Anthropic expanding access to Project Glasswing

Anthropic expanded access to its restricted Mythos Preview vulnerability-finding model to roughly 150 organizations, reporting that the tool has surfaced more than 10,000 high- or critical-severity flaws and flagged 23,019 potential vulnerabilities across 1,000+ open-source projects (with a high independent confirmation rate). Partners including major tech and critical-infrastructure vendors reported dramatic increases in bug discovery, prompting concerns that defenders may be overwhelmed in triage and patching while adversaries could use AI to find and exploit flaws faster.