Muse Glimmer: 30B-parameter model optimized for always-on local agent workflows Today • research.meta.ai
Show HN: Needle2: 14MB agentic LLM for phones, wearables, smart home and robots Today • cactuscompute.com