Muse vs a Dot for “do this and ping me”: which one is a phone app, and which one is a cloud computer you message?

Both are cloud computers you message. Meta’s Muse, announced 8 September 2026, runs on Muse Secure VM, a dedicated computer in the cloud with its own browser, and you talk to it like a person in the Muse app, in WhatsApp, or on muse.ai. An OpenAI Dot, announced 29 September 2026, runs on its own cloud computer inside ChatGPT, on GPT-6 Astra, and you message it in ChatGPT, then in Slack or Teams after you connect those on a desktop. The phone is a client for both. It is not the computer. Muse is the one whose front door is a messaging app. A Dot is the one whose front door is ChatGPT, and you cannot create that Dot on the phone.

“Do this and ping me” fits both, and the ping is not the same door. Muse keeps working after you close the app and comes back when something changes or it needs approval, such as before it sends an email or makes a purchase. A Dot can message you with progress, questions, or decisions, and the learn docs say you can tell it which channel gets which kind of update.

Where do you actually send the instruction?

For Muse, in a chat that already feels like texting someone. Meta’s announcement says talking to it works just like messaging another person, in the Muse app or directly in WhatsApp. The rollout it describes is the United States, on iOS, Android, and muse.ai, with AI glasses listed as coming soon. I am not going to add a Mac app, a button name, or a country the announcement does not list. You tell it what needs to get done. It takes action on Muse Spark, which Meta calls its model for this kind of work. For a task that takes a while, it keeps going after the app is closed.

For a Dot, the first message is on a desktop. The getting-started page says you create it in the ChatGPT desktop app or in ChatGPT on desktop web, and that you cannot currently create a Dot on mobile or on mobile web. After that, you can talk to it in the ChatGPT mobile app when mobile access is available. Slack and Teams are connected from the desktop. The announcement says you can message or call Dots in ChatGPT on desktop, web, and mobile, and that they can message you back with progress, a question, or a decision. Texting is described as coming soon in that announcement. The help article is narrower: if texting is available for your account, you connect the phone from the desktop app, then message from the phone. If your account does not show it, do not assume WhatsApp-style texting exists for the Dot. A Dot is not something you start from a phone with no computer open. Muse’s announcement presents the phone app and WhatsApp as the way you talk to it from the start.

What computer is doing the work while you wait?

Muse Secure VM. Meta says each person gets a dedicated computer in the cloud, contained so no one else’s agent can reach it. That is where Muse lives, and where the data and credentials for services you connect are stored. The research write-up describes the VM as an isolated machine with a browser, storage, and enough compute to do real work, including several sub-agents at once. You can watch the browser and take over. A separate Sentinel agent on that same machine is what Meta says must approve anything that reaches the internet, and it asks you when it needs to. Muse Confidential VM, where the VM would be encrypted with a key only you hold, is planned for later. It is not the computer you get in the September launch. Testers exist. The product you message today, if you are in the rollout, is Secure VM.

A Dot’s computer is OpenAI’s, and it is that Dot’s. Files, software, and browser sessions live there. Cloud work continues while your devices are off. Your own laptop stays disconnected until you allow access, and only one personal computer can be connected, with the ChatGPT app left open while the Dot uses it. A second Dot, so a second job would not share that disk, is still described as a future option. The primary Dot is what you have. Two “ping me when it’s done” jobs on the same Dot share the computer. They do not share it with Muse. These are different vendors and different machines. Signing both into the same mailbox would be two clients again, the same kind of collision as a Dot and a Grok Bot, and neither announcement says they know about each other.

What does the ping actually mean?

On Muse, Meta’s sentence is specific. It comes back when something changes or when it needs approval. The examples in the announcement are sending an email and making a purchase. It also shows a complete audit trail of what it has done and what it plans to do. You choose the apps, and for email you choose whether it reads or can also send. You can disconnect a service when you want. You can tell it to forget specific things it learned. I have not found a help page that names the notification toggle, the sound, or a “ping” button. “Comes back” is the behavior they published. If the app is closed, that return has to land somewhere you will see: the Muse app, WhatsApp, or muse.ai. I will not invent which of those fires a push alert on the lock screen.

On a Dot, the return is a message in a channel you already use for it. The announcement says it can message you with progress, questions, or decisions that need you. The learn page says to tell it where different updates should go. The example shape is routine progress in one place and a decision in another, such as ChatGPT for the quiet updates and Slack when a deadline is at risk. Scheduled runs and proactive updates show up in the conversation. You manage a scheduled task from the Dot’s profile or from Scheduled in ChatGPT, including whether completion notifications are on. That is as close as the docs get to “ping me.” It is an instruction plus a notification setting on the task, not a separate phone product.

A call is not the ping. The learn page says you can call the Dot from its conversation to talk something out, and that work you assigned can continue after the call ends. I have not read an OpenAI line that says the Dot will ring your phone on its own. Do not wait for a phone call from either product unless a page you are looking at says that call is outbound. Muse’s announcement does not say it places a voice call. It says it messages you by coming back.

What still needs you, even if you asked it to just do it?

Money, and a send you did not specifically allow. Muse checks with you before a sensitive action such as sending an email or making a purchase. The research note on shopping is stricter than a vibe: if payment credentials are already on file at a site, checkout asks for a person every time, with the purchase details. For a wallet path, Meta says it partnered with Stripe’s Link at launch, that a one-time card is what the merchant sees, and that Shop Pay is coming soon, along with 1Password so Muse can use logins you already have. I am not going to name a subscription price. The announcement says Muse is free for most of what people need, with subscription plans for people who want to do more. Passwords are not in the chat. Meta says Muse has no visibility into passwords or payment methods. Credentials go into secure storage. Muse can use them without seeing them, and it cannot see a password you type into the browser yourself. When you take over the browser, the agent is paused.

A Dot’s version of that line is takeover, not a Meta Sentinel. Changing a password or transferring money requires you to finish it yourself. Custom rules cannot turn that off. Asking it to draft replies does not give it permission to send them. A custom rule can say to ask first, or to hand the step to you, and the Dot can still make a mistake inside a rule. Proactive research is read-only: those tools cannot send, change app content, or control the browser or computer. A follow-up send is a separate permission. If “ping me” was supposed to mean “draft it and stop,” say that. If it was supposed to mean “send it,” you are approving a send, on whichever product you connected the mailbox to, and you should not connect send on both.

Which one do you text when you want the result later?

Text Muse if you want the instruction to live in a messaging app, in the US rollout, on a cloud computer Meta runs for that agent, and you are willing to be asked before an email goes out or a purchase happens. Message a Dot if the work already lives in ChatGPT, Slack, or Teams, and you want the ping in one of those channels, from a cloud computer OpenAI runs, after you created the Dot on a desktop. Do not pick Muse because you heard it was “the phone one” and a Dot was “the computer one.” Muse is a phone-shaped door onto a cloud VM. A Dot is a ChatGPT-shaped door onto a cloud computer, with a phone door that opens only after setup.

The job “do this and ping me” fails the same way on both if you never said what “done” is, or which send you are willing to see happen without you. The computer will keep working after the app closes. The ping is the product coming back to a chat. The approval is still yours when the next step spends money or speaks in your name.