OpenAI’s Dots Bets Premium Users Will Trust AI to Act Alone
OpenAI’s Dots Bets Premium Users Will Trust AI to Act Alone
OpenAI unveiled Dots at its Dev Day on Tuesday, framing the GPT-6 Astra-powered product as a step beyond the chatbot era and directly into autonomous work. Rather than staying tethered to a single app or device, the agents are designed to pursue user-defined goals continuously in the background, checking back only when they need clarification or approval.
The company’s pitch is deliberately ambitious. Users can name a primary agent, assign it work and eventually, OpenAI says, deploy teams of agents on their behalf. “Today, you can start with your primary dot, give it a name, and make it your own,” the company said. Sam Altman went further at the event, calling Dots “the real deal version of AI we’ve always imagined.”
The initial rollout is more selective than Meta’s approach with Muse. Starting Tuesday, Dots are available through ChatGPT and Codex to Pro, Business Premium and Enterprise users in eligible markets, with one agent per user at launch. Each Dot has its own cloud computer and can work across connected apps, ChatGPT Work and Codex on research, data analysis, documents and software development.
That focus makes developers the first test case: OpenAI imagines an agent monitoring customer feedback and shipping fixes, or a scientist’s assistant rerunning analysis as new experimental data arrives. It is also working with Microsoft on Agent 365 security controls, a signal that enterprise governance is central to the product’s appeal.
The contrast with Meta is stark. Muse launched broadly to Meta’s enormous consumer base and quickly climbed download charts; OpenAI, by contrast, is selling autonomy first to high-end subscribers. Yet the competition is converging: as Dots targets power users, Meta has begun connecting Muse to tools used by small businesses.
OpenAI’s live demonstrations captured both sides of the wager—moments of excitement, but also a noticeable lag when a presenter tried to speak to her Dot. The race is no longer about whether AI can answer. It is about whether people will trust it to act.
Write a comment