A World Where We Make Music Instead of Rent

If machines take over the labor, the wealth they create could buy back human time. Here is the math, the evidence, and the outcomes we can reasonably expect.

Almost everyone has something unfinished. A song that stalled at the second verse. A novel in a drawer. A parent they meant to visit more. A friend going through a hard stretch who needed an afternoon they couldn’t spare. The reason is nearly always the same. Rent was due, the shift started at seven, and there was nothing left by the end of the day.

For most of history that trade was the price of staying alive. Someone had to plant, haul, build, file and fix. That premise is now weakening, and the question in front of us is not a technical one. It is who owns the output of machines that work without wages, and whether any of it comes back to people as time.

This piece makes a position argument, but it tries to earn it. It draws on randomized trials and natural experiments to show what people actually do when financial pressure lifts. It then builds a transparent scenario model of what an income floor would cost, how AI driven growth could pay for it, and what outcomes we should realistically expect. Every parameter is stated so you can audit the mechanics and rerun the numbers yourself.

THE LABOR IS ALREADY MOVING

In January 2024 the International Monetary Fund managing director, Kristalina Georgieva, said artificial intelligence will affect almost 40 percent of jobs worldwide, rising to about 60 percent in advanced economies. Her breakdown was sober. Roughly half of the exposed jobs may gain productivity, while for the other half AI could take over key tasks, lowering labor demand and hiring, and in extreme cases jobs may disappear. Goldman Sachs had earlier estimated that AI could disrupt more than 300 million jobs.

What sets this wave apart is the kind of work it touches. Earlier automation mostly took over routine physical and clerical tasks. Current AI systems can draft, summarize, analyze and carry out multi step reasoning that until recently required a trained analyst. That is why the IMF found advanced economies more exposed than developing ones.

Economists disagree sharply about how fast the output gains will arrive. At the cautious end, MIT’s Daron Acemoglu estimates the macroeconomic effects at no more than a 0.66 percent increase in total factor productivity over ten years. At the optimistic end, Goldman Sachs projects a 7 percent increase in global GDP over ten years and U.S. growth running 1.5 percentage points a year faster. The gap between those estimates is the gap between two different futures, and a serious policy proposal has to work across the plausible range between them or admit where it fails.

AN OLD PROMISE FINALLY WITHIN REACH

People saw this coming long ago. In 1891 Oscar Wilde argued that all dull and degrading labor must be done by machinery, leaving people free to create. In 1930 John Maynard Keynes predicted that living standards would rise four to eight times within a century and that his grandchildren might work around fifteen hours a week.

Keynes was roughly right about productivity and badly wrong about leisure. We got richer and kept working, largely because gains flowed to owners and to consumption rather than to time. As Acemoglu and Simon Johnson argue in Power and Progress, technology improves ordinary lives only when institutions force the gains to be shared. Abundance alone frees no one. Policy does. That is the core claim here, and the analysis that follows tries to show the policy is affordable.

EVIDENCE FROM WHEN THE PRESSURE LIFTS

The standard objection is that people freed from economic necessity would drift and produce nothing. We no longer have to guess. Randomized trials measure what happens to the individuals who receive support. Natural experiments like Alaska’s capture something trials cannot: what happens when a whole economy receives it at once.

In Ireland, artists given an income made more art. Under the Basic Income for the Arts pilot, 2,000 artists received €325 a week from September 2022 to February 2026, chosen at random from about 8,200 eligible applicants. After two years, recipients were spending 11 more hours a week on their creative practice and earning over €500 more a month from their art, and an independent study found reduced anxiety and depression. Independent analysis found that every euro invested returned €1.39 to the economy. Ireland has since made the scheme permanent, running it in three year cycles with 2,000 artists selected by lottery.

In Finland, economic security improved health and focus. In 2017 and 2018, 2,000 unemployed people aged 25 to 58 received €560 a month, unconditionally. The final evaluation found small employment effects, but recipients reported fewer stress symptoms, fewer difficulties concentrating and fewer health problems than the control group.

