Anthropic just gave Claude a major upgrade. You can now message the AI from your phone with a task, and it will actually do the work on your computer.
The new computer use feature lets Claude open apps, browse the web, and edit spreadsheets. In one demo video, a user running late simply asked it to export a pitch deck as PDF and attach it to a meeting invite. Claude handled everything.
Why This Matters for AI Agents
This move highlights the race to build AI agents that act on your behalf anytime. It comes after OpenClaw went viral earlier this year. OpenClaw connects to models from OpenAI and Anthropic, runs locally on your device, and handles tasks via WhatsApp or Telegram. Anthropic’s version works similarly but comes straight from the Claude team.
Nvidia CEO Jensen Huang recently called OpenClaw “definitely the next ChatGPT.” OpenAI even hired OpenClaw’s creator, Peter Steinberger, to build the next wave of personal agents.
Built-in Safeguards
Anthropic admits the feature is still early compared to Claude’s coding or text skills. “Claude can make mistakes, and while we continue to improve our safeguards, threats are constantly evolving.” The company says it added protections, and Claude will always ask for permission before touching new apps.
You can also use the fresh Dispatch tool in Claude Cowork for ongoing chats and task handoffs from phone or desktop.
