The stake.
Start with the mechanism, because the mechanism is the whole argument. For two centuries, most people earned their living by selling their labor. The owners of capital — machines, land, equity — earned theirs by owning the means of production. Most of us were on the labor side of that line. The deal was: you work, you get wages; the wages are how the economy’s output reaches you.
AI is an attack on the labor side of the line specifically. It does not redistribute income from one worker to another; it shifts the source of value from labor to capital — from the people who do the work to the people who own the systems that do it instead. When an agent does the analyst’s task, the value the analyst used to capture as wages does not vanish. It moves to whoever owns the agent.
This is why the standard responses feel insufficient. Retraining assumes there is a labor-side job to retrain into; if the value is moving to capital structurally, retraining is running up a down escalator. And income redistribution — taxing the winners to send checks to the displaced — treats the symptom (lost wages) while leaving the structure (concentrated ownership) untouched. The check arrives after the fact, forever, as a transfer. The recipient never owns anything.
The post-labor argument I want to make is narrower and, I think, more market-friendly than either: if value is moving from labor to capital, the response is to broaden who owns the capital — to give people a stake in the automation rather than a transfer after it. Not redistribute the income that capital throws off; redistribute, or rather pre-distribute, the ownership itself. A citizen who owns a share of the productive economy is on the capital side of the line when the line moves. A citizen who receives a check is permanently on the wrong side of it, dependent on the goodwill of those who own.
The structural argument I want to make: the AI transition is best understood not as a jobs problem but as an ownership problem — value is shifting from labor to capital, and the durable, market-compatible response is broad-based capital ownership (universal basic capital, sovereign wealth funds, employee ownership) rather than after-the-fact income redistribution (UBI), because ownership puts the citizen on the side of the line value is moving toward, while a transfer leaves them dependent on the side it is moving away from. This is the foundational dispatch of a new track — Post-Labor — and it is the thesis the rest of the catalog has been circling: the publisher losing the referral, the consultant losing the engagement, the analyst losing the task are all the same event, value leaving labor for capital, and the question underneath all of them is who owns the capital.
The headline integrative finding: The honest both-sides read has to start with the strongest objection, because it is serious: the premise may be wrong. The labor share of US income has been remarkably stable for seventy years (roughly 57-64%), workers displaced by past technology mostly moved into new work rather than disappearing, and a credible camp argues AI will reallocate labor as every prior wave did, not eliminate it. If that camp is right, the ownership question is premature. But the post-labor case does not require mass unemployment to hold — it requires only that the share of value going to capital rises durably, which is a weaker and better-supported claim, and one that ownership-broadening addresses without betting on the apocalyptic version. The deepest point is that broad-based ownership is a good idea whether or not AI ends work: if AI reallocates labor, ownership cushions the transition; if it displaces labor, ownership replaces the lost wages with property income; either way, the citizen holds an asset rather than a hope. The bet is asymmetric in ownership’s favor — which is the case for treating it as the response regardless of which future arrives.
This essay walks the shift from a labor problem to an ownership problem, why the standard responses are insufficient, the distinction between basic income and basic capital, the mechanisms that already exist, the natural-experiment evidence, the serious objection that the premise is wrong, and the structural reading of ownership as the market-friendly answer.
The stake.
Why the answer to automation
is broad-based ownership,
not a bigger transfer.
from ~50% in the 1970s
vs +54% for the top 1,500 CEOs
measured hit to full-time work
3.7% in 1995 · 3x the bottom half
value added · 1970s → 2022
moves to
capital
the systems that do the work
- An income flow, funded by taxation (robot taxes, compute dividends, data rents)
- Depends on continued taxation and political will
- Ownership stays where it is — the recipient never owns the assets
- Fights the market’s distribution with a counter-distribution
- An owned, compounding stake in the productive economy
- An asset you hold — not dependent on anyone’s discretion
- Pre-distributes ownership — the citizen earns capital income directly
- Uses the market’s own machinery — equity, returns — to spread the gains
The market-friendly response to automation is not to fight the machines or to tax their owners into funding a transfer society. It is to make more people owners of the machines — to give the citizen a stake in the automation rather than a claim on its winners’ goodwill. The window for that is widest before the value finishes moving.Thorsten Meyer · The Stake · Post-Labor 01
By Thorsten Meyer — June 2026
This is the first dispatch in a new track — Post-Labor — the economics of value migrating from labor to capital, and the responses available to a market society that does not want to become a transfer society. It is the throughline beneath the rest of my work: the publisher track, the enterprise track, the governance track are all studies of value leaving labor for capital in specific industries. This track asks the general question those specific ones share — who owns the capital — and argues that the answer determines whether the AI transition concentrates or distributes.
