AI Coding Tools Boost Shipped Releases by 30%, Humans Remain Complementary
soumitrashukla9 · x · 2026-07-28
Discusses the paper "Writing Code vs. Shipping Code," which combines public GitHub records for over 100,000 developers with confidential Microsoft data to analyze the impact of different generations of AI coding tools across the software production hierarchy.
The study reveals an attenuation effect of AI productivity gains along the code production chain. While AI generates enormous gains of 1000% or more at the top of the chain (e.g., lines of code), it translates into only about a 30% increase in actually shipped releases.
This finding helps rationalize the current employment effects, explaining the lack of job displacement. It suggests that humans act as a necessary complement to AI-friendly tasks, with increased hiring observed in exposed areas.
Related event: AI Coding Tools Yield Only 30% Productivity Gain(3 posts)→
More from AGI Musings
- AI math era taught an order of magnitude more people what frontier math looks like — tszzl · 2026-09-23
- Beyond technical alignment: repligate clashes over whether AI can produce rich qualia — repligate · 2026-09-23
- Mathematicians, not just LLMs, made AI's math breakthroughs possible, scholars argue — tak3sh8 · 2026-09-23
- Why would an uncontrollable superintelligence do anything for us? Reddit debate — conn_r2112 · 2026-09-23
- X user calls for full-speed AI-driven science: braking research is 'an absurd waste' — Dr_Singularity · 2026-09-23
- Is Using LLM Output Plagiarism? A Debate Over Redefining Writing Ethics — soumitrashukla9 · 2026-09-23