pith. sign in

arxiv: 2606.05715 · v1 · pith:CGWV2XVXnew · submitted 2026-06-04 · 💻 cs.CE

AGI and the Limits of Value Production

Zichen Song This is my paper

Pith reviewed 2026-06-27 23:32 UTC · model grok-4.3

classification 💻 cs.CE
keywords AGIvalue theorysurplus valueprofit ratelabor substitutionorganic composition of capitalpolitical economy
0
0 comments X

The pith

Near-complete AGI labor substitution drives living labor, surplus value, and the profit rate all toward zero.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper constructs a political-economy model in which AGI progressively substitutes living labor with machine systems. It separates the technical ceiling of replacement from the actual share adopted once costs and frictions are considered. Under the premise that AGI only transfers value already created by labor and creates none itself, greater adoption raises the organic composition of capital, shrinks the quantity of living labor when new fields do not offset displacement, and reduces the pool of surplus value. In the limiting case of near-total adoption without compensatory labor creation, living labor approaches zero, surplus value approaches zero, and the social profit rate therefore approaches zero. The result frames full AGI substitution as a boundary condition for value production itself rather than a simple efficiency shift.

Core claim

Under the strict value-theoretic assumption that AGI transfers value but does not itself create new value, deeper AGI adoption raises the organic composition of capital, reduces the quantity of living labor when adoption outpaces the creation of new labor fields, compresses the source of surplus value, and places downward pressure on the social rate of profit. In the limiting case in which actual AGI adoption approaches complete substitution and new labor fields fail to compensate for displaced labor, living labor tends to zero, surplus value tends to zero, and the profit rate tends to zero.

What carries the argument

The distinction between technical substitutability (the feasible replacement ceiling set by AGI capability) and actual adoption (the realized replacement share chosen under cost, profitability, and frictions), applied within a value theory that treats AGI as transferring but not creating value.

If this is right

  • Deeper AGI adoption increases the organic composition of capital.
  • When adoption outpaces new labor-field creation, the quantity of living labor falls.
  • The source of surplus value is compressed.
  • The social rate of profit experiences sustained downward pressure.
  • Near-complete AGI substitution marks a boundary case for continued value production.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • Economies organized around extraction of surplus value from living labor would encounter a structural limit once AGI handles nearly all production.
  • Responses such as new forms of value accounting or non-labor income distribution become necessary if the profit-rate trajectory holds.
  • Sectors with rapid AGI uptake could serve as early indicators of whether the predicted compression of surplus value appears in aggregate data.

Load-bearing premise

AGI transfers value but does not itself create new value.

What would settle it

Empirical observation of a non-declining or rising social profit rate in an economy where AGI has displaced the large majority of labor and no offsetting new labor fields have emerged.

read the original abstract

This paper develops a political-economy model of artificial general intelligence (AGI) as a technology that progressively substitutes living labor with machine-based productive systems. The model studies the transition from the first moment at which AGI becomes economically capable of replacing labor to the later moment at which AGI becomes technically and actually capable of near-complete replacement. The central distinction is between technical substitutability and actual adoption. Technical substitutability is the feasible replacement ceiling implied by the state of AGI capability, whereas actual adoption is the realized replacement share chosen under cost, profitability, and adoption frictions. Under the strict value-theoretic assumption that AGI transfers value but does not itself create new value, deeper AGI adoption raises the organic composition of capital, reduces the quantity of living labor when adoption outpaces the creation of new labor fields, compresses the source of surplus value, and places downward pressure on the social rate of profit. In the limiting case in which actual AGI adoption approaches complete substitution and new labor fields fail to compensate for displaced labor, living labor tends to zero, surplus value tends to zero, and the profit rate tends to zero. The model therefore identifies near-complete AGI substitution not merely as an efficiency transition, but as a boundary case for value production under a strict political-economy theory of value.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

1 major / 2 minor

Summary. The manuscript develops a conceptual political-economy model of AGI progressively substituting living labor. It distinguishes technical substitutability from actual adoption and, under the maintained premise that AGI transfers value but creates none, shows that deeper adoption raises the organic composition of capital and compresses surplus value; in the limiting case of near-complete substitution without offsetting new labor fields, living labor, surplus value, and the profit rate all tend to zero.

