Why your AI code doesn’t ship: Closing the gap to production | BRK200
GitHub's product lead declares the traditional SDLC 'more or less dead,' replaced by human-agent collaboration requiring only micro-steering.
“And I think the SDLC is more or less dead. It's merging with like the product development life cycle”
At Build 2026, GitHub product lead Mario Rodriguez argued that since January AI agents have 'entered the workforce' and now work with only micro-steering rather than constant correction, collapsing the traditional software development lifecycle into human-agent collaboration. He framed the new bottleneck as shipping AI-written code to production, signaling a strategic repositioning of GitHub Copilot across the entire SDLC including deployment and live debugging.