Anthropic’s Cat Wu says that, in the future, AI will anticipate your needs before you even know what they are.

As the tech industry focuses exclusively on AI models, Anthropic is having a great year.

The company may go ahead of its main competitor, as it looks to raise tens of billions of dollars in a funding round that could put its value at $ 950 billion (OpenAI was worth $ 854 billion in its March round), and business customers are increasingly expressing their preference for Claude over ChatGPT. A recent report showed that Anthropic recently surpassed OpenAI among enterprise customers, doubling its market share by May 2025.

Cat Wu, Anthropic’s head of product for Claude Code and Cowork, has been a key figure in that success. Since joining the company in August 2024, Wu has helped shepherd Claude through a critical phase, elevating it from an information chatbot to a coding tool and beyond. Wu, who oversees the development of new features, is often paired with Boris Cherny, a key member of Anthropic’s technical staff and co-founder of Claude Code, leading the two to be seen as Anthropic’s “Batman and Robin.”

Wu sat down with me at the second annual Code and Claude conference in San Francisco, where he discussed how he thinks about the brand’s strategy, and how he hopes the experience of using Claude will change in the future.

This interview is edited for length and clarity.

When you look at the brand strategy, how effective is it for your peers or competitors? Do you think about that?

The main thing we’re designing is to stay exponential, so I think, across our team, we’re instilling in everyone the lesson that AI is going to keep getting better. For us, we just need to stay on this border. We don’t think about our competitors. I think when you think about your competitors, you end up like every two weeks, or like, a month behind as soon as you can. And so it’s usually not the best way to live on the border.

Anthropic released at least six models last year and has already released nearly as many this year. Do you expect this pace of development to continue?

Our hope is that it continues (laughs). I think models are still developing at a steady pace, so we should be able to keep sharing those with our users. I think the deployment might look a little different—like how we handled Glasswing, but as much as possible, we want this intelligence to benefit as many people as possible, and it should be handled in a very safe way, which is why we handled Glasswing. [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.]

He said in a previous interview that the future of work is basically a multi-agent workforce. It seems like that could eventually lead to a situation where agents do better, or know a job, better than a human.

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. It’s a new skill set that a lot of people are going to have to learn, but managing agents is actually a lot like being a people manager, in the sense that you have to understand, like, why did the agent make this mistake? Did you misinterpret my order? Was my request not clear enough? You must be able to debug it.

It appears that the long-term goal is to reduce the size of the group. Because if you have agents doing the work, you don’t need an intern, do you?

Ideally, I think the idea is that everyone can get more. I think that, in everyone’s work, there is always this percentage of it that is really boring. For me, it’s answering emails. I think everyone has this part of their life…So my hope is that it is [the AI agents] you actually do that, and then everyone has all these cool things that they’re going to want to build [in their spare time].

What are you guys most excited about in the next six months?

I think the next big thing is to do something. Last year we were in this world of parallel development. Currently, people are switching to habits, such as automating, for example, responses to customer support tickets. And I think the next step is that Claude understands what he’s working with, and he just sets up some of these defaults.