He claims that algorithms are not so important, but they are what propelled Google above Yahoo and co. Reddit today is very much in the vulnerable position that Yahoo was in. Someone with a recommendation engine that works well could pull the rug from under their feet.
You make a good point, although it is my experience that MIT does not echo Paul's feelings with regards to programming languages. There is obviously the group of hackers at CSAIL (Sussman et al.) that swear by LISP and the functional paradigms in general, but almost every class that involves a large amount of programming is in Java. Even 6.001 is being phased out for courses taught in Python.
As a current undergrad at MIT, I find myself sometimes struggling to determine whether I agree with Kevin or not. It does seem that my greatest selling point is my knowledge and experience with specific programming languages like Java and not any background I have in algorithms, discrete mathematics, etc. I feel that these other tools will prove more useful over the long term, but it's sometimes hard to continue to feel that way when the need for a summer job forces one to put away the common lisp book to read javadocs.
