Larry Seltzer suggests that only magic will bring the money for Philly's project, but that security will sink it: Seltzer raises the very reasonable point that the funding source for Philly's network has to materialize. The plan requires a combination of grants and non-governmental loans to fund the non-profit's somewhat optimistic five-year operating plan. Many companies are bidding, but their bids don't ensure funding, and the city has said repeatedly that they'll give the non-profit their telecom business but no money.
But Seltzer says that's just part of the problem. The real issue he raises is the total lack of security in the network as planned. There's no requirement for encryption of the links, and Seltzer leaps to suggest that unprotected nodes will wind up being launchpads for attacks and spam. They may be unsecured, but they will still require authentication for residential service.
This has been one of my key worries about municipal-scale networks. We know that adding encryption to links reduces throughput by some factor and increases complexity. But without required encryption, you have--under Philly and other cities' plans--tens of thousands of home users with their data totally exposed to anyone near a local node. That problem has already been dealt with: most mesh and metro system can employ encryption. But it has to be enabled; it has to be a choice; and it has to be done.