How VS Code Went from Monthly to Weekly Releases with AI — Harald Kirschner
This post details how VS Code transitioned from monthly to weekly releases by rebuilding its development system around AI agents. Key contributions include: making codebases 'agent-ready' by enabling agents to drive the application and use tools like Playwright for self-correction; encoding expertise into 'skills' for agents; achieving 10x faster builds with TypeScript Go; implementing mandatory AI code review and AI-driven issue triage; and transforming error stacks into auto-fix PRs. It also highlights the importance of VSC-Bench for evaluating AI performance and token efficiency, and the benefits of smaller, more focused squads for durable ownership. The talk quantifies the increase in agent-written code committed (55% to 86%) and discusses the challenges of scaling CI with multiple agents.
More from the Software Factory Engineering trail
- From Your Laptop to the Pipeline: Scaling Custom Agents with GitHub Copilot
- Generation Is Cheap, Review Is Expensive: How to Stop Shipping AI Slop — Gabriel Martinez, G2i
- Let Anyone at Your Company Ship Internal Apps with AI — Garrett Galow, WorkOS
- The Software Factory: From Bug Report to Production Code — Davis Palmie, Factory
- Why AI Agents Should Have Their Own Sandbox — Philipp Schmid, Google DeepMind