Project Glasswing
-
techAnthropic's New AI Surfaces 27-Year-Old OpenBSD Security Flaw
Anthropic's Mythos AI finds over 23,000 vulnerabilities in open-source projects, including a 27-year-old OpenBSD bug and a 16-year-old FFmpeg flaw.
-
techAnthropic AI Uncovers 23,000 Flaws in Open Source Code, Including a 27-Year-Old Bug
Anthropic's Claude Mythos detected over 23,000 vulnerabilities across 1,000+ open-source projects, with 1,726 confirmed, including a critical bug in OpenBSD that evaded detection for nearly three decades.
-
techProject Glasswing: AI Finds 10,000 Flaws in Weeks, Outpacing Decades of Human Work
Anthropic's restricted AI model, Claude Mythos Preview, uncovered 10,000+ software vulnerabilities, including a 27-year-old bug in OpenBSD.
-
techAnthropic's Claude AI Uncovers 10,000+ Software Vulnerabilities in Six Weeks
Anthropic's Project Glasswing used Claude AI to find over 10,000 critical software flaws, including zero-days in major systems like OpenBSD and FFmpeg.
-
techAnthropic Pushes for Wider Release of Controversial Mythos AI Model
Anthropic expands access to Mythos AI, capable of finding zero-day vulnerabilities, despite White House opposition and security concerns.
-
techOpen-Source Project Reverse Engineers Anthropic's Most Dangerous AI
OpenMythos reconstructs the Claude Mythos architecture from public papers, potentially unlocking its advanced cybersecurity capabilities without Anthropic's approval.
-
techNSA Tests Anthropic's Mythos AI to Hunt Microsoft Security Flaws
The NSA is evaluating Anthropic's Mythos AI to find vulnerabilities in Microsoft products, signaling a major push for AI in US cybersecurity.
-
techAnthropic Withholds Its Most Powerful AI Over Safety Concerns
Anthropic has declined to release its most powerful AI model yet, citing a lack of adequate safeguards. The decision sets a new precedent for responsible AI development.