. Shallow utopia is material post-scarcity. A deeper layer is post-work, where nothing needs to be done for pay. Deep utopia is post-instrumental: AI does everything people do, including art, science, care and friendship, better than people do it, and the claim that one's effort produces value the world would otherwise lack stops being true.

Bostrom's partial answers are tentative. Engineered enjoyment risks wireheading. AI-designed challenges risk feeling hollow once people know they were designed. An appreciative orientation asks us to love the good even when we do not produce it. Enhancement of human nature raises Parfit's question of whether the beings who flourish are still us. Social entanglement is the answer I find most plausible: relationships, kinship, raising children and mutual care keep their meaning when they produce nothing external, because producing things was never their main purpose.

The speculative long-term branches run from an Aristotelian civilisation that channels status competition into real accomplishment, through transhumanist enhancement, Robin Hanson's *Age of Em* (where easy copying drives digital minds back to subsistence wages despite abundance), inward-turned virtual worlds and outward Kardashev-scale expansion, to decadence and slow demographic fade. South Korea's total fertility rate of 0.7 in 2024 is often read as an early sign of the last branch, which the futurist literature treats as a serious possibility rather than a fringe one.

What sustains human engagement when economic necessity disappears? The evidence and the philosophy put relationships and care first, followed by curiosity and learning, creativity and craft, status competition channelled into excellence, spiritual and contemplative practice, play, stewardship and self-cultivation. Bernard Suits defined a game as the voluntary attempt to overcome unnecessary obstacles, which may describe much of what post-instrumental people will need. The mix will depend on culture, education and the choices made during the transition.

The strongest objection

A business reader could reasonably say this is too far from their decisions to matter. Strict post-scarcity is closer to a thought experiment than a forecast, the long-term section is speculation, and the basic income evidence covers a few thousand people over at most three years.

All of that is true, and the argument should be read within those limits. The near-term part does not depend on the speculation, though. The distribution of productivity gains, the share of work automated in a given function and the effect on the people whose tasks disappear are observable now, inside ordinary investment decisions. That is the part I would ask leaders to act on.

What leaders can do now

For policy, three observations follow from the evidence. Treat the capture of AI rents as a first-order question, because it separates the optimistic and pessimistic scenarios more than any technical variable. Scale basic income tests toward whole-community pilots of the GiveDirectly kind, which can measure the social and meaning effects that individual pilots miss by design. Invest in institutions that give people purpose, such as civic associations, contemplative communities, arts and sport, third places and intergenerational care, because the OpenResearch result suggests people need those structures in place when material stress falls away.

Over longer horizons, education built around autonomy, competence and relatedness, progressive consumption taxes of the kind Robert Frank proposes, and frontiers worth pursuing in science, ecology and moral repair all look like infrastructure for meaning rather than luxuries.

Inside a company, the same logic applies at smaller scale. An AI business case that counts only unit cost leaves out who gains time, income and authority, who absorbs the transition, and which new bottleneck replaces the one the technology removes.

Monday move

Add one question to your next AI investment review: if this works exactly as promised, who gains time, income, authority or optionality, and who loses it? Write the answer next to the productivity estimate, and ask the sponsor what the organisation will do for the people on the losing side.

Caveats and limitations

1. The premise is contestable. Strict post-scarcity in the maximal sense is closer to a thought experiment than a forecast, and constraints such as positional goods, attention, time and unique experiences cannot be automated away.
2. The time horizons differ in epistemic status. The twenty-to-thirty-year section rests on empirical evidence from basic income trials, automation research and well-being studies. The long-term section is philosophy and educated speculation.
3. The source pool is skewed. The post-scarcity literature is dominated by Western, mostly English-speaking, mostly secular and mostly male authors. Non-Western perspectives, such as Confucian *ren*, Ubuntu reciprocity and indigenous stewardship ethics, are underweighted here.
4. Basic income evidence tops out at three years and a few thousand participants, so saturation effects at population scale are unproven.
5. The audit assumes the broad AI capability trajectory continues. Faster or slower paths would change the timelines but not the structure of the argument.

