Meta Platforms recently unveiled the first large language model (LLM) to come out of its superintelligence laboratory. The company then went on announce that they expect Muse Spark to improve AI features available across its social media apps.
Muse Spark is being positioned as the company’s most powerful model to date. It is initially being used on the Meta AI app and website, with plans to introduce it for WhatsApp, Instagram, Facebook, Messenger and on its glasses in the coming weeks.
It is also offering preview access to the model to select partners through an API.
Announcing the move, the social media giant highlighted the LLM developed by the Meta Superintelligence Labs is “small and fast by design, yet capable enough to reason through complex questions in science, maths and health”.
Muse Spark is an “early data point on our trajectory”, with larger models in development which will build on its initial release.
Its ultimate aim is to provide “personal superintelligence”, which it defines as “an assistant that can help anyone, anywhere with the things that matter most to them”.
The company asserted the model provides “multimodal perception”, improving Meta AI’s ability to deliver more detailed responses, using the examples of health queries including graphs and other visuals, product comparisons and recommending style choices.
It expects users of Meta AI on its social media apps to be able to gather “richer, more visual results, with Reels, photos and posts woven directly” into responses, noting it would credit content creators for the information used.
Meta Platforms also pointed to the model’s ability to undertake “visual coding”, allowing the creation of custom websites or mini-games quickly from a prompt.
The company is pumping huge sums into its AI efforts as it pursues the superintelligence goals.
Previously Meta Platforms’ AI systems on its apps used its open-source based Llama series of models.
Source: Mobile World Live
Image Credit: Meta Platforms
