As AI transforms how we build software, how do developers adapt?
"The models aren't mind readers... you have to tell them what to do."
This roundtable discussion hosted by Gopal Raman features founders building the next generation of AI coding tools, bringing together Charlie (Conductor), Ben Hylak (Raindrop), Oliver Gilan (Mesa), and Harrison Chase (LangChain) to explore how the software development lifecycle is changing in the age of AI.
The conversation surfaced some shifts in the software development lifecycle: how AI-written code introduces entirely new failure modes compared to human code; how “context engineering” is emerging as a critical skill for developers learning to teach models judgment; and how the best AI tools are sometimes less about the model itself and more about the harness around it—the memory systems, interfaces, and human-in-the-loop workflows that shape the developer experience.