Memory

Both products include one — Memory (both) — and it looks like a counter-example to everything the window teaches. It isn't. The model still remembers nothing between sends. The app keeps notes about you and loads them into the window on your behalf — a saved file of facts, and in some products a reach back into your older chats for whatever looks relevant. That is the ledger confirming itself, not breaking: everything it knows about your problem is in the window, including what the app put there for you.

Which is exactly why you can open memory in settings, read it, and delete what you don't want carried. A new account's file is empty or nearly so; the point of knowing it exists is knowing where it lives, that it is readable, and that every line in it can ride into any window until you remove it.

Two working rules:

• Read the file before you trust a "remembered" fact. A note the app saved from an old chat is window text like any other — it can tilt the odds in any conversation after it, including the ones it does not belong in.
• Keep durable facts there; keep project facts in the project. Memory is for who you are across all your work; a project is for what one body of work needs. Loading the wrong one into every window is how unrelated context starts voting.

Say it back in one line: when a chat seems to remember last month, what is actually doing the remembering?