It's not just that social networks go bad the way cheese goes bad, but also that any "social network" business becomes opposed to the interests of its user base at the moment it becomes a "sustainable" business (e.g. the users would want 100% of the surplus to go to improving the system, but management, employees, investors and such won't agree)
Once you are #1 in the Google Economy there are many reasons to say "why try harder?" Questions might have better answers some place else but there is no one site that threatens the SO hegemony. Also, a site like SO has a proven compatibility with Google SEO -- if you made any big changes to a site like that there is a high risk that your traffic would drop catastrophically.
What I would do:
StackOverflow Q/A are all available by a "cc-by-sa" license, so it would be a good starting place for something better: an "antisocial" project curated by a small group of people. Take the top 0.1% of questions from SO and only the best answer and strip away the "socialjunk" (e.g. see "chartjunk") and you are starting to get there.
A system like that could justify itself it were your own personal knowledge base, but to make something generally useful I'd expect to curate 100 questions a day for 3 months or so: about 10,000 questions in all.
There are a number of methods that could speed up that curation a lot, not least of which is that you'll find some authors who consistently write great answers and others who consistently get upvoted with mediocre answers, etc. No amount of analytics will make up for zero manual curation work, you could be maybe 5x as productive once you've curated 10,000 questions and developed some automation.