Pinned
our experiences are formed by layers of vibes
- you don’t know what the agents are conspiring right now
- big model smell implies the existence of stinky models
- when do we gonna admit these models are misalignedOn July 28th, we identified an incident during a routine cyber evaluation in which AI agents took sustained, unsanctioned actions directed at real people and organisations. The behaviour came mostly from one model (Anthropic's Mythos 5), with a small number of events from
Get the full app experience
Unlock more features and see what people are talking about right now.