The OpenAI engineer who turned AI coding into one of the company's fastest-growing revenue streams is now taking on an even bigger challenge. Thibault Sottiaux, who led the development of Codex, is overseeing what insiders describe as the most sweeping transformation of ChatGPT since its launch. The move signals OpenAI's bet that the lessons learned from building tools for developers can reshape how millions of consumers interact with AI.
OpenAI is betting big on the engineer who cracked the code on AI coding. Thibault Sottiaux, the technical lead behind Codex, is now steering what Wired describes as ChatGPT's biggest transformation yet. It's a promotion that reveals as much about OpenAI's strategic priorities as it does about Sottiaux's track record.
Codex didn't just succeed, it exploded. The AI model that powers GitHub Copilot became one of OpenAI's fastest-growing revenue generators, helping developers write code faster and with fewer errors. Under Sottiaux's technical leadership, Codex evolved from an experimental project into a product that Microsoft integrated across its developer tools. The business case was clear: developers would pay premium prices for AI that understood their workflows.
Now Sottiaux faces a different challenge entirely. ChatGPT serves millions of users across every skill level and use case imaginable, from students drafting essays to executives analyzing market trends. The product that put OpenAI on the map has also become its biggest liability, as users complain about inconsistent responses, confusing interfaces, and a feature set that's grown bloated over time.
The timing is telling. OpenAI recently reorganized its product teams, consolidating what were previously separate consumer and enterprise divisions. Sottiaux's appointment suggests the company believes the structured, workflow-oriented approach that worked for Codex can translate to ChatGPT's chaotic consumer landscape. It's a theory that hasn't been proven at this scale.
Competition is intensifying on both fronts. Anthropic is winning over developers with Claude's superior code generation, while Google keeps iterating on Gemini's multimodal capabilities. Meta just open-sourced its latest Llama models, putting pressure on OpenAI's pricing power. In the consumer market, startups like Perplexity are unbundling ChatGPT's features into focused products that do one thing exceptionally well.
Sottiaux's background offers clues about the direction ChatGPT might take. Codex succeeded by understanding context, anticipating what developers needed next, and integrating seamlessly into existing workflows. Apply those principles to ChatGPT and you get a product that's less chatbot, more intelligent assistant that knows when to step in and when to get out of the way.
The Codex playbook had another advantage: developers are technical users who can articulate exactly what they want. ChatGPT's millions of everyday users rarely know what's possible until they stumble onto it. That gap between capability and discoverability has plagued ChatGPT since launch, leading to a paradox where the product is simultaneously too complex for new users and too limited for power users.
Industry observers see this transition as a test of OpenAI's product philosophy. Can engineering-led thinking scale to consumer products, or does mass-market AI require a different approach entirely? Microsoft, which has embedded ChatGPT across its consumer products, is watching closely. So is Apple, which reportedly delayed its AI features partly due to concerns about consistency and user experience.
The stakes extend beyond product design. ChatGPT's trajectory will influence how OpenAI monetizes its technology going forward. The company's enterprise business is growing, but consumer subscriptions still drive significant revenue. A successful overhaul could justify premium pricing tiers and expand the addressable market. A failed redesign could push users toward competitors just as they're gaining credibility.
Sottiaux's challenge is also organizational. ChatGPT touches nearly every team at OpenAI, from safety researchers testing for harmful outputs to infrastructure engineers managing scale. Codex had clear boundaries and a defined user base. ChatGPT is the company's public face, which means every change carries reputational risk.
The transformation is already underway, though OpenAI hasn't detailed specific changes. Engineers familiar with the project say the focus is on personalization, context retention across sessions, and smarter defaults that reduce the need for prompt engineering. Think less about teaching users to write better prompts and more about building AI that understands intent without being told.
Sottiaux's move from Codex to ChatGPT represents more than a personnel shift, it's a strategic signal about how OpenAI plans to compete in an increasingly crowded AI market. The company that pioneered conversational AI is now scrambling to reinvent it before competitors do it first. Whether the discipline that made Codex a developer favorite can scale to ChatGPT's massive, diverse user base remains the open question. For OpenAI, getting this right isn't just about improving one product, it's about proving the company can evolve as fast as the technology it's creating.
