Shannon's paper would be a good example of a great math paper that introduced deep, original ideas -- many of the higher-rated papers are good, but not original work.
I love reading the works of original authors. For example, there's a Dover edition of Georg Cantor's Theory of Transfinite Numbers which gives essentially the original development of ordinal numbers, which is based on (Cantor's own) development of set theory. Another original I recommend is Conway's On Numbers and Games, although that field (surreal numbers) has not been as influential. I think it's underrated.
About Lockhart's Lament: This guy sounds a bit angry about things. I'd love to see change in math education, but I'm afraid that basing it on strong negative emotions is not the best approach. In more detail, I'm not completely convinced that he's isolated the real source of trouble in the US math education system. A champion for math education reform I admire is Keith Devlin (http://www.maa.org/devlin/devlin_03_11.html).
Scroll down to find the (IMO) more interesting papers. As mentioned in a few comments on the linked page, people tended to submit and upvote very meta articles.
Yes, but MathOverflow is a website for research mathematicians. Studying mathematics after college is what most of the people on that website get paid to do.
I love reading the works of original authors. For example, there's a Dover edition of Georg Cantor's Theory of Transfinite Numbers which gives essentially the original development of ordinal numbers, which is based on (Cantor's own) development of set theory. Another original I recommend is Conway's On Numbers and Games, although that field (surreal numbers) has not been as influential. I think it's underrated.
About Lockhart's Lament: This guy sounds a bit angry about things. I'd love to see change in math education, but I'm afraid that basing it on strong negative emotions is not the best approach. In more detail, I'm not completely convinced that he's isolated the real source of trouble in the US math education system. A champion for math education reform I admire is Keith Devlin (http://www.maa.org/devlin/devlin_03_11.html).