I think what he argues against is a third culture, where the focus is on implementation of an algorithm in the code, rather than on the algorithm itself -- and

I think what he argues against is a third culture, where the focus is on implementation of an algorithm in the code, rather than on the algorithm itself -- and generally, the algorithm is what matters more. Just a couple months ago I've seen some code that spent most of the time traversing the entire (large) data structure and counting elements satisfying certain conditions, while 5 minutes of looking at the actual math was all that was needed to see that the count was a simple function of known quantities. But, his valid point is indeed obscured by a lot of crap and strawmen :(