An economist tested Refine, an AI tool for academic writing, on his inflation theory manuscript and found the results revolutionary. The tool, developed by Yann Calvó López and Ben Golub, provided feedback comparable to top 5% of referee reports he's received in 40 years. After using the free trial on his 80-page booklet, he plans to become a regular user.

Refine identified core arguments and offered specific, actionable critiques. For his fiscal theory of price level (FTPL) model, it suggested operationalizing the "fiscal news" narrative with dated observables rather than qualitative signals. The tool also caught logical tensions, like the distinction between completeness claims and observational equivalence with New Keynesian models, and recommended clarifying the fiscal regime distinction. It found algebraic errors and noticed inconsistencies in how transmission mechanisms were described.

Refine represents a fundamental shift in academic economics. The economist argues it will save enormous time in refereeing and improve paper quality. He plans to require authors to use it before submitting papers for review and will consult it before writing any future referee report. At $50 for the paid version, he believes the tool's ability to identify major points, logical gaps, and basic theorem violations makes it indispensable for academic economics.