The buzz is due to a recent paper from Vinay Deolalikar. I've not read the paper and honestly I don't think I have enough background for it. This seems to be a very complex work. From a quick read over the Introduction, the author says that theories from several apparently unrelated areas are used throughout the proof: logic, statistics, graphical models and others. And the proof claims that P!=NP.
It's important to notice that several other researches have tried to solve the P versus NP problem. Take a look at this page if you wanna know more about it: http://www.win.tue.nl/~gwoegi/P-versus-NP.htm