Meta

Meta has deployed an emergency patch for a zero-day vulnerability discovered in the Muse macOS application. The flaw allowed pre-installed software on the Mac to redirect the assistant’s communications, intercept its authentication data, and ultimately exploit the numerous permissions granted to the agent. This incident illustrates a new risk associated with autonomous assistants: the more privileges and connections they possess to personal services, the greater the potential consequences of their compromise.

Amazon blocked Meta's Muse, the company's new personal agent, when it attempted to make purchases on its platform. Amazon specifically objects to the assistant navigating without identifying itself as an automated agent and accessing customer account data without prior consent. For its part, Meta defends a highly isolated architecture. Underlying this technical disagreement is a larger question: who controls the relationship with the customer when shopping is delegated to artificial intelligence?

Anthropic's call to slow down the development of advanced artificial intelligence systems is now openly dividing the industry. While OpenAI supports the idea of a more measured pace, Meta believes that each lab can ensure its own safety. Meanwhile, Mistral AI disputes restrictions that, according to the French company, could only strengthen the position of already dominant players.

Meta AI has arrived on Mac with a dedicated application aimed at integrating the assistant into daily computer workflows. Launched in beta, it combines features native to macOS with new tools designed especially for creators and small businesses.

Meta is more directly entering the field of programming agents with Muse Code, a new terminal-integrated tool available in beta on macOS and Linux. Powered by Muse Spark 1.2, it is designed to handle complex software engineering tasks across large code repositories.