OpenAI’s Dots Are Always-On AI Agents—and Its Answer to Meta’s Muse (www.wired.com)

OpenAI has unveiled Dots, its new always-on AI agents, at the DevDay 2026 event, marking a significant advancement in personal AI technology. Powered by the GPT-6 Astra model, Dots function as proactive assistants that manage tasks on behalf of users, constantly gathering context from connected apps and learning user preferences over time. For instance, in a demo, a Dot efficiently planned dinner by suggesting delivery options during a busy schedule, showcasing their ability to handle multi-step projects beyond the capabilities of traditional chatbots. The introduction of Dots is particularly notable as it competes directly with Meta’s recent Muse agent, which has gained popularity among early adopters. OpenAI's strategy includes a phased rollout, starting with subscribers of its $100 monthly Pro tier, where users can control multiple Dots, eventually allowing for simultaneous management of several agents. Furthermore, Dots come equipped with safeguards, requiring user consent for sensitive actions and offering Custom Rules for setting boundaries. This innovation positions OpenAI to redefine user interactions with AI, fostering a new era of personalized digital assistance.