For the love of God, people need to get back to writing their own Readmes and not just taking LLM output unedited. I do want to read, but I don't want to read C

For the love of God, people need to get back to writing their own Readmes and not just taking LLM output unedited. I do want to read, but I don't want to read ChatGPTese. Sorry, let me ask ChatGPT to put it in terms people seem to prefer now (I don't think this stuff is actually quite right but who cares anymore): ## 1. They Optimize for Politeness, Not Usefulness ChatGPT READMEs tend to: - Over-explain obvious things - Avoid strong claims - Hedge unnecessarily The result is text that feels safe but not informative. A good README should reduce uncertainty quickly, not pad it with disclaimers and filler. ## 2. They Follow Templates Instead of Intent Most generated READMEs look structurally correct but contextually shallow: - Generic section headings (“Installation”, “Usage”, “Contributing”) regardless of relevance - Boilerplate language that could apply to almost any project - No clear prioritization of what actually matters This signals that the README was assembled, not written with purpose. ## Summary ChatGPT READMEs are usually: - Correct but unhelpful - Polished but shallow - Complete but low-signal