Tech • AI • Robotics • Game
ChatGPT Dots Finally Clicked When I Learned This
8/10TL;DR
OpenAI’s Dots is emerging less as a generic personal assistant and more as a persistent chief-of-staff style agent that monitors ongoing work, follows up across tasks, and acts inside the broader ChatGPT ecosystem.
KEY POINTS
What Dots is
Dots is described as OpenAI’s always-on agent, built into ChatGPT with access to memory, connected tools, and a dedicated cloud computer with its own browser. That means it can continue working in the background even when a user’s laptop is closed. It currently runs on GPT-6 Astra, with OpenAI indicating that additional dots may be added later.
A crowded agent market
The product enters a field where rivals such as Muse and Grockbot already offer background execution, app connections, memory, and browser-based automation. On paper, those capabilities are not unique. The distinction presented for Dots is not basic automation, but how tightly it is woven into existing ChatGPT threads, memory, and workflows.
Best use case: staying on top of everything
Rather than treating Dots mainly as a shopping or booking assistant, the stronger fit appears to be coordination and follow-through. Its value lies in continuously tracking open loops, spotting issues, and surfacing actions without waiting for a fresh prompt. That makes it closer to a digital operations aide than a one-off task bot.
Proactive interventions
Early examples show Dots acting before being asked. It flagged a potential hotel billing issue before checkout, suggested postponing a planned task until key details were finalized, and surfaced an email that had gone unanswered. The core appeal is that it identifies loose ends on its own and turns them into actionable reminders.
Following up after tasks are sent
Dots also appears designed to revisit work after an action has already been taken. In one case, after an email was drafted and sent to inquire about a missing payment, the agent later checked independently and found that the payment had in fact already arrived. That prompted a corrective follow-up, illustrating a more active style of post-task verification.
Voice tasking and multi-project coordination
The system includes a voice feature framed as a “call,” allowing users to hand off multiple assignments in one session. In a single interaction, Dots was asked to research ChatGPT Spaces for a video outline, update an app called Dayboard with lightweight editing features tied to Obsidian, and fix gameplay bugs in a project called Schmela. It then summarized those assignments and queued them across existing work threads.
Why ecosystem integration matters
The most significant advantage may be that Dots is not isolated from the rest of a user’s AI work. Because it sits inside ChatGPT, it can move across prior conversations, coding threads, notes, and connected tools instead of forcing users to recreate context. That allows it to coordinate tasks across research, software development, scheduling, and communication in a single environment.
A replacement for manual oversight
The agent’s role is framed as reducing the need for constant check-ins. Instead of repeatedly monitoring whether subtasks are progressing, a user can delegate the goal and rely on Dots to track the moving pieces, request clarification when needed, and send updates through the mobile app. The promise is less time spent managing the manager.
Limits on access
Access remains restricted. Dots currently requires a Pro plan starting at $100 per month or a business plan, and it is not yet available in the EU, the UK, or Switzerland. Those limits make the tool more of an early premium feature than a mass-market assistant.
Where the product may go next
The roadmap points toward multiple dots with distinct roles, potentially allowing users to assign specialized responsibilities to separate agents. That could move the system closer to a full team-of-agents model, with dedicated bots handling operations, research, coding, or communications on a continuous basis.
CONCLUSION
Dots appears to be most compelling not as a novelty assistant but as a persistent coordination layer for people already working heavily inside ChatGPT. Its long-term importance will depend on whether OpenAI can expand access and turn that early chief-of-staff model into a reliable multi-agent system.
Ask a question