Managing Code at Scale
summary
Companies report developers spend roughly a quarter to a third of coding time on refactoring (Microsoft 26%, SAP ~30%), averaging about two hours per week, and refactoring plus maintenance can reach 40%. Teams build tools like UI5 flintter to automate large upgrades, deploy open rewrite recipes and Madern to apply fixes at scale, and use GitHub Copilot (SAP saw ~10–15% productivity gains) — but AI increases code churn and duplication, so human review, test automation, risk scoring and renovate-style auto-merges remain essential.
AI-written summary · about
chapters (8)
- 0:00Introduction and welcome
- 2:29Refactoring time impact stats
- 7:50GenAI code creation impact
- 19:25Ensuring AI-generated code quality
- 28:55Managing PRs at scale
- 39:46Measuring tech debt impact
- 48:13GenAI for testing challenges
- 58:10Closing remarks
related talks
51mFrom Developer Pain to Impact: Establishing Large Scale Refactoring at SAP
46mLet Them Eat Code
19mIf AI Writes the Code, Who Runs Production?
29mDevTools for the decade to come
45mRefactoring in Business World... for Developers?
49mHow to sell a big refactor or rewrite to the business?
51mFuture of Software Development
21mAI Code Quality Debt: What Strong Engineering Teams Do Differently