– H1: Anthropic’s Claude can now use your computer like a human: Will it replace OpenClaw?

– H2: Introduction
• On March 23, 2026, Anthropic announced an update to Claude, enabling it to control a user’s computer to complete tasks.
• This development is significant as it highlights the advancements in AI capabilities for autonomous systems, especially in a landscape where open-source frameworks like OpenClaw are gaining traction.

– H2: Key Developments
• Claude’s new functionalities include executing tasks such as locating and sending files directly from a user’s hard drive.
• Users can initiate tasks through prompts on their phones, allowing Claude to find and utilize tools via integrations with apps like Google Calendar or Slack.
• The update is available in research preview for MacOS users subscribed to Claude Pro and Claude Max plans.
• Each app used by Claude requires user approval if a connector isn’t available, promoting a layer of consent.

– H2: How It Works / Details / Explanation
• Claude can perform automated tasks by mimicking human user behavior, such as scrolling, clicking, typing, and cursor movement.
• The new Dispatch feature in Cowork allows users to delegate tasks from their phones and return to completed work on their computers.
• Users can also initiate live coding sessions with Claude to generate code or execute tests.
• This update features functionality built on groundwork laid by Anthropic’s acquisition of Vercept, which specializes in complex agentic tasks.

– H2: Context / Background
• Anthropic first previewed computer-use capabilities in 2024 and has since been focused on enhancing AI interaction with desktop applications.
• The update comes against the backdrop of increasing concerns over job automation due to the rise of AI agents.
• OpenClaw’s rise as an open-source competitor is marked, with its ability to spawn autonomous agents capable of managing complex tasks and accessing local file systems.

– H2: Implications / Impact
• The launch is expected to reshape how tasks are completed in workplace environments, with potential disruption to traditional SaaS business models.
• Claude’s capabilities could streamline workflows and enhance productivity for users by offering a more integrated and reliable AI assistant compared to OpenClaw’s makeshift solutions.
• Security implications arise, as Claude’s permissions model necessitates user oversight, contrasting with the more permissive access requirements of OpenClaw.

– H2: Challenges / Risks / Criticism (if applicable)
• Concerns about security risks, such as prompt injection attacks, and user data privacy are underscored by Anthropic.
• The update’s reliance on AI reasoning is considered a limitation that may lead to errors in task execution, necessitating a careful approach to its deployment.
• OpenClaw’s open-source model raises fears regarding rogue agents, prompting the need for safety measures.

– H2: Future Outlook
• Continued improvements are anticipated as Claude’s computer-use feature expands to support Windows devices.
• The competitive landscape may evolve further as other companies, such as Nvidia with its NemoClaw toolkit, respond to the increasing demand for sophisticated AI automation tools.
• Observations of user adoption and feedback will shape future enhancements and adaptations of Claude’s capabilities.

– H2: Conclusion
• Anthropic’s advancements with Claude reflect a pivotal moment in AI development, posing both opportunities for enhanced productivity and challenges in security and ethical implementation.