The end of git? OpenAI CTO on the future of coding.

OpenAI is outgrowing Git as their internal code generation speeds overwhelm standard version control systems. Vijay Raji, CTO of Applications at OpenAI, explains why their team had to architect a new code repository system to handle the massive influx of AI-generated commits. We look at the breakdown of how rapid model deployment forces infrastructure changes that legacy tools simply cannot support. This conversation covers the technical lessons from the Statsig acquisition and how OpenAI manages the debt of building software that writes itself. We also look at the reality of the founder's test and what the latest model drop means for engineering workflows. Understanding how to scale AI agent development is the difference between keeping up and falling behind. Full post: https://natesnewsletter.substack.com/p/openai-pando-git-backend?utm_source=youtube&utm_medium=video&utm_campaign=openai-pando-git-backend-launch&utm_content=description Nate's Library MCP (every prompt kit and guide, inside the AI you already use): https://unlock-ai.natebjones.com/guides/how-to-connect-nates-library?utm_source=youtube&utm_medium=video&utm_campaign=openai-pando-git-backend-launch&utm_content=description My Links 🔗 👉🏻 Newsletter: https://natesnewsletter.substack.com/?utm_source=youtube&utm_medium=video&utm_campaign=openai-pando-git-backend-launch&utm_content=description 👉🏻 X: https://x.com/natebjones 👉🏻 TikTok: https://www.tiktok.com/@nate.b.jones 👉🏻 Instagram: https://www.instagram.com/nate.b.jones What's really happening inside OpenAI as AI agents write more of the code? The common story is that AI coding makes engineers faster, but the real question is what breaks downstream when they are. In this video, I share the inside scoop on what OpenAI learned when code got cheap: - Why reviews, testing, and CI/CD became OpenAI's new bottlenecks - How agents outgrew a code repository designed for offline work - What Vijaye's founder litmus test says about building on models - How OpenAI's ads team reached a billion-dollar run rate fast Faster coding is a real gain, and it only reaches customers if leaders fix the systems and decisions it overloads. Chapters: 00:00 What changed inside OpenAI 01:01 How coding changed after February 2026 01:46 Token usage up 17x at frontier labs 03:00 Reviews, testing, and CI/CD become the bottleneck 03:40 What breaks when agents push to production 04:19 Why OpenAI rebuilt its code repository 06:00 Collisions on main and sparse checkout 07:14 Handing a team the problem to solve 08:11 The next billion ChatGPT users and voice 10:30 How leaders prototype with Codex 13:19 The founder litmus test for new models 14:12 The ads team from zero code to $1B Subscribe for weekly AI strategy breakdowns, and comment which technical infrastructure challenge you want me to cover next. Listen to this video as a podcast. Spotify: https://open.spotify.com/show/0gkFdjd1wptEKJKLu9LbZ4 Apple Podcasts: https://podcasts.apple.com/us/podcast/ai-news-strategy-daily-with-nate-b-jones/id1877109372