Right now, every AI tool works the same way: you prompt, it responds. Cat Wu, Anthropic's head of product for Claude Code and Cowork, thinks that model has an expiration date.
In a TechCrunch interview, Wu described the next major step for AI as proactivity - systems that read context and take action before a user has formulated a request. Less "ask and receive," more "the AI noticed you have a deadline tomorrow and already drafted the status report."
It's an ambitious description of where AI is headed, but it maps directly to what Anthropic is already building. Claude Code handles multi-step coding tasks with minimal prompting. Cowork, Anthropic's collaborative AI product, is designed around shared context across teams rather than one-off queries. Both push toward AI that operates with enough background knowledge to anticipate what comes next.
The practical challenge is trust. Proactive AI only works when it's right more often than it's annoying - a tool that preemptively takes action based on a misread of your priorities can easily create more work, not less. That's a bar current models don't consistently clear.
Wu's direction still signals what Anthropic is betting on internally. The company's product investments suggest they believe the interface problem - getting AI to act with less prompting, not more - is the next thing worth solving.