There's a huge financial incentive to solve this problem; it would significantly improve the efficiency of many business processes and it would make it trivial to crack many contemporary forms of encryption. The solution is worth billions, if not trillions. Furthermore, you've only got to find one example where P=NP and you can reduce many other problems to this case. Furthermore, there's a very large number of abstract and practical examples to work on. So why hasn't it been solved?
Has it been solved in secret? Are we not smart enough to solve the problem? Is there little incentive to prove that P!=NP? Do foolhardy attempts produce "good enough" improvements to existing practices?
Furthermore, you've only got to find one example where P=NP and you can reduce many other problems to this case
Except that the answer is quite likely P != NP. Rather than characterizing the rewards in financial terms, IMHO the more important thing to be gained is knowledge: the answer to a fundamental question in complexity theory (perhaps the fundamental open question).
Has it been solved in secret? Are we not smart enough to solve the problem? Is there little incentive to prove that P!=NP? Do foolhardy attempts produce "good enough" improvements to existing practices?