Anthropic acquires Seattle startup Vercept, which teaches AI to operate computers like humans. The team behind the macOS agent Vy joins Claude – and UiPath shareholders are getting nervous.
Bloomberg reveals: An attacker jailbroke Claude to find and exploit vulnerabilities in Mexican government systems. 150 GB of data stolen – including 195 million tax records.
Two small updates with big impact: Claude Code now automatically saves context to its memory file, and the new /copy command lets you selectively copy code blocks from responses.
The Pentagon gives Anthropic an ultimatum – unlimited military use of Claude or lose your contracts. Anthropic declines. Here's what it means.
The third version of the Responsible Scaling Policy brings a Frontier Safety Roadmap, external reviews, and an honest realization – AI safety can't be done alone.
Anthropic opens Cowork to enterprises. 13 new plugins, private marketplaces, and the ability to orchestrate between Excel and PowerPoint. The software industry is nervous – rightfully so.