Waldek Mastykarz • 7/30/2026

We're measuring AI productivity the wrong way

This article argues that measuring AI productivity by how much time it saves an individual is misleading. It explains that while AI reduces the cost of producing content (e.g., writing documents, code), it shifts the burden to consumers (e.g., readers, code reviewers), increasing their time spent. Using examples like design documents and pull requests, the author shows that total organizational productivity can decrease even as individual efficiency rises. The piece warns against optimizing for activity over understanding and calls for a broader perspective on AI's impact.

0 comments

#software engineering #pull requests #code review