Story
October 3, 2026
OpenAI’s Dots Wants Your Workflow, but Meta Still Has the Easier Sell
OpenAI is positioning Dots as a persistent workplace agent that can carry context across tasks and teams. But its paid, enterprise-first launch leaves Meta’s free Muse with the simpler pitch for users unwilling to hand one company their digital life.
At DevDay in San Francisco, OpenAI unveiled Dots, an “always on” agent meant to work through connected apps, retain selected context and take on ongoing responsibilities rather than merely answer prompts.1 The launch arrived as personal agents become the industry’s clearest bid for loyalty: once users grant an assistant access to their calendars, files and routines, switching providers becomes much harder.
The immediate comparison was Meta’s Muse. One account cast the contest in deliberately theatrical terms, saying the two companies’ “cutesy AI agents are about to fight to the death.”2 Both products offer personable avatars, memory and computer-use capabilities, but their commercial logic sharply diverges. Muse offers a free entry point, while Dots initially sits behind paid OpenAI plans — a constraint that limits access for much of ChatGPT’s vast user base.2
Early hands-on coverage reinforced that divide. Dots can operate applications such as Blender and GIMP, access a user’s computer through the ChatGPT desktop app and even be called for spoken help. Yet it feels less like a consumer assistant than “workplace software that happens to be able to order you a burrito.”3 Sam Altman made the intended payoff explicit, saying the product improves as it learns “more of my workflow and style” and handles the chores that become a “gravity well of dread.”
4
For proponents, that persistence is the point. One DevDay reviewer found a Dot catching a recurring meeting scheduled on weekends and flagging a conflict between a DMV appointment and travel plans.5 More consequentially, OpenAI is pairing Dots with Space, where colleagues and agents can work from shared documents, turning accumulated project knowledge into revised sales material or customer proposals.5
But the same design raises the central tension. The more useful an agent becomes, the more information and responsibility it needs. Critics warn that users may resist the “complete handover” of their lives to bots — and the risk of getting hooked on a service that may later become more expensive.1 Dots is betting that useful work will overcome that hesitation; Meta is betting that free and frictionless gets there first.