Anthropic's Glasswing Unearths 27-Year-Old OpenBSD Flaw: AI Redefines Zero-Day Hunting
Anthropic just unleashed an AI that sniffed out ancient bugs in the world's toughest OSes, no humans needed. Project Glasswing isn't hype—it's a wake-up call for every dev touching code.
⚡ Key Takeaways
- Claude Mythos Preview autonomously found zero-days in OpenBSD (27 years old), FFmpeg, and Linux kernel—outpacing humans and tools. 𝕏
- Security emerges as byproduct of advanced agentic coding; benchmarks show massive leaps in SWE-bench, CyberGym. 𝕏
- Anthropic's $100M+ push aids open source, but dual-use risk looms—AI defenders today could arm attackers tomorrow. 𝕏
Worth sharing?
Get the best Open Source stories of the week in your inbox — no noise, no spam.
Originally reported by Dev.to