OpenAI is betting that the same magic that made Codex a hit with developers can win over the broader workforce. Thibault Sottiaux, who leads all of OpenAI’s core products including ChatGPT Work and Codex, sat down with TechCrunch to explain the strategy: take agentic capabilities forged for a forgiving technical audience and package them for everyone else. The launch is already showing traction, with ChatGPT Work hitting 20 million users since its debut.
Sottiaux framed the push as a diffusion problem. Codex gave OpenAI room to experiment because engineers tolerate rough edges. Now the technology has matured enough to simplify dramatically. The design philosophy is minimalism—get out of the model’s way, let it express its full utility, and adapt the interface to humans rather than forcing humans to learn the software. That instinct explains the growth of ChatGPT Voice, which Sottiaux says has seen significant adoption because talking is simply more natural than typing.
The economic logic is straightforward. OpenAI bundled Work into the $20 Plus plan because Sottiaux believes utility drives willingness to pay. If users feel like they’re extracting far more than twenty dollars of value each month, the subscription sells itself. That’s the same bet the company made with ChatGPT classic, now extended to autonomous task completion.
Analysts have pointed out that OpenAI also needs to own the application layer, not just the API. Sottiaux didn’t dispute that framing, but he steered the conversation toward public opinion. Winning over skeptics, he argued, means showing people that AI can do entire complex tasks autonomously—not just help with writing or personal advice. The question hanging over all of this is whether everyday workers are actually ready for what Wharton’s Ethan Mollick called ‘magic.’ Claude Cowork, by contrast, surfaces choices and A/B tests. OpenAI is betting on simplicity. The 20 million user number suggests that bet is paying off so far, but adoption is not the same as trust.
