Without estimating the traffic of the author's site, I still find his argument valid.

I've employed many of the tactics he describes and have achieved >15 M/s of dynamic content through MySQL. Not Facebook, but it was fast and scalable for the market my app was geared towards.