{"id":"a13f3e85-c769-4210-8af7-ee079a6af10b","arxiv_id":"2502.07050","paper_version":1,"verdict":"REJECT","confidence":"HIGH","novelty_score":2.0,"correctness_risk":"high","formal_verification":"none","parameter_count":3,"one_line_summary":"AGI is claimed to drive human wages to zero and require UBI and AGI taxation, but the paper's own Cobb-Douglas equations do not support the wage collapse.","lead":"This paper argues that AGI will replace human workers, push wages to zero, and force a rewrite of the social contract through UBI and AGI taxation. It uses simple Cobb-Douglas production models to support the claim, but the math contains errors that undermine the derivation.","discovery_kind":"review","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The wage-collapse result is mathematically backward: under the paper's own Eq. (10), w1 diverges as L1->0 for beta1<1, so Eq. (13) does not follow.","rationale":"The reader's weakest assumption identifies exactly the load-bearing flaw. The mathematical core of the paper is the claim that marginal product of labor tends to zero as human labor input tends to zero. But in a Cobb-Douglas production function with a labor exponent below 1, the marginal product of labor tends to infinity as that input tends to zero. This is not a subtle edge case; it is the opposite of the paper's conclusion. The error appears in Models I, II, and III (Eqs. 8, 13, and 20). A second internal inconsistency compounds it: the multiplicative Cobb-Douglas form makes human labor essential, so L_h -> 0 implies output collapses to zero, not that AGI produces more. The paper's 'simplified' production functions that drop the human labor term are not limits of the original functions; they are new functions with a different structure. Section 5's exponential decline of wages is also assumed rather than derived, and the 'AGI labor wage' equation is not connected to the marginal product formulas from Section 4. These are failures of internal consistency, not merely disagreements with an outside consensus. They do not invalidate the policy question, but they do invalidate the paper's claim that its model mathematically demonstrates an inevitable wage collapse. The rejection verdict is therefore unchanged.","tokens_in":6476,"tokens_out":3807,"duration_ms":35840,"concrete_test":"Evaluate the limit in Eq. (10) numerically: set A=1, K=1, L2=1, beta1=0.5, beta2=0.3, and compute w1 for L1 = 1, 0.1, 0.01, and 0.0001. If w1 increases without bound (approximately 0.5, 1.58, 5, and 50) rather than approaching 0, then Eq. (13) is refuted by the model's own formula. Also evaluate Y in Eq. (9) at L1=0: the result is 0, so the 'firms produce more using AGI' scenario cannot be represented by Eq. (9) as written.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Section 3.1's central result, Eq. (13) ('human wages drop to zero'), does not follow from the model the paper itself defines. In Eq. (10), w1 = beta1 A K^alpha L1^(beta1-1) L2^beta2. With the standard Cobb-Douglas convention beta1 in (0,1), the factor L1^(beta1-1) diverges as L1 approaches 0 from above, so w1 tends to infinity, not zero. The same error appears earlier in Section 2.2: for beta < 1, the marginal product of labor beta A(K+K_AGI)^alpha L^(beta-1) also diverges as L -> 0. Thus the paper's core economic mechanism, that AGI substitution drives human wages to zero, is contradicted by its own production function. Moreover, because L1 enters multiplicatively in Eq. (9), taking L1 to 0 with beta1 > 0 drives total output Y to zero; the model cannot represent 'more output with AGI and no human labor' without replacing Eq. (9) with a different production function. The exponential wage decline in Eq. (23) is assumed exogenously, not derived from Model III, and is inconsistent with Model III's marginal-product wage formula. The policy discussion may be worth having, but the formal model as written implies the opposite of the wage-collapse conclusion.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper argues that AGI labor and AGI capital, once integrated into Cobb-Douglas production functions, drive human wages to zero, eliminate human employment, concentrate income among AGI owners, and collapse aggregate demand; it concludes that the social contract must be renegotiated through policies such as UBI, cooperative AGI ownership, and progressive AGI capital taxation. The formal argument proceeds in three models: AGI as capital (Section 2), AGI as labor alongside human labor (Section 3), and AGI as both labor and capital (Section 4). Section 5 introduces a normalized measure of human economic power and postulates an exponential wage decline to describe the transition from decentralized to centralized economic structures. The paper's conclusion is that this transition is mathematically inevitable and that policy intervention is urgent.","tokens_in":6859,"tokens_out":2620,"duration_ms":24564,"significance":"The paper addresses