Meta launches Muse Voice Transcribe, MSL's first real-time audio perception model, with streaming automatic speech recognition, trained with 70+ languages (Meta AI Research) | TickerVault