The age of AI agents is here, and Anthropic's Claude Computer Use is one of the first to enable AI to browse the web, click buttons, and type text independently.

At Anthropic, while our models are intelligent, we must communicate how they can be useful and connect with users, not just highlight their intelligence. We need to design AI tools like Claude to be genuinely useful for people discovering large language models for the first time, providing them with concrete, practical applications in their daily lives.

Anthropic has shifted focus towards coding applications for Claude, moving away from its initial chatbot functionalities.

With the introduction of Claude 4 models, Claude Code has become better at following instructions and executing tasks accurately on the first attempt, enhancing the coding experience.

To use Claude's computer capabilities, developers must run it in a virtual machine or container and obtain an Anthropic API key, after which they can interact with Claude through a dedicated browser interface.

In the past six months, Claude's developer market share surged from 5% to 25%, indicating a significant shift in preference among developers for alternative models over OpenAI's offerings.

I write code that is more compatible with Claude, which highlights the advanced capabilities of AI compared to human cognition.

The findings in our research indicate that AI models, like Claude, exhibit complex thought processes that resemble human-like planning.

The new models, Claude Opus 4 and Claude Sonnet 4, have impressed us, particularly with their capabilities in agentic coding, which we are excited to integrate into Cursor.