I’ve been using OpenAI’s codex has my agentic harness application for personal use and at work too. And, I must say I’ve been tired of downloading new updates every single day on multiple devices.
It’s like they ship something twice or thrice a day (at-least on the macOS desktop app). And, I’ve been wondering ever since.
How do they ship at this cadence? I get it that they have a large engineering team but still…
Surely, they’re doing something right.
That led me to question “What does their SDLC look like?”
I asked this question into Codex itself!
The answer is not in this PDF but it’s breadcrumb trail to the answer:
https://cdn.openai.com/business-guides-and-resources/building-an-ai-native-engineering-team.pdf
~ updated at: 2026-10-19T15:05:33.214Z