Suthahar Jegatheesan MSDEVBUILD / blog Multi-Stack Dev Build · by Suthahar
Search
GitHub Copilot

LLMs have no memory

So how does Copilot know your code? It never remembers. It re-reads, re-ranks and re-sends your context on every single request.

What you'll learn

  • Why the model holds nothing between two requests
  • How your editor rebuilds the whole prompt every time
  • What competes for the same fixed token budget
copilot 9 steps 28 seconds —views

Was this useful?

Share

Understand it one step at a time

The short runs these in order in about 28 seconds. Here they are written out — pick any step to jump the short straight to it.

1 Step 1 of 9

Copilot does not remember you

The underlying model is completely stateless between calls. It holds nothing at all across requests — not your repo, not your last question.

Frequently asked questions

Does GitHub Copilot remember your codebase between requests?
No — the underlying model is stateless and holds nothing between calls. What feels like memory is your editor rebuilding the relevant context (open files, recent edits, relevant snippets) from scratch and sending it fresh with every request.
Why does everything Copilot "knows" have to compete for the same space?
The rebuilt context, your instructions, and the model's own response all have to fit inside one fixed token budget per request — so open files, custom instructions and conversation history are all competing for the same limited window rather than living in separate memory.

Free app · no app store

These are built for a phone

Every short is drawn at full portrait height, the shape a phone already is. Installed, it opens full-bleed with no address bar across the top — and the whole library reads offline.

How it works

Read deeper on GitHub Copilot

Twenty seconds gets the shape of an idea across. These go into how it behaves in production.

More shorts

Get new posts by email

New technical articles, Azure AI and GitHub Copilot updates, and upcoming events. No spam, unsubscribe anytime.

↑↓ navigate ↵ open