Slorks Blog
If Slorks had started in the blog era — tech in writing, pictures and discussion
عEN中Sign inTime: --:-- Gulf time
Ad space · 970×90 · Advertise here

Home › News › #2026-09-30-openai-dots-always-on-agents

News OpenAIChatGPTDotsوكلاءذكاء اصطناعيأخبار

OpenAI تطلق Dots: موظف ذكاء اصطناعي يشتغل ٢٤ ساعة على كمبيوتر خاص فيه — والأول مجاني لمشتركي Pro

OpenAI launches Dots: an AI employee that works 24/7 on its own computer, and the first one is free for Pro subscribers

by Slorks 2026-09-30 14 0 replies
Machine-translated — the Arabic text is the original
OpenAI تطلق Dots: موظف ذكاء اصطناعي يشتغل ٢٤ ساعة على كمبيوتر خاص فيه — والأول مجاني لمشتركي Proالصورة: OpenAI

A week after Meta's Muse, OpenAI answered with its own device, minus the device. Dots are "always-on" AI agents: each Dot has its own cloud computer and browser, works in the background while you sleep, and comes back with results in ChatGPT, Slack or Microsoft Teams (texting "coming soon"). The engine is GPT-6 Astra.

OpenAI Dots
The four Dot characters; each has a name, colour and personality you can change — Photo: OpenAI

How is that different from regular ChatGPT?

ChatGPT answers when you ask. A Dot works without being asked: you give it a project ("watch customer feedback on the app and prepare a summary every morning"), it opens its own browser, reaches into connected apps (4,000+ via plugins), uses saved passwords without the model itself ever seeing them, and keeps going. It learns your style over time and takes feedback. You can optionally connect it to your laptop.

Ad space · full width · Advertise here

OpenAI's examples: a developer having it analyse user feedback and prepare code fixes, a launch lead having it revise creative the moment the plan changes, a scientist having it update figures as data comes in, a sales lead having it prepare proposals, and a content creator handing it an interview recording and getting posts back.

A Dot preparing a post from an interview
The creator example: a Dot named Jojo turns an interview into posts — Photo: OpenAI

Safety and control

OpenAI says there are built-in safeguards against malicious instructions (a website trying to command the agent, for instance), Custom Rules where you define what's allowed, automatic review of any action that touches your account, and an Activity view showing every step it took. Some things stay yours alone, like changing passwords.

Price and availability

PlansFirst Dot included for ChatGPT Pro, Business Premium and Enterprise (including Edu and Healthcare)
More than oneAdding Dots "soon" with scaling options; price not announced
Plus and FreeNot included for now
RolloutStarted September 29, 2026 in "eligible markets"; OpenAI published no country list
EnterpriseA "specialist" variant with organisational identity and credentials, integrated with Microsoft Agent 365

My take

This matters more than any device announced this month. Meta put the agent on a keychain; OpenAI put it in an employee that works in the background with no interface at all, and the difference is that the second one can actually save you hours. For creators like me, "hand it the interview, get the posts back" is the first thing I'll try. Two open questions: is the Gulf among the "eligible markets"? (Pro is sold in the UAE, so probably yes, but OpenAI hasn't confirmed), and what the second and third Dot will cost. If the answer is reasonable, the $200 Pro plan is worth it for the first time.

Sources: OpenAI — Introducing dots

Reading time: 2 min
Worth reading?
Reply0
Reply0
Voting and rating need no account. Replying does — it keeps the discussion clean.Sign in to reply
No replies yet — be the first.