China just dropped a new frontier model... and nobody can explain how
Save just the original audio for offline listening. No transcription or AI generation is needed.