Anthropic's Claude Code Lead on Usage Limits, Transparency, and the 'Lean Harness'
Anthropic's head of product for Claude Code, Cat Wu, discusses the company's approach to usage limits, transparency, and model capabilities.

SAN FRANCISCO—As Anthropic's Claude Code continues to expand its reach across various surfaces, the company is facing growing demand for tokens and compute, as well as a rapidly evolving user base. Despite this, Anthropic doesn't have a long-term road map for Claude Code. However, according to Cat Wu, Anthropic's head of product for Claude Code, the company is betting that improvements in model capabilities and new signals from developers will render such a plan moot.
Wu's comments came during a 30-minute conversation with Ars, following Anthropic's second annual Code with Claude developer conference, held last week in a converted three-level car rental parking garage in downtown San Francisco. The event drew attendees eager to learn about the latest developments in Claude Code, with a keynote introducing new features for Managed Agents and announcing a compute deal with SpaceX. One of the key takeaways from the event was the doubling of usage limits for Claude Code users on the company's Pro and Max plans.
This move was a direct response to user frustration about a compute crunch, which had become particularly acute in recent weeks. According to Wu, the increased usage limits are a sign of Anthropic's commitment to supporting its users and helping them to get the most out of Claude Code. Wu also touched on the concept of the 'lean harness,' a approach that emphasizes simplicity and efficiency in the use of Claude Code.
While details on this approach are still emerging, Wu suggested that it represents a key aspect of Anthropic's strategy for Claude Code, one that prioritizes flexibility and adaptability in the face of rapidly changing user needs. As Anthropic continues to evolve Claude Code and respond to user demands, Wu emphasized the importance of transparency and communication. By working closely with developers and listening to their feedback, Anthropic aims to create a more responsive and effective platform for building with Claude Code.
The company's focus on model capabilities and developer feedback reflects its commitment to creating a platform that is both powerful and adaptable. As Wu noted, Anthropic is constantly learning from its users and iterating on Claude Code to better meet their needs.
Source: Ars Technica