OpenAI just gave ChatGPT a background employee, and it looks a lot like Meta's.
At its DevDay keynote on Tuesday, OpenAI unveiled Dots, an always-on AI assistant built on the company's GPT-6 Astra model. Dots run on their own cloud computer, giving them a web browser and access to more than 4,000 supported apps to work through tasks in the background while learning a user's preferences. People talk to a Dot through a text-message-style chat, the same format Meta uses for Muse, or by placing a voice call to it inside ChatGPT on web, desktop, or mobile. Like Muse, each Dot shows up as a cute, customizable avatar.
OpenAI is matching Meta feature-for-feature here rather than staking out new ground. Meta shipped Muse's persistent, personality-driven assistant first; OpenAI's version arrives with a similar avatar concept and the same background-agent pitch, right down to the claim that it can do nearly anything across your apps. That looks less like a technology leap and more like a competitive necessity: OpenAI cannot let Meta define what a personal AI agent looks like.
Whether Dots actually finishes your errands or just becomes one more notification to manage is the part no keynote demo can answer.