Research · Sources
Research references
Last updated: 2026-08-12
Public sources behind the essay The Reviewer Still Wins. Local package paths omitted.
Primary
- OpenAI — Measuring the performance of our models on real-world tasks (GDPval)
- Patwardhan et al. — GDPval paper (arXiv:2510.04374)
- Hugging Face — openai/gdpval gold set
2026 leaderboards & model releases
- Artificial Analysis — GDPval-AA v2 Leaderboard
- Artificial Analysis — Grok 4.6 benchmarks
- xAI — Introducing Grok 4.6
- Snorkel — Grok 4.5 on GDPval+
Historical parallels
- Robert Solow, New York Times Book Review (12 Jul 1987) — productivity paradox remark
- Paul A. David, “The Dynamo and the Computer,” American Economic Review 80(2) (1990)
Appendix map (this site)
| Topic | Page |
|---|---|
| What GDPval is | 01 |
| Dataset & grading | 02 |
| Quality | 03 |
| Productivity math | 04 |
| Implications | 05 |
| 2026 boards / Grok 4.6 | 06 |
| Question ledger | ledger |