How do you explain reasoning models inside a story about the people building them? For “How Codex Happened,” the answer was to make OpenAI itself the character doing the thinking. Between sections, imagined reasoning traces follow the company as it tries an approach, sees what works, and changes course. The device lets readers follow the company’s decisions while getting a feel for the technology behind Codex. @danshipper and his editors found a way to make the explanation part of the…