This is a great soundbite to build an essay around but ignores two decades of learning, cognitive, and psychological research. There are too many examples to try to pack into a comments, but among the most obvious:
- 3D games are have been demonstrated to have positive effects on spatial perception and to allow rehearsal-based training for real-world activities (http://www.dcs.shef.ac.uk/~daniela/SeriousGames4thePolice_Ah...)
- Medical education in virtual worlds has also demonstrated the efficacy of training in synthetic worlds (http://www.jmir.org/2010/1/e1/)
- Massively multiplayer games' positive impact on literacy (http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.114...)
Easy to go on and on, but the short answer is that it is preposterous to dismiss technologically mediated achievement as somehow fake.
The author's other points about variable-reward based training and the lure of positive feedback are also well researched fields. Since Skinner we've seen evidence that small amounts of positive reinforcement with some randomness -- called variable reinforcement -- generates the largest behavioral changes and longest engagement, so it should come as no surprise that achievement games -- like gambling -- engage a large number of people.
Of course, the 30,000' view of the article -- that variable reward games can consume a large amount of time and may not be making the player's life better -- is true. However, the simple assertions that games are fake and hurt their players are simply not supported by the data. For a very deep analysis of MMO players, you may want to peruse Dmitri William's work on the EQ2 data: http://dmitriwilliams.com/research.html