“ElevenLabs predicts that "audio general intelligence," the ability to combine different audio models like narration and singing into a single continuous stream, will be possible "very soon."”