OpenAI has launched dots, an always-on personal assistant that keeps working in the background even when you haven’t asked it to do anything. The company unveiled it at its developer day in late September. According to TechCrunch AI, Alexander Embiricos, who leads OpenAI’s enterprise product team, used TechCrunch Disrupt this month to explain how the product grew out of the company’s coding tools.
The launch puts OpenAI in this year’s always-on assistant race. The startup Instinct became an overnight sensation, and Meta has put its full marketing weight behind its own assistant, Muse.
It’s Codex, repackaged for everyone
Embiricos admits that you could get much of what dots does out of Codex, the coding app OpenAI released early this year. Doing that takes far more setup and hand-holding, though. dots takes that agentic plumbing and wraps it in something regular people can use. OpenAI did something similar when it spun ChatGPT Work out of its developer tools.
“A lot of that isn’t building some new piece of tech into the product; it’s how you frame it to the user and how it feels,” Embiricos told the Disrupt audience.
One entity, one conversation
The key design choice is presenting dots as a single persistent “entity” instead of a pile of separate chats. “If we could frame this as, it’s an entity and it’s a single conversation, not many different conversations… it starts to become intuitive,” Embiricos said. He calls the industry’s drift toward cute assistant mascots “convergent evolution.”
The name follows the same idea. It plays on the negative space inside OpenAI’s logo and is meant “to feel like just something that could eventually be ubiquitous.” One awkward detail: Elon Musk owns the dot.com domain. Embiricos says he “found out very late.”
It acts before you ask
The whole point of dots is that it doesn’t wait for prompts. Once you connect it to your calendar, email, computer and other tools, it looks for ways to help on its own. Here’s what that covers:
- Calendar and deadline alerts: It flags conflicts and upcoming due dates.
- Proactive research: On a regular cadence, it reviews what it knows about you and your conversations, then proposes ways to help.
- Incident response: Embiricos described an engineer at a drone delivery startup whose dot had already spotted an outage, analyzed it and suggested fixes before the engineer noticed anything was wrong.
This cuts both ways, and it’s worth saying plainly. An assistant that thinks for you all the time burns far more tokens than one you prompt now and then. That’s good business for OpenAI.
The timing still needs work
In TechCrunch’s testing, dots’ ability to think ahead was clear and often useful, but it isn’t quite there yet. The alerts don’t always arrive in time. Embiricos says one early warning about a calendar conflict reached him three minutes before the event. “I only have three minutes!” he recalled. He expects the timing to get better.
Privacy controls are still coarse
This is the biggest open question. dots has the same privacy settings as ChatGPT, so you can turn off model training and decide which resources it can reach on your computer. TechCrunch notes those controls aren’t granular, though, and the app can be pushy. For example, it won’t accept read-only access to cloud drives.
Embiricos says OpenAI needs to “build out super, super, super granular controls.” He also admits the company took heat at launch for the opposite problem: conservative defaults that had the agent asking for permission constantly.
The data question hangs over every assistant in this category. Muse has drawn criticism for building profiles of users’ friends and family. Rival startup Hark makes a point of saying it doesn’t collect data to feed an ad business the way OpenAI and Meta do.
Availability and pricing
dots is available in the ChatGPT desktop and mobile apps for subscribers on the Pro tier and above. Plans start at $100 a month.
Why it matters
I think the real story is the strategy behind dots. OpenAI already built the agent capabilities for developers, and now it’s competing on packaging and feel. Whether dots can scale is another matter. Businesses and people don’t always welcome dealing with agents, and the privacy tradeoff gets harder to accept the more your assistant knows. Expect the fight over persistent assistants to come down to trust as much as raw capability. TechCrunch AI’s full piece has more of Embiricos’ Disrupt conversation.