Keep in mind that some of the dynamic shifts I am concerned with are not necessarily risks associated with anticompetitive or otherwise “bad” behavior on the part of infrastructure providers. (Those risks matter too, of course.) Acting rationally, infrastructure providers will optimize their technologies, contractual relationships, and offerings to maximize their returns. Some of the troublesome shifts flow from the impact of differential pricing on how users/consumers behave.

I understand your position (re. AT&T vs. the FCC) and did not mean to suggest that you were making the argument about the need to discriminate and prioritize to supply better products and services.

Can you elaborate on this point a little bit? Keep in mind that I don’t have any contractual relationship with Comcast. TLF is hosted by my friend PJ Doland, who pays an ISP called Rackspace for his connectivity, who in turn has a peeering agreement with a backbone provider like AT&T, which in turn has a peering agreement with Comcast. Comcast never sends a bill to me, to PJ, to Rackspace, or to AT&T.

PJ chose to use Movable Type to host TLF. Are you saying that Comcast could somehow use its influence as a broadband ISP to encourage PJ to use another blogging application like Word Press (or perhaps a proprietary Comcast-developed blogging tool) that it liked better? How would it do that?

The arguments in favor of no regulation often focus on the perceived need to discriminate and prioritize to supply better products and services.

I know a lot of critics of regulation make this argument, but I don’t think I ever have. I’m a supporter of network neutrality as a technological principle. I’m just more afraid of the FCC than I am of AT&T.

But my primary point is rather simple (although perhaps disconnected from your specific concern about political content manipulation). The arguments in favor of no regulation often focus on the perceived need to discriminate and prioritize to supply better products and services. Well, it goes both ways. The common refrain that “the market” will discipline infrastructure providers that do not give consumers what they want is overstated (for a variety of reasons Barbara van Schewick and I explain in a forthcoming paper), and even if true, it misses, what I think is, the more important point: consumer willingness to pay for specific types of applications or content may be a very poor guide for allocating infrastructure access (and from a longer term dynamic perspective, structuring the online environment).

I don’t think this is quite on point. This might be a way for the ISP to increase revenues (although I’m skeptical about that) but it wouldn’t help the ISP “shape the environment” in terms of the content of blogs. Liberal and libertarian blogs could pay Comcast’s toll as easily as conservative blogs could. Indeed, it seems likely that Google and Six Apart would sign a deal with Comcast to give everybody on Blogger and Typepad priority access, so the average blogger on those services wouldn’t have to lift a finger to get fast-lane service.

Or blogs who are on Kos’s blogroll (or instapundit’s) get priority routing, and others do not.

This would mean slowing down 99.999% of the blogs in the world. Sports blogs, cooking blogs, MP3 blogs, etc, would all grind to a halt. Millions of blog readers who don’t read a single political blog would be inconvenienced.

Moreover, unless you’re proposing that they’d slow down the entire web other than Kos’s blogroll, you’d have to come up with some way to distinguish a “blog” from a generic website. Is the New York Times a blog? Is Ars Technica? What about sites that use MT as a CMS but don’t actually have a blog format? And how long will it take people to come up with WP and MT plugins to disguise the blogginess of a user’s website?

All of these would be fairly simple to implement; all would empower a specific agenda.

[Note that in practice, as I believe Ed Felten has pointed out, what this would mean is slowing down the 'others' instead of speeding up the privileged sites. But basically the same difference.]

Imagine I’m the CEO of Comcast. I want to use my control over infrastructure to push some particular agenda on my blog-reading customers. How specifically would I go about it? What routing policies or other infrastructure changes would I implement?

I agree with you that context matters and that abstracting too much from the context may be misleading. But focusing in on the existing set of blog offerings and what would happen if one suddenly disappeared also misleads. Yes, some users would switch ISPs if their chosen blog is not available. (We could open up a debate on the effectiveness of switching as a means for disciplining ISPs, but let’s not for now.) The point Benkler and others make is a dynamic one about how the information-cultural environment is shaped. Pricing infrastructure access on the basis of identity, use, or application will shape the availablity and diversity of different users, applications and content. (That is one of the main reasons advanced for opposing net neutrality regulation, right?) It seems to me that Benkler is on solid ground in arguing that infrastructure owners have both the ability and incentive to shape the environment is a manner that makes it look more “push” than “pull.” To my knowledge, Internet technology is not inherently “pull” rather than “push.” It is is malleable.