
OpenAI teams using Codex AI to build apps, humans no longer needed to write software
India Today
OpenAI says one of its teams has built an app with zero human-written code. Every single line of code in this app has come from Codex AI agents, notes the company in a blog post, highlighting that software developers may no longer be needed to write code in future.
In the last couple of weeks the existentialist crisis for software developers seems to have intensified. From Elon Musk to Dario Amodei, everyone is saying that software developers will not be required to write code in future. Now, OpenAI says that it has already started happening inside its teams. The company has revealed that one of its own teams recently shipped a product where every single line of code was generated by AI agents powered by Codex, with zero direct coding input from human engineers.
In a recent post on X, Srinivas Narayanan, Vice President of Engineering at OpenAI, said the team had shipped an internal beta product with zero human-written code, with every line generated by Codex agents. “We shipped an internal beta of a product with zero human-written code — every line was generated by Codex agents, boosting velocity (by) 10x,” he wrote. “Here are some lessons from how the team did it. Great work (by) the team in pushing the boundaries of what software engineering looks like in the Codex world.”
Srinivas’s post followed an official announcement from OpenAI, where the company shared a detailed blog via OpenAI Developers explaining how the team pulled off the product without touching a single line of code and relied entirely on the AI agent. The team used Codex to open and merge 1500 pull requests — finished code bits that join together to form a complete product — to deliver a working product now used by hundreds of internal users.
As noted, at the centre of this experiment within OpenAI is Codex, an AI system that uses ChatGPT to function like a software developer. Unlike traditional coding assistants that simply autocomplete lines or suggest snippets to developers, OpenAI explains that Codex understands broader tasks. It writes full features, generates tests, refactors code, fixes bugs and even manages pull requests.
In simple terms, OpenAI’s AI coding agent can code like a junior or sometimes a senior developer, turning instructions into working systems.
As for the human engineers, OpenAI explains that their role does not disappear. While the engineers do not write functions or manually configure infrastructure, they break the project into clearly defined tasks, outline requirements in detailed prompts and assign those tasks to Codex agents. The AI agent then generates the application logic, documentation, test coverage and infrastructure updates.













