Yea, that's because 80% of the work is getting to that level of quality and even most human devs struggle to get there, even with the necessary resources. So using a LLM which is trained on the absolutely average corpus of code produces the same level of average. Getting to that noticeable level of quality is difficult and time consuming and LLMs do not help you get there. But much like everyone thinks their baby is beautiful, they think the code they create is the same. And most of the time it isn't. Sorry about that but if quality was common, it wouldn't be called quality.