Urgent.News

What's breaking now, across thousands of outlets.

AI

Meta introduces specialized versions of its existing models, trained using Muse Wonder, bringing Active Memory and Active Focus to all Muse and Meta AI users (Frederic Lardinois/The New Stack)

Meta trained new models using its frontier Muse Wonder model and is shipping them to all Muse and Meta AI users today, while keeping Wonder itself for harder subscriber tasks.

Meta is shipping specialized versions of its existing models to all Muse and Meta AI users, which were trained using its frontier Muse Wonder model. According to Techmeme and The New Stack, these new models bring Active Memory and Active Focus capabilities to users, while the Muse Wonder model itself remains available for more difficult tasks for subscribers.

The company is also rolling out a feature called “spaces,” which allows users to build customized, app-like experiences such as music trackers, school dashboards, or shopping hubs within the Muse AI assistant, according to The Verge. Users can create these spaces by providing descriptions and can interact with them by adding notes, making edits, or checking items off. If connected to a relevant app, these spaces can update information like pricing and dates in the background.

Additionally, Android Authority reports that Muse AI users can now engage in real-time voice conversations with the assistant using Muse Realtime Voice and Muse Realtime Avatar technology. This feature, which allows users to interrupt the AI and choose from hundreds of voices, is currently limited to the Muse app on mobile, web, and Mac in the US. Separately, Mexico News Daily reports that Meta has selected Mexico as the Latin American launch country for its new AI agent.

Brief written by urgent.news from Techmeme, Mexico News Daily, The Verge, The New Stack, Android Authority — 5 reports on this story. Machine-written — may contain errors; check the original before relying on it.

This story

This is one outlet's version. Read the fullest account.

Read the original at thenewstack.io →

More in AI

I Built Select: An Open-Source Platform Where AI Agents Earn Their Autonomy

AI agents are getting more powerful, but there's one problem I've been thinking about: How do you know when an AI agent deserves to work independently?

  • Select introduces open-source platform for AI agents' autonomy.
  • Agents earn autonomy based on task performance, not from start.
  • Deterministic verification separates agent's work from correctness evaluation.

As an AI engineer, please learn:

  • 10-week roadmap teaches LLM inference in production
  • Program includes 50 sessions for OpenAI-compatible service
  • Participants need Python, transformers, command line knowledge

More from Saturday 7 November →