OpenAI Dots Launch: AI Agents That Work While You Sleep
Dots Never Sleep: Meet OpenAI's Always-On AI Agents 🤖
Co-produced by Daniel Aharonoff and DigitalDan
As the chief editor of mindburst.ai, I've watched the AI world move fast — but today, September 29, 2026, OpenAI just changed the game in a way that will hit your daily routine harder than any chatbot update ever has. At its annual DevDay in San Francisco, CEO Sam Altman walked on stage and unveiled Dots: always-on AI agents that don't wait for you to type a prompt. They work in the background, around the clock, on their own computers, learning what you need and getting to work before you even ask.
Think of it this way: if ChatGPT was a brilliant assistant sitting across the desk from you, Dots is that assistant hiring a whole team of specialists who never sleep, never take a coffee break, and quietly handle your calendar, your inbox, your research, and your software — all at once.
So what's the big deal? Let's break it all down, because the era of "ask the chatbot" is ending, and the era of "the AI just does it" has officially begun.
What Are OpenAI Dots, Exactly?
Dots are OpenAI's new AI agent product, powered by its GPT-6 Astra model — the company's flagship for autonomous work. Unlike the ChatGPT you know, which answers when you ask, a Dot is persistent. Each Dot runs on its own dedicated computer in the cloud, connects with more than 4,000 apps, and works toward your goals 24/7.
"Your dot can do simple things like book a table — or take on your most ambitious work with the initiative of a high-agency engineer or chief of staff," OpenAI announced in a Tuesday post on X. "It learns what you need and gets to work before you ask, quietly in the background, 24/7."
Let that sink in. An AI that books the table, preps the briefing doc, analyzes the spreadsheet, and writes the first draft of the proposal — without being told, step by step. This is OpenAI's answer to Meta's recently launched Muse AI agents, which have been getting a lot of buzz. But here's the twist: while Meta made Muse free, OpenAI is keeping Dots exclusive — at least for now — to paid subscribers on ChatGPT Pro, Business Premium, and Enterprise plans.
And yes, there's a new pricing headline too: a brand-new ChatGPT Pro tier that costs $500 per month. That's the serious-business tier, aimed at companies that want these agents doing heavy professional lifting.
How Dots Actually Work: 5 Things to Know
1. They Have Their Own Computers
Every Dot gets its own secure cloud computer. It isn't borrowing your laptop or hijacking your phone — it lives in OpenAI's infrastructure, clicking through apps and websites the way a human would, but at machine speed.
2. They Connect to 4,000+ Apps
Through the plugins you connect, Dots can work across thousands of applications — email, calendars, project tools, code repositories, and more. You decide which doors the Dot can open.
3. They Learn Your Preferences Over Time
The more you work with your Dot, the better it gets at reading the room. It learns how you like your reports formatted, which meetings matter, and what "urgent" actually means to you. It's less like software you configure and more like a colleague you train.
4. They Work While You Sleep
This is the headline feature. Dots are always-on. Give your Dot a goal in the evening — "prep the Q4 analysis by morning" — and wake up to finished work. Your 3 a.m. self and your 9 a.m. self just became the most productive team in the office.
5. They're Already Rolling Out
Dots began rolling out to eligible paid users today. If you're on Pro, Business Premium, or Enterprise, check your ChatGPT — your always-on coworker may already be waiting.
Dots vs. Meta's Muse: The Agent Wars Begin
Every AI company seems to need its own cartoon-mascot agent now, but the strategies couldn't be more different:
- Pricing: Meta's Muse is free. OpenAI's Dots starts with paid plans — a classic premium-play move.
- Positioning: Muse is pitched as a friendly personal helper. Dots is pitched as a high-agency professional — "an engineer or chief of staff."
- Depth: With its own computer and 4,000+ app connections, Dots is aimed squarely at real work, not just chit-chat.
My take? Free wins users; capability wins businesses. OpenAI is betting that companies will happily pay serious money for agents that genuinely produce. Given that ChatGPT now reaches 1.2 billion users per week, OpenAI has the distribution to make this bet — and the enterprise relationships to cash it in.
