AISLE found 500+ zero-days across 100+ codebases in one year, ~1% of 2026's open-source CVEs
brhydon · x · 2026-10-17
AI security research lab AISLE (stanislavfort's team) reports discovering 500+ CVE-credited zero-day vulnerabilities across 100+ codebases in its first year — roughly 1% of all open-source CVEs globally in 2026. On hardened codebases like curl, the team repeatedly matched or outperformed frontier AI labs at zero-day discovery.
Reposter Ian Goodfellow notes the lesson: the system built around an LLM matters enormously — even small, cheap, fast models can be boosted to do valuable security work.
More from Models
- Model 6.1 Sol Solves a 6D Rubik's Cube, Far Beyond Human Intuition — flowersslop · 2026-10-17
- 10 open-source repos for training small models to do your agent's job — victor_explore · 2026-10-17
- Claude Sonnet 4.5 will be fully removed from AWS Bedrock by April 2027 — repligate · 2026-10-17
- DeepSeek V4.1 Flash talks like a human while coding: "I love this", "Disaster", "Ugh" — teortaxesTex · 2026-10-17
- Theory-of-mind test: Claude spots AI-written text, ChatGPT and Gemini fail — Aizkmusic · 2026-10-17
- Rumor: Core automation to release a model literally called 'AGI' on Nov 10 — imjustnewatai · 2026-10-17