OpenAI's Dots vs. Meta's Muse: Inside the Race for Your Personal AI Agent

OpenAI and Meta have each put a personal AI agent in front of users, with OpenAI's Dots and Meta's Muse now competing for the same job: acting as the assistant that sits between you and everything you do online. A Wired reviewer who spent time with both products came away convinced that adoption is not an either-or choice, and that regular users will soon be running both. The two launches mark the point where personal agents stop being demos and start being products with real users attached.

The pitch behind both tools is essentially the same. An agent that remembers your preferences, takes actions on your behalf, and shows up across the apps you already use is a more ambitious product than a chatbot that answers one question at a time. OpenAI has spent the past two years pushing ChatGPT from a text box toward something that can browse, schedule and execute tasks; Meta has the advantage of owning the social graph, messaging apps and hardware that a persistent assistant needs to feel native rather than bolted on.

That difference in distribution shapes how the two products feel in practice. OpenAI's Dots leans on the company's model strength and its developer ecosystem, while Meta's Muse can draw on Facebook, Instagram and WhatsApp to learn who you talk to and how you spend your time. For Meta, an agent is a way to keep users inside its apps; for OpenAI, it is a way to become the layer that sits above them. Both companies have reason to move fast, because whoever owns the default assistant owns the entry point to a long list of paid services.

The reviewer's conclusion that people will use both is less a compliment to either product than a statement about how unfinished the category is. Agents still stumble on multi-step tasks, forget context between sessions, and require enough supervision that many users treat them as a second opinion rather than a delegate. That friction is why switching costs stay low right now, and why neither company can claim a lock on the market. Google, Anthropic and Apple are all building comparable assistants, so the field will get more crowded before it consolidates.

The immediate question is not which agent is smarter but which one gets permission to act. Agents that send messages, spend money or touch work accounts need trust that a chat window never required, and both OpenAI and Meta are still earning it. Watch for how quickly each company expands what its agent can do without asking, and how loudly users push back when it gets something wrong.

For now, the honest read is that personal agents have arrived as a mainstream product category without arriving as a solved one. Users who try Dots and Muse today are effectively beta testers for a business model both companies badly want to work. The next twelve months will show whether people keep two agents open out of curiosity or out of genuine need.