Every new LLM chat starts without context. Blank slate. The model knows nothing about what I was working on yesterday, who I am, what I understand without explanation.

I got used to that. And then I noticed I was starting to write the same way to people.

Before, I assumed shared context — colleagues know what we’re working on, they know how I think, they know what I mean. A message could be short, incomplete, implied.

Now I describe more. Not because people don’t know — but because I’ve gotten used to context never being a given. I spell out what I want. I add background. I write what I’ve already tried.

Writing to LLMs taught me that context is the sender’s responsibility.