3 ms·

A potential solution: Implement a comment recommendation system. Based on comments you've liked and disliked, the system will push comments you would like to t

A potential solution: Implement a comment recommendation system. Based on comments you've liked and disliked, the system will push comments you would like to the top. The quality of the community you see is then completely up to you. If you like poop jokes, you'll see poop jokes. If you like pedantic attacks, you'll see pedantic attacks. And if you like insightful comments, you'll see insightful comments. As you stabilize on a community you like, you can start responding to their comments, and then they too will have the chance to vote on your comments. If they like you they'll start seeing more of you, and a micro-community will evolve naturally for free.

ewillbefull 12y ago
If you think about it, it's pretty much the same thing as twitter, except the social graph is built more implicitly. In exchange for a larger selection of relevant content, you have to sift through more noise. But it probably feels better to vote if you know your vote counts differently for some users. You're no longer just one number in a statistic, but part of a much more complicated ranking system. Best of all, your votes significantly affect your own experience on the website. A dynamic social graph like that would probably obviate the need for even categorizing content, let alone "jumping ship" when the community decays. There are some drawbacks though. If you implemented it in reddit/hacker news, there would no longer be a "front page of x" to serve as a symbolic milestone for your submission. Without context, "2000 upvotes" has ambiguous importance. Also, a different ranking system exists for each cluster in your social graph, which sounds like a fun engineering problem.