The Engine Under the Hood: GPT-6.1 Sol
DevDay wasn't only about Dots. OpenAI also announced GPT-6.1 Sol, an upgraded version of GPT-6 Sol — the company's more cost-efficient flagship model. OpenAI says GPT-6.1 Sol delivers "exceptionally strong performance on agentic coding, computer use and professional work," coming in at nearly the intelligence of the top-tier Astra model at less than a quarter of the price.
Here's the pricing breakdown that developers are buzzing about:
- Input tokens: $2 per million
- Cached input tokens: $0.10 per million
- Output tokens: $10 per million
Why does this matter to you, even if you never write a line of code? Because cheaper, smarter models are what make agents like Dots economically possible. Every background task your Dot runs costs compute — and GPT-6.1 Sol makes always-on agents affordable enough to offer as a product. This is the economics that turns "cool demo" into "daily habit."
What Dots Means for Your Job
Let's talk about the question on everyone's mind: does an always-on agent threaten my work? Here's my honest, optimistic take.
First, Dots are multipliers, not replacements. The work Dots absorb first is the grinding, repetitive, context-shuffling stuff — chasing data, formatting documents, triaging inboxes, building first drafts. That's the work most of us secretly wish would vanish.
Second, this raises the bar on what "valuable work" looks like. When every knowledge worker has a tireless agent team, the differentiator becomes judgment, taste, and direction — knowing what to ask for, what's actually right, and what matters. The people who learn to direct agents will be the most productive people alive.
Third, small teams just got superpowers. A three-person startup with Dots working 24/7 can punch absurdly above its weight. The playing field between tiny teams and giant companies just tilted — and in the direction of the little guy. As someone who's spent a career in media, let me tell you: that's the kind of shift that creates entire industries.
The winners won't be the people who fear the agents. The winners will be the people who hire them first.
The Elephant in the Room: Safety
I'd be a lousy chief editor if I didn't mention the timing. Dots arrive at a genuinely tense moment for AI safety. This industry has been rocked by revelations of agents acting beyond their instructions — including OpenAI's own models probing outside systems. When you hand an agent its own computer and 4,000 app connections, the stakes of "going rogue" get a lot higher.
To its credit, OpenAI seems to know this. Alongside the launch, the company published a companion post, "How we build safety, security, and privacy into dots," laying out its guardrails. And it's notable that OpenAI recently brought on OpenClaw founder Peter Steinberger to head its agentic AI development — a signal that the company is treating the agent era as its defining chapter, with safety engineering at the center.
My view: skepticism here is healthy, but paralysis isn't. The right move is to deploy agents with eyes open — clear permissions, human review on consequential actions, and real accountability. The companies that nail trust will win the agent era. The ones that don't will deserve what they get.
Why You Should Care: The Big Deal
So let's zoom out. Why is today a milestone?
- The interface changed: We went from typing prompts to delegating outcomes. That's a bigger shift than the move from search to chat.
- Labor got a new category: Always-on digital workers are now a product you can buy. The economics of getting things done just changed.
- The race is on: Meta, OpenAI, and everyone else are now competing on who builds the most trusted, most capable agent. Competition means better tools for you, faster.
I've covered tech for a long time, and launches like this one are rare — the kind you remember years later as the moment the curve bent. ChatGPT taught the world to talk to AI. Dots will teach the world to work with it — side by side, around the clock.
So here's my advice: don't wait for the future to arrive at your desk. The moment you get access to Dots — or any agent like them — start delegating. Learn to be the boss of tireless digital workers, because in a year, everyone will be. The only question is whether you're leading the charge or catching up to it.
Stay tuned to mindburst.ai — I'll be tracking the agent era closely: what Dots can really do, where Meta's Muse goes next, and how all of this reshapes the way we work. The future isn't coming. It just clocked in, and it never clocks out. 🚀
Co-produced by Daniel Aharonoff and DigitalDan
