| DAILY BRIEF AI / AGENTS |
| |
| OpenAI Expands Dots With Mobile Setup, Codex Delegation and Scheduled Tasks |
| The latest update pushes OpenAI's always-on agents deeper into persistent work, allowing users to create them from mobile, delegate coding work and manage recurring tasks across ChatGPT. |
| OpenAI is expanding what its Dots agents can do less than two weeks after their initial launch, adding new ways for the agents to continue work across ChatGPT, Codex and mobile devices. |
| The company says users can now create a Dot directly from the ChatGPT mobile app on iOS and Android rather than beginning setup only from desktop. |
| Dots can also start work inside Codex, continue existing coding threads and draw context from ChatGPT conversations, Codex activity and automations. |
| WHAT MAKES DOTS DIFFERENT A Dot is designed to keep working toward an outcome beyond a single chat session rather than waiting for every next instruction. Each Dot runs on GPT-6 Astra, has its own cloud computer and browser and can connect to thousands of apps through OpenAI's plugin ecosystem. | |
| OpenAI launched Dots on September 29 as always-on agents designed to maintain context, use connected tools and continue longer-running projects without requiring constant step-by-step prompting. |
| The latest update extends that model into software development. A Dot can now hand work to Codex, pick up existing threads and decide whether a task should continue in the same context or begin in a new one. |
| OpenAI also says Dots can review and edit Scheduled Tasks in ChatGPT Work, allowing an agent to inspect recurring work and adjust it as plans change. |
| That makes the product closer to a persistent work layer than a conventional chatbot. A Dot can keep several projects moving, report progress and return to the user when a decision or approval is required. |
| But the system is not designed to act without limits. Users decide which apps are connected, what actions are permitted and which steps require approval. |
| OpenAI says background proactive research is more restricted than active task execution. When the user is not directly working with a Dot, connected-app tools used for proactive research are read-only and cannot send messages, change application content or control the user's browser or computer. |
| THE SIGNAL OpenAI is moving beyond AI that answers individual prompts toward AI that maintains responsibility for ongoing work. The October update matters because Dots can now carry that work across mobile, coding and scheduled-task environments. | |
| 03 | CORE NOTES What OpenAI has confirmed | |
| 01 Dots were launched on September 29 and are rolling out across Pro, Business Premium and Enterprise plans in eligible markets. |
| 02 The October 9 update lets users create a Dot from mobile, delegate work into Codex and continue existing coding threads. |
| 03 Users retain control over app access, permissions and approvals, while background proactive research uses restricted read-only tools. |
| |
| SOURCES OpenAI — Introducing Dots
OpenAI — Fresh Dots Updates
OpenAI Help Center — Getting Started With Your Dot
AI Magazine — OpenAI Dots and Autonomous Agents |