Significance. If the value-theoretic premise is granted, the paper supplies a clear boundary-case analysis that frames AGI adoption as a potential limit to value production rather than a mere efficiency gain. The conditional framing and the technical-versus-actual distinction are maintained throughout, giving the result internal coherence as an implication rather than an unconditional prediction.

major comments (1)
  1. [Abstract] Abstract: the limiting-case result (living labor, surplus value, and profit rate all tending to zero) is stated as the direct logical consequence of the strict premise that AGI creates no new value together with the stated adoption and compensation assumptions; the manuscript therefore illustrates rather than independently derives the outcome.
minor comments (2)
  1. The absence of any formal equations, functional relations, or parameter definitions makes it difficult to trace the precise dynamics of organic composition or profit-rate compression even under the maintained assumptions.
  2. The manuscript would benefit from explicit engagement with prior formal treatments of the falling rate of profit (e.g., Okishio, Duménil & Lévy) to clarify where the AGI-specific mechanism departs from or reproduces existing results.

Simulated Author's Rebuttal

1 responses · 0 unresolved

We thank the referee for the positive assessment, the clear summary of the manuscript's contribution, and the recommendation for minor revision. We address the major comment below.

read point-by-point responses

  1. Referee: [Abstract] Abstract: the limiting-case result (living labor, surplus value, and profit rate all tending to zero) is stated as the direct logical consequence of the strict premise that AGI creates no new value together with the stated adoption and compensation assumptions; the manuscript therefore illustrates rather than independently derives the outcome.

    Authors: We agree that the limiting-case result follows directly as a logical consequence of the maintained value-theoretic premise (AGI transfers value but creates none) together with the adoption and compensation assumptions. The manuscript's contribution is to embed this implication within the technical-versus-actual substitutability distinction and to trace its effects on the organic composition of capital and the profit rate. We will revise the abstract to state more explicitly that the zero-tendency result is presented as a boundary-case implication under the strict premises rather than an independent derivation. revision: yes

Circularity Check

1 steps flagged

Limiting-case zero-profit result is direct entailment of the no-new-value premise

specific steps
  1. self definitional [Abstract (limiting case paragraph)]
    "Under the strict value-theoretic assumption that AGI transfers value but does not itself create new value, deeper AGI adoption raises the organic composition of capital, reduces the quantity of living labor when adoption outpaces the creation of new labor fields, compresses the source of surplus value, and places downward pressure on the social rate of profit. In the limiting case in which actual AGI adoption approaches complete substitution and new labor fields fail to compensate for displaced labor, living labor tends to zero, surplus value tends to zero, and the profit rate tends to zero."

    The zero limits are entailed by construction once the premise (AGI creates no new value) is combined with the limiting adoption condition; surplus value is defined in the theory as arising solely from living labor, so the outcome restates the premise in the boundary case rather than deriving a new result.

full rationale

The manuscript's central boundary result (living labor, surplus value, and profit rate all tending to zero) is presented as following from the maintained premise that AGI transfers but creates no new value, under complete substitution with no offsetting labor fields. This reduces the claimed 'prediction' to a restatement of the input assumption via the definitions of the value theory itself, with no independent derivation or falsifiable step shown.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

The model rests entirely on one domain assumption drawn from classical value theory; no free parameters or new entities are introduced in the abstract.

axioms (1)
  • domain assumption AGI transfers value but does not itself create new value
    Explicitly identified in the abstract as the strict value-theoretic assumption on which the entire analysis depends.

pith-pipeline@v0.9.1-grok · 5743 in / 1215 out tokens · 19074 ms · 2026-06-27T23:32:00.898208+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

42 extracted references · 5 canonical work pages

  1. [1]

    Economic Policy 40(121): 13--58

    Acemoglu D (2025) The simple macroeconomics of AI. Economic Policy 40(121): 13--58

    2025

  2. [2]

    American Economic Review 108(6): 1488--1542

    Acemoglu D and Restrepo P (2018) The race between man and machine: Implications of technology for growth, factor shares, and employment. American Economic Review 108(6): 1488--1542

    2018

  3. [3]

    Journal of Economic Perspectives 33(2): 3--30

    Acemoglu D and Restrepo P (2019) Automation and new tasks: How technology displaces and reinstates labor. Journal of Economic Perspectives 33(2): 3--30

    2019

  4. [4]

    Cambridge Journal of Regions, Economy and Society 13(1): 25--35

    Acemoglu D and Restrepo P (2020) The wrong kind of AI? Artificial intelligence and the future of labour demand. Cambridge Journal of Regions, Economy and Society 13(1): 25--35

