Why scaling alone will not solve AI efficiency and how the Unconventional Grant is supporting new computational paradigms across hardware, systems, neural architectures, and theory.