On: An Argument for P=NP (2005) [pdf]

I think it is rooted in a common fallacy -- inability to intuitively understand exponential behaviour. In can be something common such as compounding interest on a loan or something like this related to computational problems.