    2020

  5. [5]

    Quarterly Journal of Economics 141(2): 1521--1579

    Acemoglu D and Restrepo P (2026) Automation and rent dissipation: Implications for wages, inequality, and productivity. Quarterly Journal of Economics 141(2): 1521--1579

    2026

  6. [6]

    Quarterly Journal of Economics 118(4): 1279--1333

    Autor DH, Levy F and Murnane RJ (2003) The skill content of recent technological change: An empirical exploration. Quarterly Journal of Economics 118(4): 1279--1333

    2003

  7. [7]

    Quarterly Journal of Economics 139(3): 1399--1465

    Autor DH, Chin C, Salomons A and Seegmiller B (2024) New frontiers: The origins and content of new work, 1940--2018. Quarterly Journal of Economics 139(3): 1399--1465

    2024

  8. [8]

    Working paper, Bocconi University and USC Marshall School of Business

    Banchio M and Mantegazza G (2022) Artificial intelligence and spontaneous collusion. Working paper, Bocconi University and USC Marshall School of Business

    2022

  9. [9]

    Working paper, The Wharton School, University of Pennsylvania, December

    Bastani H and Cachon GP (2025) The human-AI contracting paradox. Working paper, The Wharton School, University of Pennsylvania, December. SSRN 5962739

    2025

  10. [10]

    Working Paper 20941, National Bureau of Economic Research, February

    Benzell SG, Kotlikoff LJ, LaGarda G and Sachs JD (2015) Robots are us: Some economics of human replacement. Working Paper 20941, National Bureau of Economic Research, February

    2015

  11. [11]

    Review of Economic Studies 92(1): 69--96

    Beraja M and Zorzi N (2025) Inefficient automation. Review of Economic Studies 92(1): 69--96

    2025

  12. [12]

    Bhaimiya S (2025) AI job cuts: Amazon, Microsoft, and more cite AI for 2025 layoffs. CNBC

    2025

  13. [13]

    Bhaimiya S (2026) Anthropic CEO Dario Amodei warns AI may cause `unusually painful' disruption to jobs. CNBC

    2026

  14. [14]

    Technical report, Cornell Tech, April

    Bondi T and Johnson G (2026) Skill atrophy and AI productivity measurement. Technical report, Cornell Tech, April. SSRN 6169671

    2026

  15. [15]

    Econometrica 82(6): 2167--2196

    Boppart T (2014) Structural change and the Kaldor facts in a growth model with relative price effects and non-Gorman preferences. Econometrica 82(6): 2167--2196

    2014

  16. [16]

    Working paper, Stanford Digital Economy Lab, November

    Brynjolfsson E, Chandar B and Chen R (2025a) Canaries in the coal mine? Six facts about the recent employment effects of artificial intelligence. Working paper, Stanford Digital Economy Lab, November

  17. [17]

    Quarterly Journal of Economics 140(2): 889--942

    Brynjolfsson E, Li D and Raymond L (2025b) Generative AI at work. Quarterly Journal of Economics 140(2): 889--942

  18. [18]

    Budman S (2025) Salesforce CEO confirms 4,000 layoffs ``because I need less heads'' with AI. CNBC

    2025

  19. [19]

    In Essays in Persuasion, pages 358--373

    Keynes JM (1930) Economic possibilities for our grandchildren. In Essays in Persuasion, pages 358--373. Macmillan, London

    1930

  20. [20]

    In Agrawal A, Gans J, and Goldfarb A (eds.) The Economics of Artificial Intelligence: An Agenda

    Korinek A and Stiglitz JE (2019) Artificial intelligence and its implications for income distribution and unemployment. In Agrawal A, Gans J, and Goldfarb A (eds.) The Economics of Artificial Intelligence: An Agenda. University of Chicago Press, Chicago

    2019

  21. [21]

    Information Systems Research 37(1): 156--175

    Li B, Huang N and Shi W (2025) Forced to change? Media exposure of labor issues and firm artificial intelligence investment. Information Systems Research 37(1): 156--175. DOI: 10.1287/isre.2022.0402

    work page doi:10.1287/isre.2022.0402 2025

  22. [22]

    Journal of Political Economy 75(4): 321--334

    Lucas RE (1967) Adjustment costs and the theory of supply. Journal of Political Economy 75(4): 321--334

    1967

  23. [23]

