logoalt Hacker News

consptoday at 10:05 AM1 replyview on HN

I'd argue there would be two lanes as well: one where the issues are addressed in code, the other being the discussion of why people think this is a bad idea and speak so openly about it. This topic is the second I guess. Looking at the flow there is quite a bit of flamebait by the LLM and non-LLM camps which only muddies the water and doesn't resolve anything. The better discussion (imo) would be to decide if the vide coded fixes are worth it and if not, fork the project somewhere and let the distro's chip in to maintain that.


Replies

sysguesttoday at 10:46 AM

idk maybe LLM people should only commit what they actually understand, only in bite-size (maximum few lines in few files) and with at least 1~5 tests that shows the edge cases

drive-by 20-file pull-requests that ultimately end up costing maintainer's burden seems to hit hard here.