A language model can view only so much text at once. When a book or conversation is too large, chunking splits it into smaller pieces instead of trying to fit everything into one request. section 7.1
Chunking is one way to manage that limit. Summarization packs the text down; retrieval fetches relevant pieces when needed. Choose an approach that gives the model useful information without overloading it, then test speed, cost, and quality. section 7.1
