Hacker News new | past | comments | ask | show | jobs | submit login

I learned this very explicitly recently. I've had some success with project and branch prompts - feeding a bunch of context into the beginning of each dialog.

In one dialog, some 30k tokens later, Claude requested the contents of package.json... which was in the context window already - the whole file!

The strange thing was that after I said so, without re-inserting, Claude successfully read it from context to fill the gap in what it was trying to do.

It's as if a synopsis of what exists in-context delivered with each message would help. But that feels weird!




That's what these char models are already doing.

Most chat is just a long running prompt. LLMs have zero actual memory. You just keep feeding it history.

Maybe I misunderstood what you're saying but what you're describing is some kind of 2nd model that condenses that history and that gets fed; this has been done.

Really, what you probably need is another model managing the heap and the stack of the history and brining forward the current context.

But that's easy to say because we are humans.


To clarify:

I didn't 'fix' the problem by re-inserting the package.json. I just gave a reminder that it that it was already in context.

"I could confirm this root cause if I could see the contents of package.json".

"You can see it."

"Whoops, yes. Package x needs to bump from n to n+1."

The point being, even info inside the current context can actively be overlooked unless it is specifically referenced.




Consider applying for YC's Summer 2025 batch! Applications are open till May 13

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: