Perplexity Pushes AI Deeper Into Windows as Robot Curbs Tighten

Perplexity Pushes AI Deeper Into Windows as Robot Curbs Tighten
Perplexity AI is pushing its “Personal Computer” agent beyond the browser and into Windows desktops, promising to handle files, Microsoft applications and web research in a single workflow. The move highlights automation’s expanding reach just as Washington is taking a far more restrictive approach to some physical AI systems.

Perplexity says its Windows tool can open and edit local Word and Excel files, organize downloads, draw on web research and connected services, and continue a job begun on a phone. It is available for Windows 10 and 11 subscribers on Pro, Max and Enterprise plans, according to the company’s launch announcement.

The company’s case is that enterprise work is fragmented: documents live on local drives, collaboration occurs in Microsoft 365 and external information sits on the web. “Multi-model orchestration no longer stops at the browser,” Perplexity said, presenting the product as a way to turn scattered data into finished work. A separate announcement said the system can act directly on local files while giving users “remote coverage” to start work on a phone and finish it on a Windows machine.

Perplexity also stresses guardrails. It says users are alerted before sensitive actions such as sending emails or deleting files; actions are auditable, and files are created in a secure sandbox. Its broader developer push includes a remote MCP server intended to connect Perplexity with Claude Code, Cursor and VS Code using an API key.

That software-first expansion contrasts with a separate FCC ban on new foreign-made advanced mobile robots. Supporters see the measure as industrial protection: Standard Bots chief Evan Beard called it “one of the strongest technology-security actions in modern US history.”

But critics argue the restrictions could cut US researchers and startups off from inexpensive Chinese platforms before Western alternatives are ready. Robotics adviser Georg Stieler warned the policy “could slow US physical AI innovation,” while analyst Rueben Scriven said it may inhibit the domestic humanoid-robotics industry rather than strengthen it. Together, the developments show two competing instincts in AI policy: make digital agents more capable inside everyday work tools, while limiting foreign hardware seen as a security risk.

Continue reading https://foxvector.com/stories/019fc942-4c02-399b-72ac-3aa9be7c106b