332
Audio & Video Production332
Automation & Workflow223
Software Development249
Marketing & Growth204
AI Infrastructure & MLOps154
Writing & Content Creation204
Data & Analytics132
Customer Support134
Design & Creative155
Sales & Outreach123
Operations & Admin97
Photography & Imaging143
Voice & Speech132
Education & Learning122
Meta launched Muse Spark, its first new AI model since major spending on AI staff and data centers, aimed at Meta AI across Instagram, Facebook, and Threads.
In short: Meta has released Muse Spark, a new AI model designed for use inside its social media apps and its Meta AI assistant.
Meta unveiled Muse Spark on Wednesday. It is the company’s first AI model launch since CEO Mark Zuckerberg approved multibillion-dollar spending on AI researchers and large computer facilities called data centers (warehouses full of servers that run online services).
Meta says Muse Spark will power a “smarter and faster” version of Meta AI, the assistant that answers questions inside Meta products. The company says responses will be more personal and more visual, using content people share on Instagram, Facebook, and Threads. Meta also said answers will include Reels, photos, and posts, and will point back to the original creators.
Meta said the model can also help with health topics. It said it worked with more than 1,000 doctors to train the system to give more detailed replies on areas like nutrition and exercise. Meta AI will also add a “shopping mode” to help users compare prices.
Muse Spark is different from Meta’s earlier Llama models because it is not openly shared for developers to download and change. Instead, Meta describes it as a smaller, closed model, and plans a limited “private preview” with select partners.
For everyday users, this could change what “asking the app” looks like. Instead of getting a plain text answer, Meta wants the assistant to act more like a guide that pulls up relevant posts and videos from your social feeds (like a librarian who also hands you pages from your own scrapbook). It also shows Meta is trying to prove to investors that its recent AI spending will turn into features people actually use.
Source: Financial Times