Anthropic scrambles to contain self-inflicted leak of valuable AI code

Fonte: New York Post

Anthropic has been scrambling to contain a self-inflicted mess after it accidentally leaked a treasure trove of internal code that powers one of its most valuable artificial intelligence tools, according to reports. The code serves as instructions for Claude Code, an AI agent app that developers and businesses pay top dollar to use to program and build applications of their own. Anthropic’s competitors and hoards of startups and developers now have the goods to essentially clone features of Claude Code — a shortcut to reverse-engineering them, the Wall Street Journal noted.