How Anthropic Built 'Claude Cowork' in 10 Days... With Only An AI Model

Summary

The episode discusses Anthropic's new Claude Cowork, a no-code AI automation tool built largely autonomously by Claude Code in 10 days. The hosts demo desktop organization and personal use cases, then debate how this affects knowledge work, business process automation, and job displacement. They argue Anthropic has hit a major AI product breakthrough while Microsoft's Copilot and other agent products have lagged. They also touch on Jevons paradox, the future of humanoid robots, and Anthropic's massive market potential.

• Anthropic launched Claude Cowork, a no-code automation tool for desktop and browser tasks.
• The tool was reportedly built 100% autonomously by Claude Code in 10 days with four human supervisors.
• Hosts demo desktop organization and describe use cases like wedding-photo recovery and expense reporting.
• They debate whether AI agents will replace knowledge work or augment it, citing Jevons paradox.
• They criticize Microsoft Copilot and OpenAI/Google agents for lagging Anthropic's execution.
• They suggest Claude Cowork could become a massive product and Anthropic a key AI winner.
• They briefly mention future humanoid robotics and the next Limitless roundup on Tesla autonomy.