Dots in ChatGPT: always-on AI agents that you hand responsibilities to
Melvin Vivas · X video post · 2026-09-30 · 6:54 · 568 views · Open on X
Topics: AI Agents, Tool Use & MCP, Industry Trends & Job Market, AI Safety, Security & Guardrails · Level: beginner
Summary
A clip of the launch event for Dots, presented by "Sam" (probably Sam Altman, since Dots ships in ChatGPT). Dots are always-on agents inside ChatGPT, powered by a model called Astra. The demo shows a Dot running work and personal tasks for weeks at a time. It is useful for seeing where agent products are heading: each agent gets its own cloud computer and browser, connects to apps through plugins, can be reached by text or phone, and asks a person to approve important actions.
Key points
- Dots are always-on agents in ChatGPT. You give one an ongoing responsibility (e.g. 'monitor incoming bug reports, test fixes and draft PRs') instead of one-off tasks like booking a ticket.
- They run on Astra, which the presenter calls 'our most aligned model', and come with built-in protections and safeguards.
- Each Dot has its own computer and browser in the cloud. It uses the access you give it and can write and test code.
- Coding example: move an app off a legacy API before it shuts down. The Dot traces dependencies across the codebase, writes the code, runs tests and opens pull requests for the team to review.
- Dots work with more than 4,000 apps in ChatGPT and reuse the plugins you have already connected (the demo uses Teams, Google Meet notes and Slack).
- Approval is built in: the Dot checks in when something important needs your sign-off (e.g. a staged rollout). You can limit which apps it uses and what it does on the computer, and customize instructions for its actions.
- Availability: you start with one Dot, with teams of Dots planned. Texting and phone calls with your Dot are 'coming soon'.
- The Dot learns how you like to work over time. For example, it reviews what came in overnight and pings you in the morning with anything urgent.
Resources mentioned
- Dots · tool · openai.com · paid
Always-on agents in ChatGPT, each with its own cloud computer and browser, that take on ongoing work and personal responsibilities.
Also in: Claude Duet: Anthropic's Research Preview of a Proactive Agent App (Melvin Vivas on X · notes), Emoji Acknowledgement Prompt for Agents (Built Into Grok Bot) (Melvin Vivas on X · notes), Grok Bot vs Dots: Multiple Bots vs a Single Bot (Melvin Vivas on X · notes), ChatGPT Dots on Mobile: Setting Up an Always-On Proactive Agent (Melvin Vivas on X · notes) and 14 more - ChatGPT · tool · chatgpt.com · free · recommended by both Bashiri Smith & Melvin Vivas
OpenAI's AI assistant (web, desktop and mobile apps), used for chat, coding, writing and image generation.
Also in: Keep a Change Log When Multiple AI Agents Edit One Notion Doc (Melvin Vivas on X · notes), Multi-Agent Planning in Notion: Grok Bot, ChatGPT and Claude (Melvin Vivas on X · notes), Hark Pro Launch: A Proactive Personal AI Agent with Cloud Computer Use (Melvin Vivas on X · notes), Why Grok Bot Could Be the Next "ChatGPT Moment" for Personal AI Assistants (Melvin Vivas on X · notes) and 35 more - GPT-6 Astra · tool · openai.com · paid
The model announced in the quoted launch post, pitched as the developer's most capable model for work, coding, science and cybersecurity, and able to operate a computer.
Also in: Glimpse: A Visual Reasoning Model That Uses a Visual Scratch Space (Melvin Vivas on X · notes), Is AI Getting Cheaper? How to Measure LLM Cost Per Task and Per Intelligence Level (Melvin Vivas on X · notes), GitHub Copilot Adds Intelligent Local Model Routing (Project HydraFusion) (Melvin Vivas on X · notes), Comparing Opus 5.5, GPT-6 Astra and Grok 4.7 on OpenRouter (Melvin Vivas on X · notes) and 55 more
More in AI Agents, Tool Use & MCP
- GPT-6.1 Sol May Beat Luna for Subagents
- Using Jev (TypeSafe AI) as a Decision Model for Email Classification
- OpenAI Dots: Agents Inside ChatGPT That Debug, Fix, and Open PRs
- Do you still need Context7 MCP if the model browses docs itself?
- Jev Decision Model: Forecasting AI Receptionist Calls from Structure Alone
- Cloudflare's 'cf': an agentic CLI for the whole Cloudflare API