Cartesia is advancing speech synthesis by pioneering the State Space Model architecture and applying it to real time voice.