The cost of starting over
Every new chat begins from zero. A calmer way to work with assistants that don't remember.
Open a fresh chat and your assistant knows nothing about you. Not your stack. Not last week’s decision. Not the project you’ve spent three days inside. So you set the stage again: the same paragraph of background, retyped, a little different every time.
That’s not a flaw to fix in yourself. The model reasons well and remembers nothing between sessions. Memory was never part of the deal.
The usual workaround is a bigger prompt. You keep a doc of “context to paste,” and you paste it. It works until it doesn’t. The doc goes stale, it lives on one machine, and it only covers the assistant you copied it into.
There’s a better shape. Keep the context outside the conversation, in one place, and let the assistant read it directly. When the work moves forward, it writes the new piece back. Next session, that part is already there.
That’s a knowledge layer. Not a transcript of your chats. Not a second brain you tend by hand. A small, durable set of things worth remembering: who you are, how you work, what you’re building.
MCP is what makes it practical. One connection from your client to your layer, and the background comes along. Start a session and the assistant already has the shape of your work. Ask it to note a decision and it lands somewhere you’ll find again, not in a thread you’ll lose by Friday.
You’re still in charge. The layer doesn’t act on its own, and it won’t nag you to feed it. It just means the next conversation doesn’t start from nothing.
vtriv is built around this idea. If the retyping has been getting to you, give it a try: get started.