Cat Wu joined Anthropic in the summer of 2024 after trying an early Claude Code prototype. “It worked, I was hooked, and I joined the team,” she later recalled.¹

By February 2025, Anthropic had released Claude Code in research preview. Wu and the tool’s creator, Boris Cherny, became its public-facing leaders as developers embraced AI that could take on larger coding assignments. The product’s rise gave Anthropic a viral commercial engine, lifting usage of its models and sharpening its rivalry with OpenAI.

Wu and Cherny briefly left for coding startup Cursor that summer before returning to Anthropic, where they oversaw Claude Code’s rapid expansion late in 2025. Wu also took responsibility for Cowork, Anthropic’s non-coding workplace tool, and described her task as helping push Claude Code toward an “AGI-pilled” future.¹

Then came the break. Wu left Anthropic last week, announcing her departure on X on Tuesday — only weeks before the company’s anticipated IPO. She did not publicly explain why she departed, and Anthropic declined further comment. The timing leaves the startup without one of the most visible product executives behind a tool that had become central to its growth story.

On Wednesday, Wu said she was joining OpenAI to lead product management for Core Products. She cited OpenAI’s reach to more than a billion users and said the aim was to build agents that “have your back and your team’s, and keep you in control.”² Greg Brockman, OpenAI’s co-founder and president, greeted the hire with a simple public message: “welcome @_catwu to the team!!”Greg Brockman (gdb)³

For OpenAI, the hire brings in a product leader who helped make Claude Code a formidable enterprise competitor to its own Codex. For Anthropic, it is an uncomfortable defection just as investors are being asked to believe its momentum can outlast the people who built it.