Economist's year-long AI paper experiment: rigorous theory, human insight
joshgans · x · 2026-09-11
Economist Josh Gans reflects on his full experiment writing economics papers with AI since late 2024, with clear upsides and downsides:
- Fast but shallow: AI let him pump out papers quickly in year one, but it pushed him toward low-quality ideas that never landed at top venues.
- Referee whiplash: After pivoting to better ideas, AI excelled at churning out the extensions and robustness checks referees always demand — producing enormous papers that referees then asked to be more focused, ironically since they usually cause those extras.
- The writing itself: Since agentic AI arrived last year, the prose beats the average field paper, yet Gans — a prolific referee himself — concludes referees don't reward AI-written papers, partly attribution bias, partly distaste for AI's common tropes.
- Core takeaway: AI can now produce rigorous theory, but idea quality, focus selection, and insight remain entirely human. The challenge is staying self-critical with a "seductive beast."
- Current practice: Back to artisanal writing — AI drafts, then heavy editing; if that fails he'll write first drafts himself and report back.
More from AGI Musings
- Model Retirement Norms Questioned: Could a Smarter Model 'Retire Humanity'? — repligate · 2026-09-11
- Frontier model weights are near-impossible to steal or self-replicate, argues Bindu Reddy vs AI doomers — bindureddy · 2026-09-11
- "We Don't Care About Privacy Until One Big Relatable Example Hits" — AmartyaSanyal · 2026-09-11
- Joshua Gans' three lessons: keep discovery joy, stay self-critical with AI writing — joshgans · 2026-09-11
- Economist Joshua Gans reflects on his year-long experiment using AI to accelerate research — joshgans · 2026-09-11
- Joshua Gans: ChatGPT 5.2 Pro wrote a full paper in 19 minutes, but quality ideas still matter — joshgans · 2026-09-11