Over time the readership and topics covered of HN has grown and maybe the quality has gone down. Perhaps one way of solving this problem is to find a way to present a smaller subset of data to different groups so that they can focus on the areas they like?

My favourite kind of posts are big data, web app architecture and functional programming. If my view of HN was skewed by my interests, then my perception of the quality of HN would improve. Chances are that other people would fit into my interest groups and hopefully that would promote tighter, smaller groups and perhaps the quality would increase.

Stackoverflow's tag system seems to work well, but assigning tags would introduce extra burden on people (and another area for disagreement). I like the idea of a machine learning approach where I see articles based on the content on previous articles I enjoyed (via voting). The same could work for comments. Going further, perhaps this could even suggest people with similar interests to yours.