AI Built Its Own Sibling in 10 Days: Anthropic Claude Cowork Agent Launch
Anthropic has launched Cowork, a new AI agent for Claude Max subscribers that can manage files and automate tasks. It was built in just 10 days using Claude Code.
The era of self-evolving AI is officially here. On Monday, Anthropic released Cowork, a new AI agent for non-technical users. According to company insiders, the team built the entire feature in just 1.5 weeks, largely using their own developer tool, Claude Code. This rapid development cycle highlights a recursive loop where AI tools are now accelerating the creation of even more AI tools.
The Capabilities of Anthropic Claude Cowork Agent
Unlike traditional chatbots that merely suggest text, Cowork operates as a functional agent with direct access to your local files. By designating a specific folder on a macOS machine, users allow Claude to read, edit, and create documents within that sandbox. Anthropic's Boris Cherny noted that early users have deployed the underlying agent for tasks ranging from organizing messy download folders to controlling smart appliances and recovering lost photos.
The tool leverages an 'agentic loop,' meaning it doesn't just respond—it plans, executes steps in parallel, and asks for clarification when hit with a roadblock. It also integrates with Anthropic’s ecosystem of connectors, linking to services like Notion, Asana, and PayPal. Combined with browser automation via Chrome, Cowork can navigate websites and fill out forms, moving much closer to the functionality of a digital colleague.
Security Risks and Anthropic Claude Cowork Agent Pricing
Granting an AI access to your file system comes with inherent risks. Anthropic was unusually transparent in its warning, stating that Claude could take 'destructive actions,' such as deleting files, if instructions are misinterpreted. They also highlighted the danger of prompt injection attacks, where malicious code hidden in web content could hijack the agent's actions. Safety in real-world AI execution remains an 'active area of development.'
| Category | Details |
|---|---|
| Access Tier | Claude Max Subscribers |
| Subscription Fee | $100 - $200 per month |
| Current Platform | macOS Desktop App |
| Key Features | File manipulation, Browser automation, 3rd-party connectors |
Cowork is currently available as a research preview for Claude Max subscribers, a tier priced between $100 and $200 per month. While currently limited to macOS, Anthropic plans to bring the feature to Windows and add cross-device synchronization in future updates.
This content is AI-generated based on source articles. While we strive for accuracy, errors may occur. We recommend verifying with the original source.
Related Articles
Anthropic's AI cybersecurity model is reportedly available to the NSA and Commerce Department—but not to CISA, the agency responsible for defending US federal infrastructure. What that gap reveals.
Amazon has poured an additional $5 billion into Anthropic, bringing its total stake to $13 billion—with up to $20 billion more on the table. Here's what the deal really signals about the AI infrastructure race.
Amazon's fresh $5B investment in Anthropic brings its total to $13B. But the real story is a $100B AWS spending pledge and a bet on Amazon's own AI chips over Nvidia.
After two months of bitter conflict, Anthropic and the Trump administration may be thawing—thanks to a new cybersecurity AI model. What does it mean when principle meets political pressure?
Anthropic's AI cybersecurity model is reportedly available to the NSA and Commerce Department—but not to CISA, the agency responsible for defending US federal infrastructure. What that gap reveals.
Amazon has poured an additional $5 billion into Anthropic, bringing its total stake to $13 billion—with up to $20 billion more on the table. Here's what the deal really signals about the AI infrastructure race.
Amazon's fresh $5B investment in Anthropic brings its total to $13B. But the real story is a $100B AWS spending pledge and a bet on Amazon's own AI chips over Nvidia.
After two months of bitter conflict, Anthropic and the Trump administration may be thawing—thanks to a new cybersecurity AI model. What does it mean when principle meets political pressure?
Thoughts
Share your thoughts on this article
Sign in to join the conversation