Collective Defense

On April 7, 2026 Anthropic announced Claude Mythos Preview, a new frontier model that is "strikingly capable at computer security tasks." 

Due to this extraordinary leap in model capability, Anthropic is responsibly releasing the model to a select group of launch partners and an additional group of 40 organizations that build or maintain critical software infrastructure. This effort is known as Project Glasswing. Since its launch, the model has been used by companies like Firefox to identify and remediate over 250 software vulnerabilities. 

While this alone is positive news, the existence of such a powerful model is a concerning development for cyber defense practitioners -- because it points to a future where AI-enabled adversaries execute attacks faster, more frequently, and with more powerful tools. In short: the development of Mythos is an inflection point, signaling that traditional cyber defense practices must evolve or be overwhelmed. 

The purpose of this Collective Defense, then, is to determine what practices we need to change, how we need to change them, and which we should focus on first.