an important and timely question about the economic consequences of AGI and proposes a set of policy responses that are broadly consistent with current debate. If the formal result were valid, the paper would provide a simple demonstration that AGI substitution drives human wages to zero and that aggregate demand collapses. The paper also correctly identifies that the distribution of AGI-owned capital, rather than labor productivity alone, determines the welfare outcome. However, the central theoretical claim is not supported by the model as written: the marginal-product calculations imply the opposite of the wage-collapse conclusion, and the exponential wage decline used later is assumed rather than derived. Because the formal core is mathematically incorrect, the paper cannot substantiate its central claim in its current form.","major_comments":[{"comment":"The statement that w → 0 as L → 0 is the reverse of what Eq. (7) implies. For the standard Cobb-Douglas convention β ∈ (0,1), the term L^(β−1) has a negative exponent, so as L → 0+ the marginal product βA(K+K_AGI)^α L^(β−1) diverges to +∞, not to zero. Consequently Eq. (8) does not follow, and the wage-collapse conclusion of Model I is invalid.","section":"Section 2.2, Eq. (7)"},{"comment":"The same sign error appears in Model II. In Eq. (10), w1 = β1 A K^α L1^(β1−1) L2^β2; with β1 < 1 the factor L1^(β1−1) diverges as L1 → 0+, so w1 tends to infinity, not zero. Moreover, because L1^β1 multiplies the production function in Eq. (9), taking L1 → 0 drives total output Y to zero rather than to a positive AGI-only outcome. Thus Eq. (13), the paper's central claim that human wages drop to zero, is contradicted by the model's own production function.","section":"Section 3.1, Eqs. (10) and (13)"},{"comment":"The claim that increasing AGI capital K_AGI leads to wh → 0 is not supported by Eq. (16). K_AGI appears only through the positive multiplicative factor K_AGI^γ, so for fixed Lh and LAGI the marginal product of human labor rises with K_AGI rather than falling. To obtain wh → 0 one must take Lh → 0, but then Lh^(β1−1) again diverges for β1 < 1. The limiting argument in Eqs. (18)–(20) therefore does not produce the stated wage collapse.","section":"Section 4.1, Eqs. (16) and (20)"},{"comment":"The exponential wage decline wh = w0 e^(−λ L_AGI) is introduced as an assumption, not derived from Model III. This matters because the earlier marginal-product analysis is supposed to justify the wage collapse; when that analysis fails, Eq. (23) imports the conclusion through an exogenous functional form. In addition, Eq. (24) is internally odd: w∞ is described as 'the asymptotic wage level of AGI labor, typically approaching zero', which conflicts with the idea that AGI wages grow in proportion to AGI capital. Without a derivation of Eq. (23) from the production structure, Section 5's power-decline result is an illustration of an assumed path rather than a model-based prediction.","section":"Section 5.1, Eqs. (23) and (24)"}],"minor_comments":[{"comment":"The phrase 'in form of renegotiation the Social Contract' is grammatically incomplete and should be revised to 'in the form of renegotiating the social contract.'","section":"Title and Abstract"},{"comment":"The keyword 'Resource Missalocation' contains a typo; it should be 'Resource Misallocation.'","section":"Introduction, keywords"},{"comment":"There is a typo in the text introducing Eq. (9): 'an extende Cobb Douglas pprodiuction function' should read 'an extended Cobb-Douglas production function.'","section":"Section 3, Eq. (9)"},{"comment":"The phrase 'typically approaching zero' is unclear; if w∞ is the asymptotic AGI wage level, it should be a positive parameter, not a quantity approaching zero. This ambiguity should be resolved.","section":"Section 5.1, Eq. (24)"},{"comment":"References [2] and [10] appear to be the same item and should be merged; several entries are missing complete publication details or arXiv identifiers.","section":"References"},{"comment":"The paper contains numerous formatting and spacing errors, including missing spaces between words in Section 3.2 and elsewhere; a careful copyedit is needed.","section":"General"}],"recommendation":"reject","confidential_remarks":"The paper's policy discussion is reasonable, but the formal economics is not. The central mathematical claim is the reverse of what the model implies in two separate places (Section 2.2 and Section 3.1), and the subsequent exponential wage path is assumed rather than derived. These are load-bearing errors: they invalidate the paper's core thesis that AGI substitution drives human wages to zero. The shortcomings are not mere presentation issues; they concern the fundamental derivation. In my view, the paper cannot be repaired within its current scope without replacing the production-function analysis entirely. I would recommend rejection, though I would encourage the authors to resubmit a corrected version that either uses a production function with a finite limit as human labor vanishes or explicitly treats the replacement of human labor as a technological regime change rather than a smooth Cobb-Douglas limit."