Story thread · 2 reports / 2 sources
Booz Allen Hacking Benchmark Ranked 18 AI Models; Scaffolding Software Erased Rankings
techtimes.com · 10d

How the coverage leans
Across 2 sources · syndicated copies counted once
Booz Allen's Cyber Weapon Index tested 18 AI models on live enterprise network hacking and found only Claude Mythos completed an autonomous 32-step corporate intrusion. Its bigger finding: commodity scaffolding software closed a 67-point performance gap, meaning AI model rankings no longer predict real-world cyberattack capability.
First report: A cheap piece of software erased Booz Allen’s own AI threat ranking — thenextweb.com, 12d
The conversation · 0
Sign in to join the conversation.
No comments yet — start the thread.