Anthropic's super-scary bug hunting model Mythos is shaping up to be a nothingburger
Anthropic's Mythos AI model, designed to detect vulnerabilities, was accessed by unauthorized users through a third-party vendor. The breach highlights the challenges of keeping code secure, especially when working with contractors and vendors.
Anthropic's Mythos AI model was accessed by unauthorized users through a third-party vendor environment. The breach occurred despite Anthropic's controlled release of the model to select organizations under Project Glasswing. A group of users gained access to Mythos by making an educated guess about its online location, based on information revealed in the Mercor data breach. The unauthorized users reportedly belong to a private Discord channel and have been 'playing around' with the model. The incident highlights the challenges of keeping code secure, particularly when working with contractors and vendors. Anthropic is investigating the breach, but so far, there's no evidence that its systems were affected.
This content was automatically generated and/or translated by AI. It may contain inaccuracies. Please refer to the original sources for verification.