Meta AI Now Plans and Acts on Tasks via Muse Spark 1.1

Meta announced Muse Spark 1.1, a new AI model that powers the Meta AI app and meta.ai, designed to move beyond answering questions to actually executing tasks on behalf of users. The model is built to plan, work with apps, and follow through on multi-step actions from start to finish.

The new capabilities span several domains. In home renovation, users can tell Meta AI about their kitchen style, budget, and preferences, and it will search Marketplace for matching furniture and fixtures, then generate a mood board. For fitness, a user can ask for a half-marathon training plan, and Meta AI will create a week-by-week schedule, adjust for the user's availability, and send weekly plans every Monday morning. For event planning, it can find restaurants, check the user's calendar for conflicts, and suggest dinner options. Meta AI can also deliver personalized daily briefings—pulling calendar items, finding relevant updates (such as spotting double bookings or changed plans), and summarizing them at a user's preferred time. Once a task is set up, Meta AI continues executing it autonomously: weekly meal plans, alerts on sneaker drops, or trend updates arrive without needing a fresh request each time.

The model also synthesizes research, pulling from web sources, research papers, and content shared on Meta's apps, and can generate visual presentations (slides) from the research. Users can steer tasks in real time while Meta AI is working—shifting focus, changing tone, or removing sections—and the model adjusts on the spot. All outputs (training schedules, slide decks, mood boards) are stored in one place, allowing users to revisit, build on, or share them. Meta frames this as a step toward "personal superintelligence: an AI that knows your context, is there for you whenever you need it, and handles things so you don't have to."

Rollout begins today in select markets through the Meta AI app and meta.ai. Meta plans to expand to more countries and surfaces, including WhatsApp, over the coming weeks. The company emphasizes user choice—whether to seek quick answers, get contextual responses, shop for inspiration, or use Incognito chats for fully private conversations.

Meta is moving beyond the conversational AI assistant paradigm into autonomous task execution. Muse Spark 1.1 represents a fundamental shift: instead of responding to discrete queries, the model can now retain context across multiple steps, learn user preferences, and execute plans that unfold over time. This capability—planning a half-marathon schedule week by week, or automatically delivering Monday morning training updates—requires the model to understand user intent, anticipate needs, and maintain state across interactions without constant re-prompting.

Meta frames this as progress toward "personal superintelligence," an AI that knows its user's calendar, style, budget, and habits. The emphasis on real-time steering (adjusting focus mid-generation) and persistent storage (all outputs in one place) suggests Meta is designing for collaboration rather than pure automation—users remain in control while the AI handles the execution work. The rollout strategy (starting in select markets, expanding to WhatsApp in coming weeks) indicates a staged deployment, likely to manage infrastructure and gather feedback before broader release.

No comments yet. Be the first to share your thoughts!