Post #2736477

At work today I reviewed yet another PR that looked cursed by an LLM. Not fully vibe-coded but noticeably unpolished or sloppy (in the non-"AI" sense). I trust that my colleagues are well-intentioned and that they review and rework the generated code. Yet over the past months I've noticed a drop in quality… I'm wondering what the cause is. Is the initial generated code so bad that their standards shift without them realizing? Or so bad that their typical amount of polishing effort is no longer sufficient? Either way… this is tedious. I'm gonna need this bubble to pop ASAP. Or even yesterday.