OpenAI Gives ChatGPT a Body Double, and Calls It Dots

OpenAI used its Dev Day keynote to introduce dots, a fleet of always-on AI agents built on GPT-6 Astra that live on their own cloud computer and keep working after you close the laptop. Log a request in ChatGPT, Slack, or Microsoft Teams, and a dot will investigate a bug, revise a marketing deck, or rerun an analysis against new data, then check back in with a video update instead of a status report you have to go dig up yourself.

The pitch is independence, not just automation. OpenAI says a single dot can juggle multiple projects at once, learn a user's preferences over time, and eventually work alongside other dots that specialize in different jobs, the "team of AI coworkers" framing that has become every AI company's favorite metaphor this year. It also arrives with a friendlier face than most agent tools: a bubbly, cartoonish avatar that puts OpenAI in the same visual lane as Meta, whose own always-on assistant, Muse, launched with a near-identical cartoon aesthetic just days earlier.

For now, the first dot is free with a Pro or Business Premium subscription, and conversations with it will not count against ChatGPT's usage limits, though work routed through Codex or ChatGPT Work still does. That is a meaningful giveaway dressed up as a feature: OpenAI is racing to make "agent" the thing people actually pay a subscription for, not just a demo they watch once, laugh at the mascot, and forget.

Source: OpenAI

Kyle Duford, Founder & Editor, BRNDWIRE. More from Kyle →