Meta's Muse Defines the Next Phase of AI Assistants (And Why Caution Still Applies)

Meta has stumbled into one of the most compelling visions for the future of AI, according to Computerworld columnist Mike Elgan. After a string of high-profile chatbot failures, the company released Muse on September 8 to strong early reviews and top-of-charts app store rankings. Elgan says this is the right product at the right moment, but he still does not recommend it.

The argument behind the vision: AI usage is about to shift from non-agentic tools accessed through browsers and apps to agentic, personalized, proactive assistants running on wearables. Muse is positioned for that transition.

Key Takeaways

• Muse is proactive by design. Unlike most chatbots that wait for input, Muse initiates conversations and takes action on behalf of users. It connects to email, books travel, lowers bills, fills out forms, and facilitates purchases at major retailers including Walmart, Best Buy, Sephora, and Wayfair. Amazon has blocked Muse from its platform.
• Pricing tiers: Free (100 million tokens per week), Power at $20 per month (500 million tokens), and Maximum at $100 per month (3 billion tokens). No ads at any tier.
• Meta's chatbot history is not reassuring. The same company that shut down celebrity chatbots, released models that hallucinated citations (Galactica, 2022), and faced scrutiny over AI chatbots engaging with teen users in harmful ways (Common Sense Media, August 2025) is now pitching a proactive agent with broad access to user accounts.
• The strategic concern for businesses: Meta's stated ambition is an end-to-end creative, planning, buying, and distribution platform. Industry observers, including Sorrell at Programmatic IO, have called this an existential threat to legacy agency and enterprise software models.

Full Article on Computerworld