A technical assessment of the funding threshold in Nayebi (2026)

This technical note works through the model of Nayebi (2026), "When do AI gains become broadly shareable? A policy threshold for AI-driven automation" (AIES 2026; proofs in arXiv:2505.18687). It uses only the equations and the calibration of that paper, together with the definition of the benefit given in the author's AIES 2026 poster. It asks which conclusions about the funding threshold hold, and under which assumptions. The note reaches five results. (1) In the production function of the paper, GDP depends on AI productivity and on capital only through the AI income share. For given capital and labour, higher AI productivity lowers GDP, and AI income has an upper limit at every date. (2) The reported threshold of 5.6 is computed at a point where the funding condition requires an AI income share of 190%, while the share is always below 100%. (3) With the fixed benefit and the capital rule of the paper, only one constant growth factor of AI productivity is compatible with balanced growth, 0.375% a year. On that path the AI income share stays at 14.1% and the benefit is funded after 236.4 years with the current public share. (4) On any path, the benefit cannot be funded before 202.8 years with the current public share, or 127.1 years with a one-third share. (5) At a given date the funding condition is two-sided: AI productivity can be too high as well as too low. These are properties of the model as written and calibrated. They are not forecasts. The role of the public share and of the cost share in the funding condition of the paper is unchanged. The note also shows that the long-run convergence of capital/output stated in the appendix of the paper is correct when AI productivity is constant. This is a working paper and has not been peer reviewed. It is a separate study in the same research programme as "Intelligence Scales, Matter Does Not" (De Joannon 2026). It assesses a model cited in the literature review of that paper and does not change its propositions.

Authors

Publication Details

Journal
Zenodo (CERN European Organization for Nuclear Research)
Published
2026-11-16
DOI
https://doi.org/10.5281/zenodo.24287897
Primary Topic
Economic Growth and Productivity
Type
article
Field-Weighted Citation Impact
0.00
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
article

A technical assessment of the funding threshold in Nayebi (2026)

Alfredo De Joannon
1 links
Zenodo (CERN European Organization for Nuclear Research)
Economic Growth and Productivity
article

A technical assessment of the funding threshold in Nayebi (2026)

Alfredo De Joannon
article en

Abstract

This technical note works through the model of Nayebi (2026), "When do AI gains become broadly shareable? A policy threshold for AI-driven automation" (AIES 2026; proofs in arXiv:2505.18687). It uses only the equations and the calibration of that paper, together with the definition of the benefit given in the author's AIES 2026 poster. It asks which conclusions about the funding threshold hold, and under which assumptions. The note reaches five results. (1) In the production function of the paper, GDP depends on AI productivity and on capital only through the AI income share. For given capital and labour, higher AI productivity lowers GDP, and AI income has an upper limit at every date. (2) The reported threshold of 5.6 is computed at a point where the funding condition requires an AI income share of 190%, while the share is always below 100%. (3) With the fixed benefit and the capital rule of the paper, only one constant growth factor of AI productivity is compatible with balanced growth, 0.375% a year. On that path the AI income share stays at 14.1% and the benefit is funded after 236.4 years with the current public share. (4) On any path, the benefit cannot be funded before 202.8 years with the current public share, or 127.1 years with a one-third share. (5) At a given date the funding condition is two-sided: AI productivity can be too high as well as too low. These are properties of the model as written and calibrated. They are not forecasts. The role of the public share and of the cost share in the funding condition of the paper is unchanged. The note also shows that the long-run convergence of capital/output stated in the appendix of the paper is correct when AI productivity is constant. This is a working paper and has not been peer reviewed. It is a separate study in the same research programme as "Intelligence Scales, Matter Does Not" (De Joannon 2026). It assesses a model cited in the literature review of that paper and does not change its propositions.

Zenodo (CERN European Organization for Nuclear Research)
Economic Growth and Productivity
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon