Maybe we should make a list of "common knowledge" (topics that have been discussed to death and for which there seems to be a pretty strong consensus) and then refrain from posting/upmodding articles about those subjects if they don't bring something truly new to the table...
Here's a start of the list:
Ideas don't matter, execution does.
Stop thinking about doing and just start doing.
Release early, release often.
We might live in a computer simulation.
Unit-tests let you make change to the codebase while remaining reasonably confident you didn't break anything.
It's very hard to start a startup if you still have a regular job so consider quitting.
News.yc is, among other things, an educational resource -- and teaching is about saying the same things over and over and over, to a new group of students each time. That's how it works.
If you want to read a site where "we've already discussed that, so you can't publish it again" is a valid criticism, you need to switch to reading scientific journals. I don't think that the alternative -- trying to turn news.yc into a site that behaves like a scientific journal -- is going to fly. We don't have a required curriculum that prospective upmodders must read. We don't make them pass an entrance exam. They're allowed to just turn up and vote. That's how this system works.
On this and other sites, I've always craved a moderation system whereby I could nominate a certain collection of modders whose votes I trust above others. For example, if I could click a button and see what news.yc would look like if only the YC class of 2007 were allowed to vote, I might be able to filter out the articles which were of interest only to freshmen in CS. Unfortunately, I've never really seen such a thing. Obviously, allowing an arbitrary set of such queries would risk creating a huge scaling problem, but you'd think that a site could offer a choice of two or three ways to weight the moderations. You could divide the incoming readers into "cohorts" based on year of arrival (or even other demographic information) and offer options for reading social news within your own cohort.
How are the new folks to know what ideas have been discussed to death? And, sometimes there is a delta of thought that does add value, specifically for frequently/commonly discussed topics.
Here's a start of the list: Ideas don't matter, execution does. Stop thinking about doing and just start doing. Release early, release often. We might live in a computer simulation. Unit-tests let you make change to the codebase while remaining reasonably confident you didn't break anything. It's very hard to start a startup if you still have a regular job so consider quitting.