Vibe Coding Review Reconciles +26% More Tasks Per Week With a 19% Measured Slowdown and +441% Code-Review Time

arXiv 2608.20446·high signalXBlueskyLinkedInCopy link

Vibe Coding: Practice, Performance, Productivity, and Risk (arXiv 2608.20446, Aug 20) assembles seventeen months of empirical evidence since Karpathy named the practice in February 2025, spanning software engineering, HCI, labour economics, security and education. It confronts three contradictory productivity readings head-on: peer-reviewed field experiments reporting +26% more tasks per week, independent randomised trials measuring a 19% slowdown, and team telemetry showing code-review time up 441%. The review argues these are consistent once measurement method and time horizon are held constant, and closes with a falsifiable conjecture that gains are real on new code and shrink or reverse on mature codebases.