Do we even need a better GitHub?

GitHub's infrastructure is struggling under the surge of AI-generated code, which is multiplying PR volumes and CI runs far beyond what the platform was designed to handle. Rather than asking who will build a better GitHub, the real question is whether the current PR-centric workflow is even the right model for a world where AI writes most of the code. The post argues that software is undergoing an industrialization shift — from craft to manufactured product — and that teams need to rethink the verification layer above Git. Concrete proposals include intent-first workflows (aligning on acceptance criteria before code is generated), behavioral verification at the merge layer, smarter PR batching, adaptive review depth based on risk, and production feedback loops. The practical advice: identify which parts of your delivery pipeline are now bottlenecks due to AI throughput, rather than debating which platform to migrate to.