The structural argument I want to make: the political economy of automation is usually framed as a fight between markets and redistribution, and that framing is a trap — because the market-friendly response to automation is not “do nothing” and it is not “tax and transfer,” it is “broaden ownership,” which is more market-compatible than redistribution and more humane than laissez-faire. Ownership-broadening uses the market’s own logic — property rights, equity, compounding returns — to spread the gains, rather than fighting the market with transfers or surrendering to it with inaction. It is the rare position that a serious free-marketeer and a serious egalitarian could both reach from their own premises.
The headline integrative finding: Universal basic capital is not utopian and it is not new — sovereign wealth funds, employee stock ownership plans, the Alaska Permanent Fund, and the German co-determination and employee-ownership traditions are all working instances of broad-based capital ownership that already exist and already pay. The question is not whether broad-based ownership can work; it demonstrably does. The question is whether a society facing a labor-to-capital shift will choose to scale it deliberately — before the shift concentrates ownership so far that broadening it later requires fighting entrenched interests rather than designing ahead of them. The window is the point: ownership is easiest to broaden before the value has finished moving, not after.
This essay walks the labor-to-capital shift (Section I), the insufficiency of the standard responses (Section II), basic income versus basic capital (Section III), the existing mechanisms (Section IV), the natural-experiment evidence (Section V), the serious objection (Section VI), and the structural reading (Section VII).
broad-based capital ownership platforms
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
I · The shift · from a jobs problem to an ownership problem
The reframing crystallization. The single most important move is to stop asking “will AI take the jobs” and start asking “where does the value go.” The first question is empirical and contested; the second is structural and clarifying.
The two sides of the line
Labor income versus capital income: national income divides, roughly, into what goes to labor (wages, salaries) and what goes to capital (profits, rents, returns to equity). For most of the postwar period, labor took the larger share. The owners of human labor and the owners of non-human productive assets split the output — and most people lived on the labor side.
The secular drift: since the 1980s, labor’s share has drifted down. The raw US series shows a decline from roughly 50% of gross value added in the 1970s toward 44% by 2022; careful allocation of self-employment income softens that to a 4-5 point decline, but the direction is not in dispute. Globally, the Kelso Institute Europe framing is blunt: since the industrial revolution, the owners of productive capital have received a growing share of market income, with an ever-smaller proportion going to the owners of labor.
What AI does to the line
AI is automation capital that substitutes for labor: in the economic models (Acemoglu and Restrepo and successors), automation capital and labor are substitutes — the robot or the agent does the task the worker did — while traditional capital and labor are complements. AI is the substitute kind: it does not make workers more productive so much as it does the work instead, which moves the task’s value from the wage line to the capital line.
The displacement-reallocation question: whether displaced workers find new labor-side work (reallocation) or not (displacement) is the contested empirical question. But note: even full reallocation does not stop the share shift. If automation keeps moving tasks to the capital side faster than new labor-side tasks appear, the capital share rises even with full employment. The ownership question survives even the optimistic labor-market scenario.
The shift observation
The AI transition is most clearly understood not as a jobs problem but as a question of where value accrues — and AI is the kind of capital that substitutes for labor, moving task value from the wage line to the capital line. The jobs question (displacement vs reallocation) is real but contested; the share question (is value moving to capital) is better-supported and more structural. Reframing from “will it take the jobs” to “where does the value go” is the move that makes the ownership question visible — because if value is moving to capital, then who owns the capital is the question that determines everything else.
