Writing Code vs. Shipping Code: Productivity Effects Across AI Tool Generations
This NBER working paper by Mert Demirer, Leon Musolff, and Liyuan Yang explores the productivity effects of AI coding tools across different generations. It distinguishes between the writing code and shipping code phases, providing insights into how AI tools impact software development efficiency.
We thank Joel Becker, Alexander Bick, Adam Blandin, David P. Byrne, Tom Cunningham, Jan De Loecker, Anders Humlum, Sanjog Misra, Frank Nagle, Sida Peng, Nate Rush, Suproteem Sarkar, Max Schnidman, and Chad Syverson. This paper benefited from comments by seminar participants at BIG.AI@MIT, the Chicago AI Workshop, the RAP Symposium at Harvard & MIT, the PSU AI and the Economy Initiative, the St. Louis Fed, the Government Office of the Czech Republic, Wharton’s AI and the Future of Work Conference, and the Columbia MAD conference. This project was funded in part by the Mack Institute For Innovation Management at Wharton and the Chicago AI Incubator. The views expressed herein are those of the authors and do not necessarily reflect the views of the National Bureau of Economic Research.
Mert Demirer
Mert Demirer previously held a postdoctoral research position at Microsoft and now works as a paid research consultant for the company.
Leon Musolff
Leon Musolff previously held a postdoctoral research position at Microsoft and now works as a paid research consultant for the company.
Copy Citation
Mert Demirer, Leon Musolff, and Liyuan Yang, "Writing Code vs. Shipping Code: Productivity Effects Across Generations of AI Coding Tools," NBER Working Paper 35275 (2026), https://doi.org/10.3386/w35275.
Download Citation
Feldstein Lecture
Presenter:
N. Gregory Mankiw
Methods Lectures
Presenters:
Raj Chetty
& Kosuke Imai
Panel Discussion
Presenters:
Oleg Itskhoki,
Paul R. Krugman
& Linda Tesar