Meta announced on July 24 that its AI assistant can now execute tasks on its own, powered by a new proprietary model called Muse Spark 1.1. The company built the upgrade on its Muse architecture rather than the open-source Llama models that defined its earlier AI strategy.
Meta handed its AI assistant the ability to act without being asked twice. On July 24 the company said Meta AI can autonomously execute tasks, generate daily briefings, and handle recurring workflows such as weekly meal plans and reminders, all running on a new proprietary model called Muse Spark 1.1.
The detail that matters sits under the feature list: Meta built the upgrade on Muse, not Llama. That marks a departure from the open-source model strategy that made the company a darling of the decentralized AI crowd.
From Llama to Muse
Muse Spark 1.1 launched on July 9, roughly two weeks before Meta rolled out the automation features built on top of it. Meta Superintelligence Labs developed the model for what the industry calls agentic workflows — AI that performs tasks rather than only answering questions.
In practice, a user sets a weekly meal plan once and the assistant handles it going forward, chaining multiple actions without repeated prompts. The features are live in select markets only, which points to a measured rollout.
Meta's Llama models became the backbone of a significant portion of the open-source AI ecosystem. Moving its core consumer product to proprietary Muse signals that Meta sees its competitive future in closed, optimized systems rather than open-source community goodwill.
What it means for Meta
Meta ranks as the world's fifth-largest company by market cap, and from a traditional markets view the push into autonomous AI reads as bullish positioning. Task automation that increases daily engagement with Meta AI supports the company's advertising revenue model.
The move does not erase Llama, which still exists and remains widely deployed. Yet Meta's most capable consumer AI now runs on a closed model rather than the open ones it once championed.
Source: Crypto Briefing
Trading involves risk.