Anthropic’s Cat Wu says that in the future, AI will predict your needs before you know them.

Anthropic is having a very good year as the tech industry focuses specifically on AI models.

The company aims to raise tens of billions of dollars in a funding round that would value it at about $950 billion (OpenAI was valued at $854 billion in a March round), and enterprise customers are increasingly preferring Claude over ChatGPT, potentially putting it ahead of major competitors soon. Anthropic recently surpassed OpenAI among enterprise customers, quadrupling its market share since May 2025, according to a recent report.

Cat Wu, Anthropic’s head of product for Claude Code and Cowork, is the driving force behind that success. Since joining the company in August 2024, Wu has guided Claude through key stages and helped take it from a purely informational chatbot to a coding tool and beyond. Wu, who oversees the development of new features, is a core member of Anthropic’s technical staff and is often paired with Boris Cherny, creator of the Claude Code, so much so that the pair has been characterized as Anthropic’s “Batman and Robin.”

Wu sat down with me last week at the second annual Code with Claude conference in San Francisco to discuss how he thinks about product strategy and how he hopes the experience of using Claude will change in the future.

This interview has been edited for length and clarity.

When you consider your product strategy, how much of it is responsive to your peers and competitors? Have you given any thought to that?

Our main design objective is to continue to grow exponentially. So I think across the team, we’re instilling in everyone a lesson that AI will continue to improve. All we need to do is stay on this frontier. We don’t think about our competitors. If you think about your competitors, you’re always going to be two weeks or a month behind how fast you’re running. Therefore, staying on the frontier is usually not the best idea.

Anthropic released at least six models last year, and has already released about the same number this year. Do you think this pace of development will continue?

I hope that continues (laughs). I think the models are still improving at a very steady pace, so we should be able to continue to share them with our users. The deployment may look a little different, such as how we handled Glasswing, but we handled Glasswing because we want this intelligence to benefit as many people as possible, and we want it to be handled in a very secure way. [in the way that we did].

[Glasswing is an initiative that Anthropic launched in April that invited a small consortium of partner organizations — including companies like Amazon, Apple, CrowdStrike, and Microsoft — to gain access to its new cybersecurity model, Mythos. Unlike many of Anthropic’s other AI models, Mythos is not being given a general public release. The company has claimed that it fears the model — which is designed to scan codebases for software vulnerabilities — is too powerful and could be weaponized by bad actors.]

You’ve said in a previous interview that the future of work will essentially be about staff managing large numbers of agents. If this happens, it seems possible that a situation may arise where agents are better at their jobs or know more about their jobs than humans.

I think it’s very difficult to manage agents if you can’t do the work yourself. I think managers still need to be experts in their field. This is a new skill set that many people have to learn, but managing agents is actually very similar to being a manager of people in the sense that you need to understand things like why an agent made this mistake. Did you misunderstand my instructions? Was my request poorly specified? I need the ability to debug it.

However, the long-term goal appears to be to reduce the size of the team. After all, if you had an agent do your work, you wouldn’t need an intern, right?

Ideally, I think the idea is that everyone can achieve more. I think everyone’s job has a certain amount of really boring parts. In my case, it’s a reply to an email. I think everyone has a part like this in their life…so my wish is that [the AI agents] If you actually do that, you’ll have all the amazing things anyone could ever want to make. [in their spare time].

What are you most looking forward to in the next six months?

I think the next most important thing is positivity. Last year, we were in a world of synchronous development. Nowadays, people are moving toward routine and like to automate responses to customer support tickets, for example. And I think the next step is for Claude to understand what you’re working on and set up some of these automations.