OpenAI Dots: a persistent assistant built around a project
OpenAI introduces Dots, where each dot has its own cloud computer and browser, is powered by GPT-6 Astra, and carries context across ChatGPT, Slack and Teams.
OpenAI Dots puts an ongoing project at the center of the assistant experience. Each dot has its own cloud computer and browser, connects to your apps, and carries context across ChatGPT, Slack and Teams. GPT-6 Astra powers it. OpenAI’s developer example is concrete: monitor customer feedback, scope fixes, build and test them, then bring back PRs with demonstration videos for review. For builders, I’d start with one recurring responsibility: turn incoming bug reports into reviewable patches. Define the repository, allowed changes, required tests and stopping conditions. Measure accepted patches, review time and rework over a week. One important boundary: unprompted “proactive research” uses read-only tools. It cannot send messages, modify connected apps or control a computer. Assigned work follows action rules and approvals. Dots began rolling out on September 29 to eligible Pro and Business Premium users, with an admin-enabled Enterprise beta. The test I care about: how much useful work reaches review without someone repeatedly rebuilding the context. https://openai.com/index/introducing-dots/ https://x.com/OpenAI/status/2104984504133918973/video/1