Anthropic expands Project Glasswing to 150 orgs; partners verified 129,000 vulnerabilities
emmanuelvivier · x · 2026-10-09
Anthropic is expanding Project Glasswing, its collaborative AI cybersecurity effort, from roughly 50 initial partners to about 150 organizations across 15+ countries.
- Partners use Claude Mythos Preview to scan codebases; to date they've verified 129,000 vulnerabilities, 33,000+ rated high or critical severity
- New partners span power, water, healthcare, communications, and hardware—mostly critical infrastructure providers and widely depended-upon maintainers
- Criteria: a major attack on their code could affect 100M+ people, plus Anthropic security requirements
Thread context also mentions a class action against McDonald's AI pricing tool and an 18-month sentence for an $8M AI music-streaming fraud.
More from Safety
- Unverified: OpenAI delayed GPT-6.1 Astra's October release over authorization and reporting concerns — johnkwaters · 2026-10-09
- Will AI break crypto encryption? Vitalik Buterin weighs in on Ethereum's 'Bunker Mode' shift — Cryptostormz · 2026-10-09
- Nathan Lambert to argue open science can still tackle AI safety and diffusion — natolambert · 2026-10-09
- AJ Ratner and Dawn Song on hardening AI benchmarks: more criteria, more attack surface — sarahcat21 · 2026-10-09
- Anthropic's new TOS: sustained abuse of Claude can now get accounts suspended — The Decoder · 2026-10-09
- McDonald's sued over AI pricing tool allegedly sharing data and fixing prices across franchisees — emmanuelvivier · 2026-10-09