In Alaska, a permanent universal dividend did not reduce employment. Every resident has received an annual dividend since 1982. Using a synthetic control method, Damon Jones and Ioana Marinescu found no effect on employment and a 1.8 percentage point rise in part time work, about 17 percent. Their analysis suggests the cash stimulated the local economy, a general equilibrium effect that offset any individual tendency to work less.

In the United States, the largest randomized trial produced mixed results. In the OpenResearch study, 1,000 low income people received $1,000 a month for three years. Recipients cut their work by one to two hours a week, partners cut by a similar amount, and leisure was the largest gain in time use. On the other side, parents spent more on their children and reported better parenting, with the strongest effects among the lowest income and single parent households.

The underlying mechanism is that scarcity taxes the mind. Anandi Mani and colleagues showed in Science in 2013 that financial worry measurably reduces cognitive performance. Taken together, the evidence supports a measured conclusion. Cash does not produce a utopia, but when support is stable, people use freed time productively, and the one universal, permanent design we can observe did not shrink aggregate employment.

THE ECONOMIC MODEL

The model starts from 2025 baseline figures. U.S. nominal GDP totaled $30.76 trillion for a population of roughly 342 million, which leaves an estimated 265 million adults. The 2026 poverty guideline for a single person is $15,960 a year. Federal revenue reached $5.2 trillion. The average private payroll workweek was 34.3 hours.

THE COST OF A NATIONAL FLOOR

Paying this benefit to every adult, plus a half benefit to every child, has a gross cost of $4.84 trillion. That figure misleads, because most recipients also pay the taxes that fund the program. The number that matters is the net cost: what is actually transferred to people who come out ahead.

With a $15,960 benefit and a 50 percent clawback applied to market income, the break even point is $31,920. A person earning nothing keeps the full benefit. Someone earning $10,000 has half of that clawed back, leaving a net gain of $10,960 and a final income of $20,960. Above $31,920, people become net contributors. Every extra dollar earned still raises final income, so work always pays.

Using Census microdata, economist Karl Widerquist estimated the net cost of a poverty line basic income at 2.67 percent of GDP in 2024, down from 9.35 percent in 1967. At 2025 GDP that is about $0.82 trillion a year, roughly half the current cost of Social Security.

PROJECTING GROWTH

To test whether AI growth could pay for the floor, the model projects real GDP ten years out under four scenarios. The baseline assumes 2 percent annual real growth. A cautious scenario, following Acemoglu, adds roughly 1 percent to the level of GDP after a decade, reaching $37.87 trillion with $0.37 trillion attributable to AI. An optimistic mainstream scenario, following Goldman Sachs, grows at 3.5 percent a year and reaches $43.39 trillion, with $5.89 trillion attributable to AI.

A transformative scenario assumes 7 percent annual growth, reaching $60.51 trillion with $23.01 trillion attributable to AI. This is a stated hypothetical, not a forecast. Growth like that would require AI to move well beyond isolated tasks, for example through multi agent systems in which autonomous agents coordinate to run whole interconnected workflows. Whether that happens, and how fast, is exactly what economists disagree about.

CAPTURING THE GAIN

To fund the floor, some share of the new AI output has to return to the public. Call that share the capture rate. It equals the cost of the floor divided by the output attributable to AI.

In the cautious scenario, AI growth alone cannot pay for the floor. It would require a 270 percent capture rate, which is impossible. If AI turns out to be a modest productivity tool, a floor would have to be funded like any other major program, through ordinary tax choices.

In the optimistic mainstream scenario, capturing 19.7 percent of the AI gain funds a poverty line floor. For comparison, federal revenue already takes roughly 17 percent of GDP. In the transformative scenario, capturing just 7.0 percent is enough.

