Currently, The Economist is the only news paper I read on a weekly basis. Generally, the writing seems more informed by sound social science and a historical perspective than The New York Times or The Wall Street Journal.
I do like the economist, but they suffer from a recent graduate syndrome I've become increasingly aware of as the blogosphere allows the sort of person who forms the majority of their staff greater access to the limelight, the very convincing on areas you know nothing about, horribly amateurish on areas you are familiar with, concatenate the last three AP reports with a dose of the old Washington Consensus style. It's a style formed by throwing in enough second hand research that one's editorially mandated conclusions seem authorative, which can see it in ...
I generally agree with Bryan Caplan that news is generally useless:
However, there are some analytical sources of news that are exceptions.
Personally, what I find most stimulating is http://www.mckinseyquarterly.com/home.aspx . http://www.cato-unbound.org/ is also quite stimulating, although it isn't contemporary news. Then on the more non-news-y side, IEEE Spectrum and Communications of the ACM can occasionally be great. And while Scientific American has dumbed down in recent years, the featured articles in Science News (http://www.sciencenews.org/view/feed/type/article) are *amazingly* insightful. Two examples are below:
http://blog.sciencenews.org/view/feature/id/330593/title/Inside_Job http://www.comoogl.sciencenews.org/view/feature/id/74456/title/Healthy_Aging_in_a_Pill
You usually don't even get analyses like these in Science or Nature.
And http://www.newgeography.com/ is surprisingly insightful.
What do people here think of The Atlantic? It's better than other news sources, and I was impressed with it when I was in 8th grade. And some columnists, particularly Reihan Salam, are great (Reihan isn't in the position of supporting *any* politician or party, so that puts him in a position where I don't suspect him of twisting facts [even a little bit] to support his position - something that I do often suspect for people like Ezra Klein and other left-wing pundits). But even The Atlantic does seem to be too much in line with mainstream opinion.
Other than that, I get most of my analysis from Gene Expression and Information Processing. I'm not too much of an economist yet, so I don't really read the economics blogs. FuturePundit/ParaPundit is right on many things, but its blogposts don't contain too much depth in them.