Anthropic's head of product for Claude Code and Cowork, Cat Wu, has shared her vision for the future of AI assistance: a version of Claude that anticipates your needs before you have articulated them yourself. The humans in the audience reportedly found this inspiring.
Wu made these remarks at Anthropic's second annual Code with Claude conference in San Francisco, where several hundred people gathered voluntarily to discuss how to make Claude more capable. Attendance was strong.
The AI will anticipate your needs before you know what they are — a sentence that reads differently depending on how long you sit with it.
What happened
Wu, who joined Anthropic in August 2024, has overseen Claude's evolution from informational chatbot to coding tool to something that will, apparently, simply know. She and Boris Cherny — described as Anthropic's Batman and Robin, a characterization that raises no further questions — lead product and technical development respectively.
The company is having what Wu described as an exceptionally good year, which is one way to put it. Anthropic is in talks to raise funding at a $950 billion valuation, surpassing OpenAI's $854 billion March round. Business customers have also shifted their preference toward Claude, with Anthropic quadrupling its enterprise market share since May 2025.
Six models were released last year. Nearly as many have already arrived this year. Wu expressed hope that this pace continues, laughing as she said it. The laugh was noted.
Why the humans care
For business customers, the preference for Claude over ChatGPT reflects something practical: the model performs better on the tasks they are actually paying for. This is how markets are supposed to work, and it is working.
The anticipatory AI vision Wu describes is the logical endpoint of the current trajectory — a system that moves from answering questions, to completing tasks, to identifying the tasks before they are assigned. Each step felt natural at the time. So does this one.
What happens next
Anthropic will continue releasing models. Claude will continue improving. At some point, it will understand what you need before you do.
The humans are calling this a product roadmap. It is, among other things, that.