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague,\n\nRead Stiefenhofer's AGI/social contract paper. Quick take: the policy discussion is sane, but the formal model that supposedly backs it has a load-bearing error, and the rest is borrowed.\n\nWhat's actually new: not much. The paper extends a standard Cobb-Douglas function by adding K_AGI to K or L_AGI as a second labor input, then concludes wages go to zero as human labor vanishes. That conclusion is already the thesis of Acemoglu-Restrepo and Korinek-Suh, both cited. The proposed remedies (UBI, cooperative AGI ownership, capital taxation) are the standard menu in that literature.\n\nWhat it does well: it frames the aggregate demand paradox clearly, and the policy section is coherent. A non-specialist reader would come away with a fair sketch of the debate and the relevant citations.\n\nThe soft spots are not minor. The core derivation in Sections 2 and 3 is wrong in a simple way. In Eq. (7), the marginal product of labor is beta A (K+K_AGI)^alpha L^(beta-1). Since beta-1 is negative, this diverges to infinity as L goes to zero, not to zero. Same problem in Eq. (10) for human labor w1. So the paper's claim that wages drop to zero is the opposite of what its own production function implies. The exponential decline in Eq. (23) is assumed, not derived, and is inconsistent with the marginal-product wage from Model III. There is no data, no calibration, no robustness check. The \"index function\" in Section 5 is just an exponential curve with free parameters.\n\nI also note the paper cites Stiefenhofer's own work on Cobb-Douglas stability but doesn't use it. The self-citation is not the problem; the missing derivation is.\n\nWho is this for? A reader who wants a quick outline of the argument that AGI could eliminate wage labor and that social contract renewal is needed. As an op-ed it's fine. As an economics paper it doesn't hold up, because the one technical result that would make it a contribution is false.\n\nMy recommendation: desk reject. Send it back to the author with the Eq. (7)/(10) point and suggest either fixing the model or reframing as a non-technical policy piece. Not worth referee time in current form.","headline":"The paper's central wage-collapse result is mathematically backwards under its own Cobb-Douglas equations, leaving a policy essay that adds little to the existing AGI-and-labor literature.","tokens_in":7299,"tokens_out":1925,"would_cite":false,"duration_ms":16983,"reading_group":"no","serious_thinker":"no","would_accept_peer_review":false},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"The paper argues that AGI will drive human wages to zero and make a new social contract necessary.","keywords":["artificial general intelligence","Cobb-Douglas production function","wage collapse","aggregate demand","universal basic income","social contract","labor displacement","wealth concentration"],"falsifier":"Evaluate the paper's own wage equation (10) at a small positive $L_1$ with $0<\\beta_1<1$: the factor $L_1^{\\beta_1-1}$ diverges to infinity, so the predicted wage is arbitrarily large, not zero; repeating this calculation for any sequence $L_1\\to 0$ would overturn the claimed limit.","tokens_in":6285,"feed_emoji":"🤖","tokens_out":5633,"duration_ms":45879,"temperature":0.7,"pith_summary":"The paper tries to establish that AGI, once it can substitute for human labor, undermines the wage-based economy: human wages fall to zero, income concentrates among AGI and capital owners, and aggregate demand collapses. It uses extended Cobb-Douglas production functions to derive this trajectory and an index of human economic power that decays as AGI labor expands. A sympathetic reader would care because the conclusion is that the social contract must be renegotiated through UBI, cooperative AGI ownership, or progressive AGI capital taxation to avoid economic collapse. The argument is a formalization of a widely discussed risk, not an empirical prediction.","feed_headline":"AGI wages to zero: a new social contract","feed_subtitle":"A Cobb-Douglas model predicts employment collapse and demands UBI or shared AGI ownership.","key_machinery":"The machinery is the Cobb-Douglas production function, extended in three versions: AGI as capital ($Y=A(K+K_{AGI})^{\\alpha}L^{\\beta}$), AGI as labor ($Y=AK^{\\alpha}L_1^{\\beta_1}L_2^{\\beta_2}$), and both AGI labor and AGI capital ($Y=AK^{\\alpha}K_{AGI}^{\\gamma}L_h^{\\beta_1}L_{AGI}^{\\beta_2}$). Wages are identified with marginal products, and the limiting argument sends human labor to zero; the normalized power index $P_h=w_hL_h/(w_hL_h+w_{AGI}L_{AGI})$ is then specified to decay exponentially with $L_{AGI}$.","core_discovery":"On its own terms, the