A note on authorship

This article is my synthesis of a thirty-five-page independent research document I wrote in May 2026. It draws on AI tools for drafting, reference assembly and stress-testing the argument; the framing, the synthesis and every editorial judgement are my own. Independent research, written in a personal capacity. Personal website. Views are my own.

References

The full list of cited sources, in roughly the order they appear in the audit:

• Keynes, J. M. *Economic Possibilities for our Grandchildren.* 1930.
• *Revisiting Keynes* (Stiglitz, Phelps, Freeman, Frank, Becker, Solow, Friedman, Baumol). MIT Press.
• Brynjolfsson, E. and McAfee, A. *The Second Machine Age.* W. W. Norton, 2014.
• Susskind, D. *A World Without Work.* Allen Lane, 2020.
• Frey, C. B. *The Technology Trap.* Princeton University Press, 2019.
• McKinsey Global Institute. *Generative AI and the Future of Work.* 2023.
• OECD. *Employment Outlook 2023.*
• Brookings Institution. *Automation and AI: How Machines Are Affecting People and Places.*
• Hirsch, F. *Social Limits to Growth.* Harvard University Press, 1976.
• Frank, R. H. *Luxury Fever.* 1999; *Falling Behind.* 2007.
• Veblen, T. *The Theory of the Leisure Class.* 1899.
• OpenResearch. *Unconditional Cash Study.* 2024 (n=3,000; $1,000/month × 3 years).
• Kangas, O. et al. *The Basic Income Experiment in Finland 2017-2018.* 2019.
• West, S. et al. *Stockton SEED Demonstration Preliminary Analysis.* 2021.
• Jones, D. and Marinescu, I. *The Labor Market Impacts of Universal and Permanent Cash Transfers.* AEJ:EP, 2022.
• GiveDirectly. Kenya village-level saturation UBI trial.
• Bélisle-Pipon, J.-C. *AI-elite UBI advocacy as symbolic violence.* Frontiers in Artificial Intelligence, 2025.
• Brickman, P., Coates, D., Janoff-Bulman, R. *Lottery winners and accident victims.* JPSP, 1978.
• Diener, E., Lucas, R. E., Scollon, C. N. *Beyond the hedonic treadmill.* American Psychologist, 2006.
• Easterlin, R. A. 1974; updated IZA paper, 2020.
• Kahneman, D. and Deaton, A. PNAS, 2010.
• Killingsworth, M. A. PNAS, 2021. Killingsworth, Kahneman and Mellers. PNAS, 2023.
• Deci, E. L. and Ryan, R. M. Self-Determination Theory; foundational 2000 paper.
• Csikszentmihalyi, M. *Flow.* 1990.
• Aristotle. *Nicomachean Ethics.*
• Frankl, V. *Man's Search for Meaning.* 1946.
• Sartre, J.-P. *Being and Nothingness.* 1943.
• Camus, A. *The Myth of Sisyphus.* 1942.
• Parfit, D. *Reasons and Persons.* 1984.
• Pearce, D. *The Hedonistic Imperative.* 1995.
• Sahlins, M. *Stone Age Economics.* 1972.
• Hanson, R. *The Age of Em.* Oxford University Press, 2016.
• Bostrom, N. *Astronomical Waste.* 2003. *Letter from Utopia.* 2008. *Superintelligence.* 2014. *Deep Utopia.* Ideapress, 2024.
• Suits, B. *The Grasshopper.* 1978.
• Case, A. and Deaton, A. *Deaths of Despair.* Princeton University Press, 2020.
• Twenge, J. M. *Generations.* Atria, 2023.
• *An AI Capability Threshold for Rent-Funded Universal Basic Income in an AI-Automated Economy.* arXiv working paper, May 2025.