Anthropic's Claude gets remote control

· Source: The Rundown AI · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Robotics & Autonomous Systems, Emerging Technologies & Innovation · Depth: Intermediate, medium

Summary

Anthropic has released a research preview enabling Claude to directly control macOS desktops, allowing users to assign tasks from a mobile device via Dispatch and have Claude execute them by clicking, typing, and navigating applications. This functionality is currently available to macOS users on Pro or Max plans through Cowork and Claude Code, with a Windows version planned. This development follows Anthropic's acquisition of computer use startup Vercept in February, marking their first product launch in four weeks. Concurrently, Luma AI introduced Uni-1, an image model that processes text and visuals through a unified pipeline, demonstrating real-world understanding and outperforming competitors in human preference rankings for style and editing, with an API price of ~$0.09 per image at 2K resolution.

Key takeaway

For Computer Vision Engineers evaluating new AI tools, Anthropic's Claude with remote desktop control and Luma AI's Uni-1 image model represent significant advancements. You should investigate Claude's new capabilities for automating macOS tasks and consider Uni-1's unified architecture for creative applications, especially given its competitive pricing and performance in image generation and editing.

Key insights

AI models are rapidly evolving into autonomous agents capable of direct computer control and advanced creative generation.

Principles

Method

Anthropic's system prioritizes direct app integrations and browser access before resorting to screen control for task execution. Luma AI's Uni-1 processes text and images via a single pipeline.

In practice

Topics

Best for: Computer Vision Engineer, AI Engineer, Machine Learning Engineer, AI Product Manager

Related on AIssential

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by The Rundown AI.