The stress test tells the most important part of the story. Suppose displacement is so severe that almost nobody earns above break even, pushing the cost of the floor toward its full gross amount of about $5.09 trillion. In the transformative scenario, capturing 22.1 percent of the AI gain still covers it. The more work AI takes over, the more affordable it becomes to guarantee that people are fine without that work. The real danger is not that we can’t afford the floor. It is that the gain is captured privately and no floor gets built.

THE TIME DIVIDEND

If some productivity gains were taken as time rather than consumption, the workweek could shrink substantially. Assuming output scales with hours, producing baseline output in the optimistic mainstream scenario would let the average workweek fall from 34.3 hours to about 29.6. In the transformative scenario, it could drop to about 21.3 hours, giving people 13 extra hours a week. That is close to the 11 hours Irish artists actually redirected into their work, and within reach of Keynes’s fifteen hour week.

A PRACTICAL FIRST STEP

A national floor is a large political lift. A sectoral one is not. Scaled to the U.S. population, the Irish arts program works out to roughly 126,000 artists receiving about $19,760 a year, for around $2.5 billion, which is less than 0.01 percent of GDP. It would be a large scale test of the future described above, at a cost the federal budget would barely register.

CARE IS THE BIGGEST JOB ON EARTH

The automation conversation fixates on paid jobs, but the largest workforce on Earth is unpaid. The International Labour Organization found that 16.4 billion hours a day go to unpaid care work, the equivalent of 2 billion people working full time, worth about $11 trillion at minimum wage. Women perform 76.2 percent of those hours. This is the work that holds societies together, and it is the work machines are least suited to replace. An income floor quietly compensates it.

MECHANISMS FOR WEALTH CAPTURE

Capturing this wealth requires concrete mechanisms. A sovereign wealth fund holding a public stake in AI returns would follow the Alaska model, with automated production taking the place of oil. Taxes on compute, or on the excess profits of frontier AI firms, would target the specific income stream that is growing. The Windfall Clause, proposed by researchers at Oxford’s Centre for the Governance of AI in 2020, would have AI companies commit in advance to share profits above a very large threshold of world GDP. Folding existing income security programs, which cost $397 billion in fiscal 2025, into a simpler universal floor would offset part of the net cost without touching Social Security.

One further idea is worth exploring: adapting the attribution modeling used in marketing analytics to estimate how much value automated systems contribute across digital supply chains. If it could be made reliable, it would allow capture to track AI output more precisely than blunt corporate taxation. No government uses anything like it today, so it belongs on the research agenda rather than in the budget.

ADDRESSING THE CRITIQUES

Critics rightly argue that cash underperforms its utopian promises. The OpenResearch trial found that recipients’ well-being rose in the first year and then returned to the control group’s level. Advocates reply that $1,000 a month in a high cost economy is a cushion, not freedom from bills, and that a temporary trial cannot reproduce the security of a permanent floor. The evidence on large, permanent, poverty line designs is still thin, which is exactly why the national arts pilot matters.

Others point out that work supplies meaning and identity. An income floor has to be paired with institutions where purpose lives outside a paycheck: arts funding, care cooperatives, education and community spaces.

There is also the plain possibility that transformative growth never arrives. The model shows that under cautious estimates, AI cannot pay for this on its own. Even if the wealth does arrive, raising demand for scarce goods like housing lets landlords capture part of the gain, so any cash transfer needs aggressive housing supply reform alongside it. Finally, the IMF warned that in most scenarios AI will likely worsen overall inequality. That is not a reason to abandon the goal. It is the reason deliberate, structural policy is required.

PROJECTED OUTCOMES

Bringing the model and the evidence together gives a reasonable projection for the United States a decade after adopting a poverty line floor, under the optimistic or transformative scenarios. Official poverty would effectively end for those enrolled. Aggregate employment would likely hold roughly steady, with more part time work. The average workweek could fall by 5 to 13 hours. Creative output would likely rise, reported stress would likely fall, and unpaid care would gain real hours. These are projections built on stated assumptions, not guarantees. Change the growth rate, the clawback or the capture share, and the results change with them.

