Technology

OpenAI's New ChatGPT Helper Keeps Working After You Close the Chat

Martin HollowayPublished 2month ago2 min readBased on 7 sources
Reading level
OpenAI's New ChatGPT Helper Keeps Working After You Close the Chat
source:openai.com

OpenAI has announced Dots, a new ChatGPT helper that keeps working for you even when you are not chatting. The company shared the news at its annual DevDay conference in San Francisco, California. CEO Sam Altman showed the system onstage. Dots is powered by GPT-6 Astra, according to reporting on the launch The Verge.

A Dot is an always-on helper in ChatGPT that takes on ongoing jobs and keeps making progress between chats, as described in OpenAI's support documentation OpenAI Help. Like an office assistant who keeps working through your to-do list after you leave the room, it does not reset when the chat ends. It remembers where it left off and continues in the background.

Each Dots agent has its own cloud computer, which is a private workspace online, and learns from feedback over time OpenAI.

Altman called the agent "remarkably capable, always-on" AP. The helpers look like colorful blobs with eyes that you can personalize. OpenAI has limited Dots to users on its $20/month Pro plan and up. ChatGPT has 1.2 billion users The Verge. Altman said Dots are starting as a premium product because they use a lot of computing power. He also said the agent has been safety tested Reuters.

OpenAI is working with Microsoft to add Dots to its Agent 365 platform. Special versions trained for accounting and legal work are planned to launch soon.

DevDay 2026 had more than 20 announcements, including GPT-6 Astra OpenAI. OpenAI also introduced GPT-6.1 Sol as a major upgrade to GPT-6 Sol. Dots was presented as the user-facing packaging for that model work, described in one report as a bubbly agentic avatar TechCrunch. Meta's Muse offers a free dedicated virtual machine, a private computer online, that holds both the agent and the user's data.

The broader context here is cost and control. Helpers that keep working use computing power even when you are away. Someone must pay for that, and someone must set rules for what data the helper can use. OpenAI's answer for now is paid access plus a separate cloud computer per helper. Meta's answer is a free online computer on its own system.

In my view, the practical questions will decide use. How well memory lasts over weeks of feedback. How well Dots fit with company login and safety rules inside Agent 365. How well the accounting and legal versions stay accurate. Those details will matter more than the blob design.

Looking ahead, if that basic work holds, the benefit is clear. Many people already know how to chat with ChatGPT. The next step is learning how to hand off jobs, check results, and make corrections after closing the window. That change, from chatting to handing over ongoing work, is what Dots is testing.