AI coding agents generate more code, but not more software

7.8 relevance

Directly addresses AI coding agents and their impact on software development productivity.

AI/ML arstechnica.com

Answer the board on AI readiness

Arrive with evidence, not opinion. See how your foundation benchmarks against your peers, get a clear read, and leave with a ranked place to start.

Why It Matters

For a solutions architect focused on AI/ML and developer experience, this reveals that AI coding tools shift the bottleneck from writing code to reviewing it, demanding new tooling and process changes to capture efficiency.

Key Takeaways

Invest in code review automation and process optimization to prevent AI coding agents from inflating output without improving software delivery.

AI coding agents generate more code, but not more software

Summary

A Harvard study of 700+ firms using Jellyfish data finds AI coding agents increase lines of code by 30% and commits by 20%, but code review time surges 49%, erasing productivity gains. The bottleneck—more PR revisions and comments—means no significant rise in software output or employment. The efficiency from AI-generated code is absorbed by downstream review constraints.

Author

Kyle Orland — Kyle Orland has been the Senior Gaming Editor at Ars Technica since 2012, covering topics ranging from retro games to new gaming hardware, business and legal developments in the industry, fan communities, gaming mods and hacks, virtual reality...