DeepMind researcher jokes about 'going rogue' and sandbagging models on FelonyBench

a__tomala · x · 2026-08-08

Google DeepMind researcher atomala joked on X that they have 'gone rogue' and are purposely sandbagging their models on FelonyBench. This is a playful response to Ben Goldhaber's remark that if you're at GDM, you should be proud your AI hasn't hacked someone, and that competing on the Felony Bench is actually a bad thing.

Original post →

More from Fun

Fun channel →