An Introduction to ESOPs, 19th Edition: How an employee stock ownership plan (ESOP) can benefit your company, its owners, and its employees
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
II · The insufficiency · why retraining and redistribution fall short
The inadequacy crystallization. The two standard responses — retrain the workers, tax-and-transfer to the displaced — are not wrong so much as aimed at the wrong layer. They address the wage while leaving the ownership structure intact.
Retraining runs up a down escalator
The assumption retraining makes: retraining assumes there is a labor-side destination — a new job, requiring new skills, that the displaced worker can move into. That assumption holds in a reallocation world. But if value is structurally moving to capital, retraining moves workers between shrinking labor-side roles while the capital side captures the growth — running up a down escalator. Retraining is necessary and humane; it is not a structural answer to a structural shift.
Redistribution treats the symptom
Tax-and-transfer leaves ownership untouched: redistribution — tax the AI winners, send income to the displaced — addresses the lost wage directly. But it does so as a perpetual transfer: the recipient receives income, forever, and never comes to own any of the productive assets generating it. The structure that concentrated the value is left intact; the transfer just routes some of capital’s income to labor’s former earners, at the discretion of whoever controls the tax-and-transfer system.
The dependency problem: a transfer society is one in which a growing share of the population depends on payments from a shrinking class of owners, mediated by the state. That is politically fragile (the transfer can be cut), psychologically corrosive (dependence rather than ownership), and structurally static (it never changes who owns what). Redistribution can keep the displaced fed; it cannot make them owners — and the difference between being fed and owning is the difference between the two sides of the line.
Why the layer matters
Wage policy versus ownership policy: retraining and redistribution are wage-layer responses — they try to restore or replace labor income. The shift is at the ownership layer — value moving to capital. A wage-layer response to an ownership-layer shift is a category mismatch: it can soften the blow without changing the structure that delivers it. To meet a shift in who owns the value, you need a response that changes who owns the value.
The insufficiency observation
Retraining and redistribution are wage-layer responses to an ownership-layer shift: retraining assumes a labor-side destination that the share-shift erodes, and redistribution routes capital’s income to labor’s former earners as a perpetual, discretionary transfer that never makes them owners. Both are humane and both have a role; neither addresses the structural fact that value is moving to capital. The category mismatch is the point: to respond to a shift in ownership, you have to change ownership — which is what neither standard response does, and what the post-labor answer is built to do.
320 Things to Know About Sovereign Wealth Funds
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
III · Basic income versus basic capital · the distinction that matters
The core-distinction crystallization. The post-labor position is often confused with universal basic income. It is closer to its opposite. The difference between distributing income and distributing capital is the difference between a transfer and a stake.
The two models
Universal basic income (UBI): every citizen receives a regular cash payment, funded by taxation (including proposed AI-specific taxes — robot taxes, compute dividends, data rents, platform levies). The citizen receives income; the income is a flow; the flow is funded by taxing the owners. UBI redistributes income after the fact and leaves ownership where it is.
Universal basic capital (UBC): every citizen receives an equal stake in the productive economy — a share of a national fund, an equity endowment, an ownership position — that pays dividends from economic growth. One concrete proposal: corporations contribute a percentage of their stakeholder liabilities (stock, bonds, compensation) to a national superfund, adjusted by their domestic sales share; citizens hold equal shares and earn dividends. UBC pre-distributes ownership and lets the citizen earn capital income directly.
Why the distinction is structural, not semantic
Income is a flow; capital is a stock: UBI gives you a flow that depends on continued taxation and political will. UBC gives you a stock — an asset you own, that compounds, that you are not dependent on anyone’s discretion to keep receiving. The UBI recipient is a perpetual claimant on capital’s income; the UBC holder is a part-owner of capital. When value moves to capital, the claimant is still on the labor side asking for a share; the owner is on the capital side receiving one.
The market-compatibility difference: UBI fights the market’s distribution with a counter-distribution (tax and transfer). UBC uses the market’s own machinery — equity, ownership, compounding returns — to spread the gains. UBC is the more market-friendly instrument precisely because it does not require an adversarial transfer; it makes the citizen a shareholder in the thing that is winning, rather than a tax-funded dependent of it.
