You can write code faster. Can you deliver it faster?
summary
Generative AI increases code volume and more mispredictions are inevitable so delivery degrades unless pipelines change. DORA data shows 1.5% delivery drop and 7% stability drop per 25% AI adoption. Required fixes are measurable: predictive test orchestration, smarter failure grouping and faster root cause in local dev, universal caching and intelligent parallelization, and policy automation to avoid large high‑risk batches. Gradle BuildScan and improved observability shift left to stop incorrect or insecure code before CI.
AI-written summary · about
chapters (6)
- 0:00AI impact on software delivery throughput
- 2:10Biology and cancer research systems analogy
- 6:23Treating LLM hallucinations as system mispredictions
- 12:12Software engineering forces and continuous delivery
- 16:19Continuous delivery strategies for AI-generated code
- 23:43Summary: managing AI hallucinations with engineering practices
related talks
36mDeveloper Experience Trends & Predictions
38mGenerative DevOps
33mAI assisted software delivery
26mAI-augmented SDLC: Rethinking how we deliver software
29mHow to use AI to fight your DevOps transformation!
19mWith great power comes great responsibility in code quality
39mDevOps in the Age of AI: Human Powered Evolution
34mBeyond CodeGen: How AI & Machine Learning Are Transforming DevOps
