Reproducing “Economic Scenarios for Transformative AI”
Reproducing “Economic Scenarios for Transformative AI”
reproduce.sh installs the pinned environment with uv, downloads the public
data behind the calibration rechecks, runs the test suite that gates every table, figure, and
quoted number of the report and its supplement on the code, and builds the site. Nothing here
solves a household, so the whole path takes minutes. reproduce_min.sh runs the model’s
internal identities, the published-number tables, and the quit order alone. The README describes
the report, the layout of the code, and what each check establishes.
Details
Authors
Abstract
An Econ-ARK REMARK that reimplements the monthly model of Korinek, Jones, Sacher, Cotter, and McCrory (2026) and records the twelve inputs, definitions, and orderings a reimplementation needs beyond the paper's text, most of which can be read from the authors' scenario explorer. The reimplementation agrees with the authors' scenario explorer in every series and month to rounding error and reproduces all 226 published numbers within the paper's printed rounding. One script fails when any number stops reproducing.
Actions
No GitHub repository in metadata.
Metadata
| Key | Value |
|---|---|
| DOI | 10.5281/zenodo.22839070 |
| abstract | An Econ-ARK REMARK that reimplements the monthly model of Korinek, Jones, Sacher, Cotter, and McCrory (2026) and records the twelve inputs, definitions, and orderings a reimplementation needs beyond the paper's text, most of which can be read from the authors' scenario explorer. The reimplementation agrees with the authors' scenario explorer in every series and month to rounding error and reproduces all 226 published numbers within the paper's printed rounding. One script fails when any number stops reproducing. |
| authors | Alan Lujan |
| authors-original-paper | Anton Korinek Charles I. Jones Szymon Sacher Tess Cotter Peter McCrory |
| cff-version | 1.2.0 |
| date-released | 2026-09-18 |
| doi | 10.5281/zenodo.22839070 |
| keywords | artificial intelligence macroeconomic scenarios labor reallocation household saving reproduction Econ-ARK |
| license | MIT |
| message | If you use this REMARK, cite it as below. |
| references | {"authors"=>[{"family-names"=>"Korinek", "given-names"=>"Anton"}, {"family-names"=>"Jones", "given-names"=>"Charles I."}, {"family-names"=>"Sacher", "given-names"=>"Szymon"}, {"family-names"=>"Cotter", "given-names"=>"Tess"}, {"family-names"=>"McCrory", "given-names"=>"Peter"}], "title"=>"Economic Scenarios for Transformative AI", "type"=>"report", "url"=>"https://www.anthropic.com/institute/econ-scenarios", "year"=>2026} |
| remark-name | econ-scenarios |
| repository-code | https://github.com/econ-ark/econ-scenarios |
| tags | REMARK Reproduction Notebook |
| tier | |
| title-original-paper | Economic Scenarios for Transformative AI |
| type | software |
| version | 0.1.0 |
| date |