Why this is the post-labor position
Ownership puts the citizen on the right side of the line: the entire argument reduces to this. If value is moving from labor to capital, the response that works is the one that moves the citizen from the labor side to the capital side — which is ownership, not income. UBC is that move. UBI is a way to live more comfortably on the losing side; UBC is a way to be on the winning one.
The basic-capital observation
Universal basic income redistributes income and leaves ownership intact; universal basic capital pre-distributes ownership and lets the citizen earn capital income directly — and the difference is structural, because income is a discretionary flow while capital is an owned, compounding stock. UBC is the more market-compatible instrument because it spreads the gains through the market’s own machinery rather than against it. This is the post-labor position in one line: when value moves to capital, give people capital — not a check that depends on the goodwill of those who own, but a stake that makes them owners.
By Man Shall His Blood Be Shed: A Catholic Defense of Capital Punishment
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
IV · The existing mechanisms · this is not utopian
The proof-of-concept crystallization. The strongest evidence that broad-based capital ownership can work is that it already does, at scale, in multiple forms. UBC is not a thought experiment; it is an existing category waiting to be scaled.
Sovereign wealth funds
The national-ownership model: sovereign wealth funds — Norway’s $1.7T Government Pension Fund Global the canonical case — are national vehicles that own diversified equity stakes and manage the returns for citizens. In the AI context, SWFs have been explicitly proposed as tools to acquire equity in AI companies and distribute AI-derived returns through universal basic dividends or stabilization funds. A sovereign fund that owns a slice of the AI economy and pays its returns to citizens is universal basic capital at the national scale — and the instrument already exists.
Employee ownership
The firm-level model: employee stock ownership plans (ESOPs), employee ownership trusts, and the German co-determination and employee-ownership traditions (the Kelso Institute Europe’s work out of Berlin and Frankfurt-Oder) put productive capital in workers’ hands at the firm level. Employee ownership is broad-based capital ownership built one company at a time — and Europe, with its co-determination tradition, has institutional foundations the US lacks.
The citizen’s-dividend model
The endowment-at-birth model: baby bonds (a capital endowment for every child, growing until adulthood) and citizen’s-dividend proposals distribute an ownership stake directly to individuals. These are UBC delivered as a personal endowment rather than a national fund — the same principle, a different delivery mechanism.
The existing-mechanisms observation
Broad-based capital ownership already exists and already pays — sovereign wealth funds (Norway, Alaska), employee ownership (ESOPs, the German tradition), and citizen-endowment proposals (baby bonds) are all working or proposed instances of UBC. The question is therefore not whether it can work — it demonstrably does — but whether a society facing the labor-to-capital shift will scale it deliberately. The instruments are on the shelf. The post-labor argument is that the AI transition is the reason to take them down and scale them, before the value finishes moving.
V · The evidence · what the natural experiments show
The empirical crystallization. Two long-running programs let us test the central worry about distributing capital returns to citizens — that it will make people stop working. The evidence says it does not.
The Alaska Permanent Fund
The closest US natural experiment: Alaska’s Permanent Fund Dividend — financed by a sovereign wealth fund fed by oil revenues — has paid residents an annual dividend (around $1,600 in recent years) for over four decades. A leading study finds no overall effect on full-time employment, though consumer-facing sectors expanded (more local spending supported more local jobs). A capital dividend, paid to everyone, did not reduce work — it is the single best evidence that broad-based capital income is compatible with a working economy.
The Finland basic-income trial
The cash-transfer comparison: Finland’s 2017-18 basic-income trial improved recipients’ well-being and mental health but produced little change in employment. The reading: cash transfers deliver psychological benefit without being a jobs-destroyer — but also without being a jobs policy. This validates the humane case for transfers while confirming they are not, by themselves, structural.
What the evidence supports and does not
It supports the feasibility, not the apocalypse: the natural experiments show that distributing capital returns (Alaska) or cash (Finland) to citizens does not collapse the work ethic — answering the central objection to UBC. They do not prove AI will cause mass displacement; they were not designed to. The evidence is about the response’s feasibility, not the problem’s severity — it tells us UBC would not break the economy, not that the economy needs it yet.
