Try it

Where does this get remembered?

Pick something your dot learns. The place that holds it lights up, and you'll see who else can use it and how to get rid of it.

Your dot learns that... You tell it: "I prefer morning meetings" It reads in Drive: "the Q4 deck is in Finance" ChatGPT already knows: "you're vegetarian" You say: "use the attached sheet for this report" You mention a deadline in a Slack DM You type a password into its sign-in form You send it a screenshot

You tell it: "I prefer morning meetings"

Kept in: Your dot's notes. A preference is what the notes are for. What you tell your dot can also add to ChatGPT memory.
Who uses it: Your dot, wherever you reach it: ChatGPT, Slack or a call. Regular ChatGPT chats too, if the preference reached ChatGPT memory.
Can you see it?: Not in the notes. If it reached ChatGPT memory, it shows up there.
To remove it: From ChatGPT memory: delete it in ChatGPT's memory settings. From the notes: only by deleting your dot.

It reads in Drive: "the Q4 deck is in Finance"

Kept in: Your dot's notes. Your dot reads connected apps without being asked, and can save what it finds.
Who uses it: Your dot, in later conversations and jobs.
Can you see it?: No. Ask your dot what it knows about the deck and it will tell you in words.
To remove it: Disconnecting Drive stops new reading. It leaves what your dot already saved. Only deleting your dot removes that.

ChatGPT already knows: "you're vegetarian"

Kept in: ChatGPT memory. Your dot starts with the relevant parts of it.
Who uses it: Regular ChatGPT chats and your dot.
Can you see it?: Yes, in ChatGPT's memory settings.
To remove it: Manage it in ChatGPT's memory settings. A change there doesn't necessarily change notes your dot has already made.

You say: "use the attached sheet for this report"

Kept in: This conversation. It's context for the reply or the job in front of your dot.
Who uses it: Your dot, for this job. A task it hands to Work or Codex gets the instructions and context for that job, not your whole history.
Can you see it?: Yes. Scroll up: it is in the conversation.
To remove it: Deleting your dot deletes its conversation. The notes aren't a full transcript, so a one-off detail like this may never reach them.

You mention a deadline in a Slack DM

Kept in: Your dot's notes. Slack and ChatGPT reach the same dot, with the same notes.
Who uses it: Your dot, in every channel. It can use the deadline to prepare an update in ChatGPT.
Can you see it?: The Slack message stays in Slack. It isn't copied into your ChatGPT conversation.
To remove it: Only by deleting your dot. Using the deadline is allowed. Posting it in a team channel still needs your permission.

You type a password into its sign-in form

Kept in: Nowhere in its memory. Your dot's saved context doesn't keep credentials.
Who uses it: The website you signed in to. Your dot's browser stays signed in until you sign out or the site ends the session.
Can you see it?: There is nothing to see.
To remove it: Nothing to remove. One caution: this covers the sign-in form. A password pasted into the chat is an ordinary message.

You send it a screenshot

Kept in: Nowhere in its memory. Your dot's saved context doesn't keep images or screenshots.
Who uses it: Your dot, while it works on what you sent.
Can you see it?: The message you sent is still in the conversation.
To remove it: Nothing in the notes to remove. Files your dot makes live where they were saved and follow that place's rules.