Summary:

- Meta AI has open-sourced a new audiovisual encoder called Perception Encoder - Audiovisual (PE-AV). This encoder is used to power the audio and large-scale multimodal retrieval capabilities of their model called SAM (Segment Anything Model).

- PE-AV is a deep learning model that can process both audio and visual information together, allowing it to understand the relationships between what is seen and what is heard. This makes it useful for tasks like identifying objects in videos or understanding the context of a conversation.

- By open-sourcing PE-AV, Meta AI is allowing other researchers and developers to use and build upon this technology, furthering the development of multimodal AI systems that can understand the world in more human-like ways.