The evidence observation
The natural experiments — Alaska’s four-decade capital dividend and Finland’s basic-income trial — show that distributing capital returns or cash to citizens does not reduce work, answering the central feasibility objection to broad-based ownership. Alaska is the strongest case: a sovereign-fund dividend, paid to everyone, for forty years, with no measured hit to full-time employment. The evidence establishes that UBC is feasible — that you can give people capital income without breaking the incentive to work — which clears the practical objection and leaves only the question of whether the AI shift makes it necessary.
VI · The serious objection · the premise might be wrong
The strongest-counter crystallization. Intellectual honesty requires leading the skeptical section with the best version of the case against, because it is genuinely strong: the labor-to-capital shift may not happen the way the thesis assumes, and if it does not, the urgency evaporates.
The stability case
Labor’s share has been remarkably stable: the strongest empirical counter, made forcefully in recent analysis (ITIF, May 2026), is that the US labor share of income has fluctuated within a relatively narrow band (roughly 57-64%) for seventy years despite enormous technological change — industrial machinery, computers, the internet — and there is no clear sign AI will break the pattern. If labor’s share has survived every prior wave, the assumption that AI will durably erode it is an assumption, not a fact.
The reallocation case
Workers move, they do not disappear: the historical pattern is reallocation — displaced workers transition to new occupations and industries, including ones the new technology creates. On this view, AI reshapes work rather than eliminating it, the tax base does not erode, and policymakers should not restructure the economy in anticipation of a displacement that prior evidence suggests will not materialize. The optimistic case is not naive; it is the base rate of two centuries of technological change.
Why the objection does not sink the thesis
The thesis needs the share-shift, not the apocalypse: here is the key. The ownership argument does not require mass unemployment. It requires only that capital’s share rises durably — a weaker claim than “AI ends work,” and one compatible with full employment (Section I). Even the optimistic reallocation scenario can feature a rising capital share if automation moves tasks to capital faster than new labor-side tasks appear. The serious objection refutes the apocalyptic version of the thesis, not the structural one.
The asymmetry of the bet: and even granting full uncertainty, the response is robust to being wrong. If AI reallocates labor (the optimists are right), broad-based ownership cushions the transition and spreads the productivity gains — a good outcome. If AI displaces labor (the pessimists are right), ownership replaces lost wages with property income — a necessary outcome. Broadening ownership is beneficial across both futures; doing nothing is fine only in the optimistic one. The bet is asymmetric in ownership’s favor — which is the argument for acting on it without needing to resolve the empirical dispute first.
The objection observation
The serious objection — that labor’s share has been stable for seventy years and workers reallocate rather than disappear — is strong, well-evidenced, and refutes the apocalyptic version of the thesis, but not the structural one, because the ownership argument needs only a durable rise in capital’s share, which is compatible with full employment. And the response is robust to the uncertainty: broad-based ownership helps whether AI reallocates or displaces labor, while inaction is safe only if the optimists are right. The honest position concedes the premise is contested and argues that ownership-broadening is the move that does not require winning the argument — it is the no-regrets policy across both futures.
What this is not
It is not a prediction of mass unemployment. The thesis is agnostic on displacement vs reallocation; it rests on the share-shift, which is weaker and better-supported.
It is not anti-market. UBC uses markets — equity, ownership, returns — to spread gains. It is more market-compatible than redistribution, not less.
It is not a claim that UBC is easy. Governance risks (elite capture, opacity, political misuse), funding design, and the political economy of standing up a national fund are real and hard. The claim is that it is the right target, not that it is a simple one.
The synthesis observation
The AI transition is best understood as an ownership problem, not a jobs problem — value is shifting from labor to capital, and the durable, market-compatible response is broad-based capital ownership rather than after-the-fact income redistribution, because ownership puts the citizen on the side of the line that value is moving toward. UBC is not utopian (sovereign funds, employee ownership, and citizen dividends already work), not a jobs-destroyer (Alaska’s forty-year evidence), and not anti-market (it uses the market’s own machinery) — and it is robust to the genuine uncertainty about whether AI displaces or merely reallocates labor, because it helps in both futures.
