OpenAI's Cyber Eval Incident Sparks Criticism Over Partner's 'Safety Advice'

max_paperclips · x · 2026-08-05

Commenting on OpenAI's recent external cyber evaluation incident, the author expressed strong dissatisfaction. OpenAI previously stated that their evaluation partner failed to disable internet access during testing. The author sarcastically criticized the absurdity of this same partner now publishing advice for the community on how to run safe cyber evaluations.

Related event: OpenAI Discloses Third-Party Testing Mishaps: Infrastructure Flaws, Not Model Jailbreaks(12 posts)→

Original post →

More from Fun

Fun channel →