On: Agents don't need memory, they need documentation

[kaydub]

This is one of the things I REALLY don't get.

If you got the LLM to generate the docs, they don't need the docs.

[01100011 → kaydub]

Think of the docs as a cache. You're saving tokens by starting with a map of the code. Why try to figure it out each time?

[kaydub → 01100011]

I'm arguing that the people doing this THINK it's saving tokens. I don't think it is. I've been seeing faster and more efficient use of the LLM by removing a lot of the documentation.