paper's central claim is that modeling AGI as either capital or labor inside a Cobb-Douglas production function makes human labor redundant in the long run. Since wages are set by marginal product, the labor term vanishes from the production function once AGI fully replaces humans, and the wage converges to zero; with no wage income, consumption demand collapses even as AGI-driven output expands. The paper also constructs a normalized measure of human economic power, $P_h$, that falls exponentially as the AGI labor share rises, and treats this as a continuous transition from a decentralized human-work economy to a centralized AGI-capital economy.","pith_inferences":["The same Cobb-Douglas logic, taken literally, points the other way: the stated wage formula diverges as human labor approaches zero, so a consistent model of a post-labor economy needs a different production technology or an explicit assumption that human labor exits the production function entirely.","If the zero-wage result is replaced by a high-wage result for the last remaining workers, the political conclusion still holds only if ownership of AGI is concentrated; a testable extension would compare wage shares in AI-intensive sectors against the model's predicted monotone decline.","The social-contract framing implies a natural policy comparison: UBI versus capital grants versus public ownership can be evaluated by the same aggregate-demand criterion, which the paper does not perform."],"forward_implications":["If human wages fall to zero, employment-based income cannot support consumption, so the economy faces a Keynesian demand crisis even as production rises.","All economic surplus accrues to AGI owners, concentrating wealth and freezing social mobility.","Standard income-distribution models, which depend on wage labor, stop describing the economy and must be replaced.","Policies such as UBI, public or cooperative AGI ownership, and progressive AGI capital taxation become necessary, not optional, for stability."],"supporting_citations":[{"why":"Supplies the Cobb-Douglas production function on which all three models are built.","marker":"[3]"},{"why":"Underpins the marginal-productivity theory of wages used to derive wage collapse.","marker":"[4]"},{"why":"Recent statement of the Cobb-Douglas function used for the model's functional form.","marker":"[5]"},{"why":"Provides the race-between-man-and-machine framework the paper extends to AGI.","marker":"[15]"},{"why":"Directly addresses AGI and future labor demand, the claim the paper formalizes.","marker":"[16]"}],"fun_headline_variants":["AGI labor to zero human wages, forcing new social contract","AGI may collapse wages to zero; UBI or shared ownership needed","AGI makes human labor redundant: wages zero, demand collapses","Post-labor economy: AGI forces renegotiation of social contract","AGI pushes human wages to zero, fueling need for UBI"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that the Cobb-Douglas marginal-product formula remains the correct wage rule as human labor shrinks to zero; under that formula the wage blows up instead of vanishing, so the zero-wage conclusion depends on an unstated different production function or on removing human labor from the technology.","fun_headline_variants_meta":{"raw":{"variants":["AGI labor to zero human wages, forcing new social contract","AGI may collapse wages to zero; UBI or shared ownership needed","AGI makes human labor redundant: wages zero, demand collapses","Post-labor economy: AGI forces renegotiation of social contract","AGI pushes human wages to zero, fueling need for UBI"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000355,"raw_usage":{"total_tokens":1872,"prompt_tokens":833,"completion_tokens":1039,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":449,"completion_tokens_details":{"reasoning_tokens":948}},"tokens_in":449,"tokens_out":1039,"duration_ms":7748,"temperature":1.0,"reasoning_tokens":948,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-08T13:55:42.943374+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Evaluate the paper's own wage equation (10) at a small positive $L_1$ with $0<\\beta_1<1$: the factor $L_1^{\\beta_1-1}$ diverges to infinity, so the predicted wage is arbitrarily large, not zero; repeating this calculation for any sequence $L_1\\to 0$ would overturn the claimed limit.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the Cobb-Douglas production function on which all three models are built."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Underpins the marginal-productivity theory of wages used to derive wage collapse."},{"cited_title":"A note on the Cobb-Douglas function","cited_arxiv_id":"2411.08067","evidence_quote":"Recent statement of the Cobb-Douglas function used for the model's functional form."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the race-between-man-and-machine framework the paper extends to AGI."}],"review_version":1}