Why are detailed comments considered bad?
retained text
Reading over the threads in this sub, the majority of people dislike that the models are adding blocks of comments over function, variables etc. Sure, I agree that reading code in this style is a bit painful. But, if you are mostly coding with LLMs and not deep diving into what the model wrote (which I believe most of us are), isn't this a good thing? I see this as the model is leaving context for itself in it's future sessions so the next time it is working on a feature over something it touched before, it has some context on that code from before. So in general I tend to leave all the comments, and so far the results have been good for me.
document fetched 2026-09-03 00:41 · canonical https://reddit.com/r/ClaudeCode/comments/1w5a84l/why_are_detailed_comments_considered_bad