Cat Wu has joined OpenAI to lead product management for its Core Products division, marking a major development in the artificial intelligence industry. Her move comes just two days after she announced her departure from rival AI company Anthropic on November 3, 2026.
Wu announced her new position in a post on X on November 5. She expressed enthusiasm about joining OpenAI and praised the company's ability to scale AI technology, bring its tools to more than a billion people and make product development a central part of its mission.
In her announcement, Wu wrote: "I’m excited to work with @thsottiaux, an incredible PM team, and everyone across OpenAI. Together, we’ll build agents that have your back and your team’s, and keep you in control, so you get more time for the people and work you love. Lots to learn, and lots to build. Please keep your feedback coming!"
OpenAI President and co-founder Greg Brockman publicly welcomed Wu to the company. Thibault Sottiaux, who leads OpenAI's Core Products division and is known as @thsottiaux on X, also welcomed her and expressed his interest in working together.
From Claude Code to OpenAI
Wu brings considerable experience in developing AI-powered products. She spent two years working on Claude Code at Anthropic, joining when the product was still a relatively small internal project.
During her time at Anthropic, she played a role in shaping the coding assistant and exploring how AI could independently handle complex assignments.
One notable example involved her asking Claude Code to build a reinforcement-learning environment. She reportedly guided the AI through the process for dozens of hours without manually writing code herself.
Following her departure, Wu acknowledged the contributions of Claude Code, Cowork, the teams behind these products and users whose feedback helped shape their development.
Her experience working with AI tools that can independently execute lengthy tasks is expected to be relevant to her new responsibilities at OpenAI.
OpenAI's plans for persistent AI agents
Wu's emphasis on keeping users in control aligns with OpenAI's broader ambitions to develop more capable AI agents.
In June 2026, OpenAI announced plans to acquire Ona, a startup specialising in secure cloud-based execution. The proposed integration aims to bring Ona's technology into Codex, OpenAI's AI-powered coding platform.
Codex reportedly has more than 5 million weekly users, representing a 400 per cent increase compared with earlier figures from the year.
The platform has also expanded beyond software development to support research, data analysis and workflow automation.
OpenAI's next objective is to develop what it describes as persistent AI agents. These systems would be capable of continuing to work on tasks for several hours, even after users close their laptops, while operating securely within customers' own cloud environments.
Users would be able to monitor progress remotely, provide additional instructions, make decisions and review completed assignments without having to remain actively involved throughout the process.
Why persistent AI agents matter for businesses
For businesses, maintaining control over autonomous AI systems is particularly important. Organisations need clear restrictions governing where AI agents operate, which information they can access and how their activities are recorded and reviewed.
By combining Wu's experience in building products that delegate complex tasks to AI with OpenAI's investments in cloud-based execution, the company aims to expand the capabilities of autonomous AI systems while retaining human oversight.
The broader objective is to develop AI agents that can independently manage increasingly complicated assignments without removing users' ability to direct their activities, inspect their work and maintain control over important decisions.
