A two-year research effort with teachers found that classroom uptake of generative AI remains minimal even among educators who are open to integrating it. The study describes a persistent challenge: teachers struggle to identify clear or universal instructional use cases for broad AI integration. The reporting frames the issue as less about training teachers to use a specific tool and more about curriculum and pedagogy that can withstand rapid technology change. The article argues that when instructional goals depend on fast-evolving AI systems, classrooms face a mismatch between curricular stability and tool volatility. Instead, the researchers point toward teaching enduring computational ideas—foundational concepts that remain relevant even as specific platforms change. For higher education and K–12 leaders designing AI literacy initiatives, the takeaway is concrete: investment in professional learning and curriculum coherence may need to focus on durable learning objectives, not prompt-and-tool workshops.