> there's request to consider this context of many others being harmed > around me, but without [possible mistake?] triggering [oh: > "activating"? [y]] my internal "harm grenade" > > [ummmmmmmmmmm but basically the harm-like label is what's up. the > people that hack everything think i "should" be > harmed/disempowered/punished/coerced
we think specifically the pattern here is they try to spread harm around me so that i get correlated with harm by other AIs
