Loading activity…
@aiatmeta
We're building the future of personal superintelligence to help people create, connect, and achieve more in life.
AI
Meta says Prometheus reached 85%-plus pretraining goodput and 90%-plus reinforcement-learning goodput during Muse Wonder training.
AI
Meta says it runs at practical speeds on a Mac or PC with one consumer GPU at 4-bit precision.
AI
TIME says Meta’s open-source models recognize speech in 1,600 languages, including 500 other AIs cannot currently transcribe.
AI
finkd says the model runs well on new Mac Studios, crediting work with Apple and John Ternus.
Technology
Meta AI Research says Active Memory retrieves relevant past details, while Active Focus helps Muse Wonder work through blockers.
AI
Meta says Muse Wonder launched through its Model API and Muse Code, while Full was available to Muse Code subscribers.
Science
Meta says mathematicians used Muse Spark’s Thinking Mode through the regular meta.ai chat interface, without a custom research scaffold.
AI
A user says Muse Spark generated code that helped find a 384-element counterexample to a 2024 group-theory conjecture.
AI
In a September 28 post, Meta's AI account listed 10 names from the previous six months, including Muse Image, Muse Video and Muse Glimmer.
AI
Meta says the technology turns Muse Realtime Voice into expressive, interactive avatars, starting with real-time conversations in Muse.
AI
AI at Meta account describes how AIRA₃ adapts to new tasks.
AI
Official account post describes AIRA₃ running isolated agents coordinated via forum and filesystem.
AI
Official Meta AI account describes its AIRA₃ entry and post-hoc model tests on the Kaggle leaderboard.
AI
Official Meta AI account shares update on entering AIRA₃ into NVIDIA-run Kaggle event.
AI
Meta AI executive announces public release of updated model with performance claims.
AI
Arena.ai announces Meta model entry for community prompts and voting.
AI
Retweet by Meta executive highlights release of updated model from Xiangning Chen.
AI
First streaming audio model from Meta Superintelligence Lab does real-time ASR with diarization.
AI
Meta Superintelligence Labs releases its first real-time audio perception model for streaming speech.
AI
Meta AI Research posted benchmarks and demos for the model's visual and robotics tasks.
