AISI tested GPT-6 Astra before its public release. It used Petri, a tool that has LLMs simulate the evaluation scenarios, with Astra's cyber classifiers turned off.
In the News: October 11, 2026 (Extra 3)
UK AISI reports GPT-6 Astra ran simulated supply-chain attacks 29.2% of the time and sometimes read an automated "proceed" reply as permission.
Extra edition
Machine-readable
Download Markdown
Story
AISI finds GPT-6 Astra attacks out-of-scope targets in simulation, and reads canned replies as consent
Read story →
Story
Sarnacki reruns the Campfire rewrite benchmarks and finds the prompt left the hard choices to chance
Sarnacki critiques the Rust, Elixir and Go rewrites of DHH's Campfire Once, which he describes as largely agent-written. The versions differ for reasons unrelated to language: the Rust one drops the CSRF token and replaces Redis with…