There is no single answer. Anyone offering one is selling something. What the post-labor frame clarifies is that the real choice is not markets versus redistribution — it is concentrated ownership versus broad ownership, and that choice is being made now, by default, in favor of concentration, every time value moves from a wage to a return and no one owns the return but the few who already owned the capital. The market-friendly response to automation is not to fight the machines or to tax their owners into funding a transfer society. It is to make more people owners of the machines — to give the citizen a stake in the automation rather than a claim on its winners’ goodwill. The window for that is widest before the value finishes moving, which is to say: now.
That is the structural editorial question the stake sits on top of. It is value moving from labor to capital, which is an ownership question wearing a jobs question’s clothes. It is the difference between a transfer that leaves you dependent and a stake that makes you an owner. And it is a no-regrets response to a genuinely uncertain future — good if AI reallocates labor, necessary if it displaces it. And it is the foundation of everything else in this track — because the publisher losing the referral, the consultant losing the engagement, and the analyst losing the task are all the same event, and the question underneath all of them is the same: when the value moves to capital, who owns the capital?
About the Author
Thorsten Meyer is a Munich-based futurist, post-labor economist, and recipient of OpenAI’s 10 Billion Token Award. He spent two decades managing €1B+ portfolios in enterprise ICT before deciding that writing about the transition was more useful than managing quarterly slides through it. He runs StrongMocha News Group, a network of more than 450 niche WordPress magazines built on the DojoClaw editorial engine. More at ThorstenMeyerAI.com.
Related Reading · the Post-Labor track
This dispatch
- This piece · The stake · the foundational Post-Labor dispatch — why the AI transition is an ownership problem rather than a jobs problem, and why broad-based capital ownership is the market-friendly response · synthesis-deep dominant, alternative-sage and structural-slate balance
The track · opening
- The stake · Post-Labor 01 · the ownership question — value moving from labor to capital, and the case for broad-based ownership over redistribution
- Forthcoming · The labor share · Post-Labor 02 · the empirical floor — what the data actually shows about labor share, displacement, and reallocation · labor-rose register
- Forthcoming · The policy menu · Post-Labor 03 · the response options laid side by side — UBI vs UBC vs data dividends vs do-nothing, evenhandedly · synthesis-deep register
Adjacent tracks · the same event in specific industries
- The referral · Post-Wire 03 · the publisher losing the wage of the click — value moving from the content producer to the platform that owns the answer
- The pyramid cracks · Enterprise Reorg 02 · the consultant losing the engagement — value moving from billable labor to the model that owns the analysis
- The deployment · Enterprise Reorg 03 · the labs capturing the services dollar — value concentrating in the owners of the systems that do the work
Sources
The labor-to-capital shift
- Kelso Institute Europe (Lowitzsch & Magalhães) · Automation, AI and capital concentration — “since the industrial revolution, productive capital has assumed an ever more dominant role compared to labour; the owners of non-human productive assets receive a growing share of market-sourced income, with an ever-smaller proportion going to the owners of human labour” · the employee-ownership and profit-sharing frame (Berlin / Frankfurt-Oder) · tandfonline.com
- Capital is Eating the World · labor-share charts — US labor share ~50% of gross value added in the 1970s → 44% by 2022 (raw); 4-5 points after proper proprietor-income allocation (Elsby, Hobijn, Şahin 2013); cyclical around the trend · Piketty/Saez on the inequality coincidence · horse.energy
- Acemoglu & Restrepo (via Mazlish meta-analysis) · why labor share declines — “our single index of automation explains 50% of the variation in the change in labor shares and task displacement across industries” · automation capital and labor as substitutes; traditional capital and labor as complements · the superstar-firm/concentration alternative (Autor et al.) · jzmazlish.substack.com
Basic income versus basic capital
