First, I think libertarians like Mark should be siding with the FCC—in this case, siding with the government agency best maximizes liberty.
I’m not exactly an expert on the tenets of modern libertarianism. I assume, though, that reduced government intrusion is a means to an end rather than an end unto itself. That is, maximizing liberty is the ultimate goal, and limiting government is—generally speaking—the way to achieve that end. (I’m assuming most libertarians are utilitarians—my argument doesn’t work if you think government action is inherently unfair).
In the world of Internet access, though, this general principle doesn’t work. Libertarians are going to have to pick their poison. Either the government must impose some basic foundational regulations upon the physical network provider, or the network provider will impose its own restrictions on the application layer (content providers, innovators, etc.). Indeed, if Comcast wins, its market-distorting interventions are just the tip of the iceberg to come.
So pick your poison—the government or Comcast.
To me, this isn’t a remotely close question. Government action in this context furthers the goals of libertarianism many times over by providing the foundation for a nearly infinite number of free markets that have extremely low entry costs and are perpetually contestable. In this respect, open networks are analogous to basic foundational property rights that (I assume) most libertarians concede are necessary to market freedom.
I also want to respond to a couple of Mark’s more specific points about the broadband “duopoly.” He writes that we should try to “remove the regulations that ensure and enforce the duopoly.” That certainly couldn’t hurt—my recommendations would be to provide strong protections for muni broadband and to buy out broadcasters’ spectrum.
But the duopoly (at the physical layer) has deeper structural causes that go beyond regulation. It’s impossible to construct entirely new networks from scratch—there are various barriers to entry, such as enormous fixed sunk costs coupled with extremely low marginal operating costs. In fact, the networks we use today—telephone and cable networks—were publicly subsidized monopoly networks that became broadband networks by accident. It's not something we can repeat today.
For the foreseeable future, then, cable and phone are all we got. Wireless is nowhere close to competing as a substitute service, and probably never will be.
For this same reason, I tend to be very skeptical of arguments that rely on “new networks” emerging. If they were economically rational to build, we’d see them. But they’re not, so we haven’t. More broadly, these arguments tend to ignore the massive structural barriers to entry that exist even without incumbent-protecting regulations.