The author of that article doesn't have much of a clue about efficiency.

Some algorithms scale well, some don't. The ones that don't can scale really, really badly. So much so that even a speed increase of 2^30 isn't going to help.

Failing some fundamental shift in the nature of reality, thinking about O(f(n)) will always be important - even if Moores law continues until the end of time.

This kind of shallow technobabble irks me. As a computer scientist you have to read things from a position of skepticism. There are good reasons for this.

-Andrew.