“Tokenmaxxing” is making developers less productive than they think
TechCrunch
Developers using AI are prioritizing high token usage, but data suggests this increases code churn and lowers overall productivity.Summary
While AI coding tools are generating more code, industry data from firms like Waydev, GitClear, and Jellyfish suggests that this volume comes at the cost of high technical debt. Developers are increasingly suffering from 'code churn,' where AI-generated code is frequently revised shortly after acceptance. Evidence indicates that while token-heavy workflows may produce more pull requests, they often yield higher costs and lower quality, highlighting a gap between perceived AI productivity and actual software development value.
(Source：TechCrunch)