Recently, the AI company Anthropic officially launched its new programming assistant "Cowork," a tool that was almost entirely developed using its own AI model Claude. Within just one and a half weeks, the first version of Cowork was released, marking another important milestone in AI programming tools.
According to Business Insider, Cowork is designed to provide more friendly AI support for non-programming users. Users can easily authorize Cowork to access specific files on their computers and complete various tasks through simple natural language instructions, greatly reducing the barrier to programming. Boris Cherny, the head of the Claude Code project at Anthropic, revealed that the core code of Cowork was almost entirely generated by Claude. The team mainly focused on discussing architecture and product direction, while the implementation of specific features, bug fixes, and research work were completed in parallel using multiple instances of Claude.
Felix Rieseberg, the product manager, said that this innovative development approach allowed Cowork to complete its first version in an extremely short time. Especially during the holiday season, users began to frequently use Claude for non-programming tasks, which provided an opportunity for the creation of Cowork. However, he also admitted that Cowork is still in the Alpha stage and has some imperfections.
Currently, Cowork is only available to Mac users who subscribe to Claude Max. After its release, it quickly attracted widespread attention. Simon Willison, co-founder of Datasette, pointed out that the potential value of Claude Code has not been fully realized, and Cowork provides a more practical way to get started. Alexis Ohanian, co-founder of Reddit, called it a "major event."
However, Anthropic also reminded users to be aware of potential risks when using Cowork. Since Cowork may perform destructive actions such as deleting local files after obtaining file access permissions, users should ensure their instructions are clear and precise.
When talking about the topic of over-investment in AI, Dario Amodei, CEO of Anthropic, said that focusing on enterprise customers helped the company establish a more sustainable business model, making them more rational in investment decisions.