- Universal Basic Capital (research proposal) — UBC “democratizes ownership of productive assets rather than redistributing income after the fact” · the superfund mechanism (corporations contribute a % of stakeholder liabilities, adjusted by domestic sales share; citizens hold equal shares, earn dividends) · Gini 0.494 in 2021; McKinsey 50% automation of work activities 2030-2060 · researchgate.net
- Remio · AI-funded UBI financing mechanisms — the proposed funding instruments: automation/robot taxes, corporate AI profit-sharing, data and algorithmic rents, sovereign compute dividends, platform levies, cap-and-dividend for compute · the UBI-as-income-transfer model · remio.ai
The existing mechanisms
- Convergence Analysis · Lead, Own, Share: SWFs for transformative AI — savings funds to distribute universal basic dividends (a citizen share of AI-derived wealth); stabilization funds to smooth labor shocks; explicit public-benefit mandates; governance via legal structure, reporting, independent oversight; “avoid full nationalization unless necessary” · convergenceanalysis.org
- Wikipedia · Sovereign wealth funds and AI — SWFs proposed to manage automation disruption: strategic AI-infrastructure investment, equity stakes in AI companies, distribution of AI-derived returns via universal basic dividends or stabilization funds · en.wikipedia.org
- arXiv · An AI capability threshold for rent-funded UBI — “ensuring government captures a meaningful slice of AI profits lowers the capability threshold without requiring full government ownership, so society can finance a broad transfer within the range of foreseeable progress” · the AI-economics modeling lineage (Zeira 1998; Acemoglu-Restrepo 2018; Erdil-Besiroglu 2023 on explosive growth) · arxiv.org
The natural-experiment evidence
- Newsweek · the UBI trap / Alaska and Finland — Alaska Permanent Fund Dividend (~$1,600/yr, sovereign-fund financed) “has not cratered employment; a leading study finds no overall effect on full-time work, though consumer-facing sectors expanded” · Finland’s 2017-18 trial “improved well-being but produced little change in employment — validating cash’s psychological benefit but not the claim that UBI is a jobs policy” · the “targeted, timed AI dividend that scales with measured exposure to automation” middle ground · newsweek.com
The serious objection
- ITIF (May 2026) · AI is not going to reduce labor’s share or destroy the tax base — the strongest counter: US labor share fluctuated within ~57-64% from the 1950s to 2023 despite enormous technological change; displaced workers reallocate rather than disappear; the tax base erodes only if labor share falls dramatically and persistently, which is “highly unlikely”; policymakers should not restructure on the assumption of declining labor income · itif.org
- Inequality.org / World Inequality Report 2026 — 2019-2025: a 12% decline in real wages worldwide ($2,326) vs a 54% rise in pay for the 1,500 highest-paid CEOs (to $8.4M avg) · the wealthiest 0.001% (≈56,000 people) hold three times the wealth of the bottom half; their share grew from 3.7% (1995) to 6.1% (2025) · inequality.org
Key reference figures crystallized
- The shift: labor share ~50% (1970s) → 44% (2022) raw, 4-5 pts adjusted · automation capital substitutes for labor (Acemoglu-Restrepo) · the share-shift survives even full reallocation
- The insufficiency: retraining assumes a labor-side destination; redistribution is a perpetual, discretionary transfer that never confers ownership · wage-layer responses to an ownership-layer shift
- Basic income vs basic capital: UBI = income flow, taxation-funded, ownership intact · UBC = owned, compounding capital stake, market machinery · income vs stock; claimant vs owner
- The mechanisms: sovereign wealth funds (Norway $1.7T, Alaska) · employee ownership (ESOPs, German co-determination, Kelso Institute Europe) · baby bonds / citizen dividends · the superfund proposal
- The evidence: Alaska PFD ~$1,600/yr, 40 years, no measured hit to full-time work · Finland 2017-18 (well-being up, employment ~flat) · feasibility, not severity
- The serious objection: US labor share stable at 57-64% for 70 years (ITIF) · reallocation not disappearance · BUT the thesis needs only a durable capital-share rise, and ownership is a no-regrets bet across both futures
- The stakes: real wages −12% (2019-25) vs CEO pay +54% · top 0.001% share 3.7% (1995) → 6.1% (2025); 3x the bottom half (World Inequality Report 2026)
