Lessons learnedEditorial digest
Cross-repository work is a context problem
gwerbin describes synthesizing information across three repositories and many business documents. In their experience, weaker reasoning settings missed details, while longer outputs became difficult to inspect. They wonder whether a retrieval-based memory system could help.
Original excerptOriginal discussionAnthropic appears to be A/B testing reduced effort levels in Claude Code“I found that I need the big model and the high reasoning effort on tasks where I have a large amount of details of”
Read the full original