    RAND Journal of Economics 17(1): 48--58

    Mankiw NG and Whinston MD (1986) Free entry and social inefficiency. RAND Journal of Economics 17(1): 48--58

    1986

  24. [24]

    Journal of Political Economy 110(5): 1035--1070

    Matsuyama K (2002) The rise of mass consumption societies. Journal of Political Economy 110(5): 1035--1070

    2002

  25. [25]

    Journal of Political Economy 97(5): 1003--1026

    Murphy KM, Shleifer A and Vishny RW (1989) Industrialization and the big push. Journal of Political Economy 97(5): 1003--1026

    1989

  26. [26]

    Palmer A (2026) Block laying off about 4,000 employees, nearly half of its workforce. CNBC

    2026

  27. [27]

    Journal of Financial Economics 131(2): 484--505

    Chod J, Lyandres E and Yang SA (2019) Trade credit and supplier competition. Journal of Financial Economics 131(2): 484--505. DOI: 10.1016/j.jfineco.2018.08.008

    work page doi:10.1016/j.jfineco.2018.08.008 2019

  28. [28]

    Journal of Law and Economics 3: 1--44

    Coase RH (1960) The problem of social cost. Journal of Law and Economics 3: 1--44

    1960

  29. [29]

    Econometrica 89(1): 311--374

    Comin D, Lashkari D and Mestieri M (2021) Structural change with long-run income and price effects. Econometrica 89(1): 311--374

    2021

  30. [30]

    Quarterly Journal of Economics 103(3): 441--463

    Cooper R and John A (1988) Coordinating coordination failures in Keynesian models. Quarterly Journal of Economics 103(3): 441--463

    1988

  31. [31]

    Review of Economic Studies 73(3): 611--633

    Cooper RW and Haltiwanger JC (2006) On the nature of capital adjustment costs. Review of Economic Studies 73(3): 611--633

    2006

  32. [32]

    Review of Economic Studies 90(5): 2261--2291

    Costinot A and Werning I (2023) Robots, trade, and Luddism: A sufficient statistic approach to optimal technology regulation. Review of Economic Studies 90(5): 2261--2291

    2023

  33. [33]

    Science , volume =

    Eloundou T, Manning S, Mishkin P and Rock D (2024) GPTs are GPTs: Labor market impact potential of LLMs. Science 384(6702): 1306--1308. DOI: 10.1126/science.adj0998

    work page doi:10.1126/science.adj0998 2024

  34. [34]

    Econometrica 84(5): 1645--1704

    Farhi E and Werning I (2016) A theory of macroprudential policies in the presence of nominal rigidities. Econometrica 84(5): 1645--1704. DOI: 10.3982/ECTA11883

    work page doi:10.3982/ecta11883 2016

  35. [35]

    Quarterly Journal of Economics 101(2): 229--264

    Greenwald BC and Stiglitz JE (1986) Externalities in economies with imperfect information and incomplete markets. Quarterly Journal of Economics 101(2): 229--264. DOI: 10.2307/1891114

    work page doi:10.2307/1891114 1986

  36. [36]

    Guerreiro J, Rebelo S and Teles P (2022) Should robots be taxed? Review of Economic Studies 89(1): 279--311

    2022

  37. [37]

    American Economic Review 83(4): 685--709

    Jacobson LS, LaLonde RJ and Sullivan DJ (1993) Earnings losses of displaced workers. American Economic Review 83(4): 685--709

    1993

  38. [38]

    Review of Economic Studies 23(2): 83--100

    Kaldor N (1956) Alternative theories of distribution. Review of Economic Studies 23(2): 83--100

    1956

  39. [39]

    Working paper, Boston University and National University of Singapore, January

    Keppo J, Li Y, Tsoukalas G and Yuan N (2026) On the fragility of AI agent collusion. Working paper, Boston University and National University of Singapore, January. SSRN 5386338

    2026

  40. [40]

    Son H (2025) Goldman Sachs is piloting its first autonomous coder in major AI milestone for Wall Street. CNBC

    2025

  41. [41]

    economic prospects: Secular stagnation, hysteresis, and the zero lower bound

    Summers LH (2014) U.S. economic prospects: Secular stagnation, hysteresis, and the zero lower bound. Business Economics 49(2): 65--73

    2014

  42. [42]

    American Economic Review 105(5): 60--65

    Summers LH (2015) Demand side secular stagnation. American Economic Review 105(5): 60--65

    2015