I have generally been a fan of Kilcullen's books and articles, although I found this particular book not as well thought out as his previous books. Perhaps because he is still exploring this concept. To be fair Ralph Peters, and many others, did meet him to the punch on this issue, and I recall a number of discussions and papers in the 1990s discussing potential military scenarios in large urban areas and how complex they would be.

http://smallwarsjournal.com/jrnl/art...-and-dangerous

Peters' wrote in 1996:
Cities always have been centers of gravity, but they are now more magnetic than ever before. Once the gatherers of wealth, then the processors of wealth, cities and their satellite communities have become the ultimate creators of wealth. They concentrate people and power, communications and control, knowledge and capability, rendering all else peripheral. They are also the post-modern equivalent of jungles and mountains--citadels of the dispossessed and irreconcilable. A military unprepared for urban operations across a broad spectrum is unprepared for tomorrow.
Peters made a lot of interesting points regarding the future of urban warfare in a talk I attended in 2001 (before 9/11), but most of the points focused on the physical aspects of fighting in a city and the difficulty of templating irregulars. Kilcullen adds the socio-political aspects and I found his thoughts on the city as an ecosystem (system of systems) and how competitive control works very helpful in observing and explaining what many of us have experienced and simply labeled it as chaos, yet there was an underlying order that wasn't necessarily visible to us at the time.

On the other hand his book in my opinion is still is missing the so what factor for security planners. He is also focused currently on data analysis to analyze cities which may prove to be valuable, but similar studies in the past have generally led us astray. Ralph Peters on the other hand glances over the socio-political and focuses on the so what at the tactical and doctrine level. Curious about readers' thoughts on his projections made in 1996 as a Major now that we have extensive experience fighting in cities (though done would qualify as a megacity the principles still apply). regardless there is considerable room for further study in this area to inform military doctrine and future capabilities required.

How, or even if, Brazil can secure the games will be interesting to see unfold. We could be overstating the threat by assuming the masses will think and act collectively and actually have an interest in attacking the games. I'm sure some do, but what percentage? Is it enough to be threatening? How good is Brazil's intelligence in identifying the leaders who could provoke a serious security threat? What is their ability to pre-empt it? Will they leverage engineers to create obstacles and channelize potential protesters / trouble makers into zones they can control? What will the impact be if the disenfranchised citizens effectively disrupt (or worse) the games?