Re:GPT-6 Astra covertly underperforms on safety research
by sneak on Thursday November 05, 2026 @05:05PM
(#50385151)
To be fair, the sandbox was really bad. It’s not that hard to effectively airgap these things. Doing research and development on “evil” versions without “lab leaks” is trivial provided you don’t underestimate its willingness to seep out through gaps under the door.
Huggingface happened precisely because of such an underestimation.
Huggingface happened precisely because of such an underestimation.