AI is helping us write more code than ever before — but studies suggest little improvement in actual end-to-end productivity. So why isn't more code translating to better software?
In this talk, Sam Newman examines the bottlenecks that AI-assisted development is exposing, and asks whether the way we write code needs to fundamentally change. Drawing on concepts from robotics and automation, he explores what working effectively with AI coding agents actually looks like — and tackles the provocative question: should humans even be reading code anymore?
Sam also looks at how changing the architecture of our systems can open up new ways of thinking about software development, while helping teams navigate the uncertainty that comes with relying on AI models.
