The news of November 6thMeta Chief Executive Mark Zuckerberg published a string on Threads social platform todayDECLARE PERSONAL AI ASSISTANT Muse The official live-on-line voice dialogue and a performing virtual image function。
The above functions are first sent to U.S. users, covering mobile, web-based and Mac applications. A user clicks on the top right telephone icon to talk to Muse like a phone and can interrupt it at any time during its response. 1AI Attach the following relevant screenshot:
In terms of technical support, the update is mainly supported by the Muse Realtime Voice voice model, which generates both “what to say” and “what to say” information, and the Muse Realtime Avatar virtual image model, which produces a real-time expression, mouth and body action based on the same data stream to synchronize sound with the image。
Meta states that the system is current in output of 448 x 768 resolution, 25 frames per second vertical screen video, with a delay of approximately 870 milliseconds from the time the user finished speaking to the first audio video response。
In the interactive mode, the user can choose to speak only, display a virtual image next to the chat interface, or switch to a full-screen face-to-face mode. Muse's default image is a cartoon character nicked “Jolly”, which can also be replaced by a person, animal or other character; and a voice that provides hundreds of styles, including “New York hospitality” and “piracy winds”。
Meta positions this capability as a more natural personal intelligence interaction: users can arrange assignments orally while walking, with their hands busy, and Muse continues to handle it backstage; it can also be used for brainstorms, travel planning, interviews or speech exercises. Meta also shows how to dance and share in Instagram, Facebook or whatsApp with Muse virtual images。