THE CHOICE IN FRONT OF US

We are close to having machines that can carry much of the labor that once consumed human lives. What we do with that is a moral decision wearing an economic costume. We can let the gains pool at the top and tell displaced workers to compete harder for what is left. Or we can decide that a person’s worth was never their hourly rate, and that the point of all this capability is to return the hours people spent merely surviving.

The math says that in any future where AI does most of the work, a floor under everyone is affordable. The evidence says people given security make more art, feel less strain, and show up better for their children. The song you never finished is still there. So is the parent you meant to visit. A world where machines do the labor could be a world where we finally have time for both, but only if we build it on purpose.

METHODS AND LIMITATIONS

All figures are in constant 2025 dollars over a ten year horizon. The adult share of the population (77.5 percent) is an approximation. The growth scenarios are illustrative compound growth projections. Scenario A approximates Acemoglu’s cautious estimate as a 1 percent level effect on GDP. Scenario B applies Goldman Sachs’s projected 1.5 point annual boost. Scenario C is a hypothetical transformative case. Net cost uses Widerquist’s 2024 estimate of 2.67 percent of GDP and assumes that share holds as the economy grows. The stress case uses gross cost, grown 0.5 percent a year for population, as an upper bound. The time dividend assumes output scales proportionally with hours. The labor response relies on the cited trials and does not model feedback on tax revenue, inflation or regional cost of living differences, each of which a full policy analysis would need.

About the Author

Press enter or click to view image in full size

Brian Curry is a Kansas City-based machine learning and AI strategist, marketing economist, and applied researcher. He leads Vector1, an independent research studio working on cognitive AI, causal inference, marketing economics, and knowledge engineering, with a focus on simulation-driven intelligence systems.

Over more than twenty years, Brian has led data science and analytics initiatives at enterprises ranging from $4B to $125B in revenue, including Koch Industries, Tractor Supply, Vail Resorts, Hallmark, Garmin, and AT&T. His work applies economics and causal inference to connect marketing investment to structural outcomes, operational decisions to economic value, and AI adoption to durable competitive advantage.

He is also the founder of KC AI Lab, a 1,500-member AI and ML education community in the Midwest.

Research and Open-Source Projects

The modeling approach in this paper draws on PyCausalSim and Papilon.

Causal Inference and Economics

• PyCausalSim: Simulation-driven causal discovery and evaluation, using counterfactual scenarios to isolate structural drivers.
• Papilon: Complex-system simulation and optimization, with applications in media mix modeling and adaptive architectures.
• Daedalus: Causal analytics for money-flow mapping and macroeconomic tracking.
• The Autonomy Economy: A macroeconomic framework for the combined effects of rapid productivity growth and structural price deflation in an AI-automated economy.

Forecasting and Applied Modeling

• Regimecast: Time-series regime detection, structural breaks, and economic forecasting.
• CarryNet: Financial modeling for commodity carry trade analysis.
• Foodmesh-ml: Machine learning, optimization, and simulation for distributed food systems.

Cognitive AI and Semantic Architecture

• Memory-Node Encapsulation (MNE): A cognitive AI architecture and data structure for artificial episodic memory (NeoCortex-M).
• MeaningFlow: Semantic content modeling, embeddings, clustering, and coverage gap analysis.

Contact: brian@vector1.ai | GitHub @Bodhi8 | Substack

The text of this article may be shared with attribution under Creative Commons Attribution 4.0 International (CC BY 4.0). Code and models referenced here are available separately to paid subscribers.

