I think Dots is one of those AI releases that sounds much smaller than it actually is.
You give it something to look after. You close your laptop. You go to sleep, attend a meeting, work on something else, whatever.
And the work does not necessarily stop with you.
Your Dot can keep checking what changed, read the apps you gave it, prepare work, start other AI tasks, and come back when it needs a decision from you.
That is the part worth paying attention to.
Because we have spent the last few years learning how to get better answers from AI.
Dots is about something else:
giving AI something to keep responsible for.
This little thing has its own computer
That is what makes Dots weirdly interesting.
Every Dot runs on GPT-6 Astra and gets its own computer and browser in OpenAI’s cloud. It can connect to more than 4,000 apps through plugins, work with ChatGPT Work and Codex, use background agents, and keep working even when your own computer is off.
So instead of opening ChatGPT every morning and typing:
What do I need to do today?
you can give your Dot the responsibility:
Keep track of my important work. Check my calendar, selected emails and project updates. Tell me only what needs my attention.
Then it keeps doing that job.
A creator can have it watch new interview transcripts and prepare content.
A founder can have it keep a launch up to date.
A developer can have it investigate small bugs and send the work into Codex.
A consultant can have it prepare the same client report every Friday.
That is where Dots starts to make sense.
And the setup is much easier than it sounds.
Behind the paywall: the exact Dots setup, cost, plugins, memory, permissions, automations, copy-paste prompts, real workflows, background agents, Astra/Sol and the few settings that decide whether your Dot actually becomes useful while you’re away.
Step 1: Give your Dot one real job
Do not begin with:
Help me with everything.
Technically, a Dot can handle several ongoing responsibilities. But for your first setup, give it one job that keeps happening.
Read more
