Developer Productivity in the Age of GenAI
summary
Netflix’s developer productivity team measures GenAI impact with OKRs and DORA-style metrics, runs experiments with predictive test selection and test containers, and treats LLM outputs as code that needs CI, progressive rollouts and evals. Calls out supply-chain risk like slop squatting 5.2% to 21.7% from hallucinated package names, rising costs ($155B) and signals (Atlassian $1B DX acquisition). Describes a stack of agents and a RAG pipeline that breaks the inner dev loop and creates a “pandemonium of parrots” to guard against.
AI-written summary · about
chapters (8)
- 0:00Introduction to GenAI and developer productivity
- 3:13Defining developer productivity work through OKRs
- 5:40The importance of measuring developer productivity
- 7:46Testing strategies in the age of LLMs
- 11:38Dependency management and supply chain security
- 13:23Maintaining developer flow and inner dev loop
- 15:32Safely adopting GenAI tools and systems
- 18:03Recap and closing thoughts on innovation
related talks
30mDeveloper Productivity for GenAI Apps
38mFrom code to culture: How AI is reshaping the engineering organization
27mSoftware engineering with GenAI
01h 23mOpening
39mDevOps in the Age of AI: Human Powered Evolution
38mGenerative DevOps
26mAI-augmented SDLC: Rethinking how we deliver software
34mBeyond CodeGen: How AI & Machine Learning Are Transforming DevOps