REFERENCES

Georgieva, K. AI Will Transform the Global Economy. Let’s Make Sure It Benefits Humanity. IMF Blog, January 14, 2024. Coverage: https://www.arabnews.pk/node/2442401

Goldman Sachs. The Potentially Large Effects of Artificial Intelligence on Economic Growth. March 2023. Summary: https://economic-policy.org/79th-economic-policy-panel/macroecon-of-ai/

Acemoglu, D. The Simple Macroeconomics of AI. NBER Working Paper 32487, 2024. https://www.nber.org/papers/w32487

Wilde, O. The Soul of Man Under Socialism. 1891.

Keynes, J. M. Economic Possibilities for Our Grandchildren. 1930, in Essays in Persuasion.

Acemoglu, D. and Johnson, S. Power and Progress. PublicAffairs, 2023.

The Irish Times. Fairness of basic income for arts scheme under scrutiny as new funding round opens. April 15, 2026. https://www.irishtimes.com/culture/art/2026/04/15/fairness-of-basic-income-for-arts-scheme-under-scrutiny-as-new-funding-round-opens/

RGMII. Thousands more artists join Ireland’s basic income plan. June 2026. https://rgmii.org/blog/thousands-more-artists-join-irelands-basic-income-plan/

Datawrapper. Investment over cuts: Ireland’s plan to save the arts. April 2, 2026. https://www.datawrapper.de/blog/basic-income-arts-ireland

Ministry of Social Affairs and Health Finland and Kela. Results of the basic income experiment. May 6, 2020. https://stm.fi/en/-/perustulokokeilun-tulokset-tyollisyysvaikutukset-vahaisia-toimeentulo-ja-psyykkinen-terveys-koettiin-paremmaksi

Kela. Basic income experiment. https://tietotarjotin.fi/en/information-package/158070/basic-income-experiment

Jones, D. and Marinescu, I. The Labor Market Impacts of Universal and Permanent Cash Transfers: Evidence from the Alaska Permanent Fund. American Economic Journal: Economic Policy 14(2), 2022, pages 315 to 340. https://home.uchicago.edu/j1s/Jones_Alaska_2022.pdf

Vivalt, E., et al. The Employment Effects of a Guaranteed Income. NBER Working Paper 32719, 2024. https://www.nber.org/papers/w32719

OpenResearch. Parenting findings. Business Wire, July 21, 2025. https://www.businesswire.com/news/home/20250721662125/en/

Mani, A., Mullainathan, S., Shafir, E. and Zhao, J. Poverty Impedes Cognitive Function. Science 341, 2013, pages 976 to 980. doi:10.1126/science.1238041

Widerquist, K. The Cost of Basic Income: Back of the Envelope Calculations. Basic Income Studies 12(2), 2017. https://ideas.repec.org/a/bpj/bistud/v12y2017i2p13n4.html

Widerquist, K. The Falling Cost of Basic Income in the United States, 1967 to 2024. Summary via BIEN: https://basicincome.org/news/author/karl-widerquist/

International Labour Organization. Care Work and Care Jobs for the Future of Decent Work. 2018. https://www.ilo.org/asia/media-centre/news/WCMS_633284/lang--en/index.htm

U.S. Bureau of Economic Analysis. GDP, 4th Quarter and Year 2025. https://www.bea.gov/news/2026/gdp-second-estimate-4th-quarter-and-year-2025

HHS/ASPE. 2026 Poverty Guidelines. https://aspe.hhs.gov/sites/default/files/documents/b1bfa16b20ae9b89d525bc35de7c1643/detailed-guidelines-2026.pdf

Congressional Budget Office. Mandatory Spending in Fiscal Year 2025. https://www.cbo.gov/system/files/2026-03/61951-mandatory.pdf

Committee for a Responsible Federal Budget. CBO Estimates 2025 Deficit Totaled $1.8 Trillion. https://www.crfb.org/blogs/cbo-estimates-2025-deficit-totaled-18-trillion

O’Keefe, C., et al. The Windfall Clause: Distributing the Benefits of AI for the Common Good. Centre for the Governance of AI, 2020.