He just said today he was going to unilaterally start changing immigration policy. He's been behaving like an autocrat for quite some time now. Your X is bad therefore Y is good logic does not work, they're both the wrong lizard.

Is this a republicans vs democrats thread in disguise? Just because one side of the discussion is arrogant doesn't mean the other is not. Google has a long history of failed projects because they're not afraid to over promise and blindly charge into a project. I think the ignition recall is a good illustration that the automotive industry doesn't have that luxury. My Google TV appliance, which is now an abandoned project, isn't going to kill me. An abandoned self driving car project might, even if it's not my car.

The logic is: "People have been wrong about things for hundreds of years. You are a person. Therefore, you are wrong."

So better performance and lower costs are not a benefit to the ISP and its customers? I'm not speculating here, just yesterday we deferred a project costing $300,000 because installing a Netflix OpenConnect cache dropped our peak bandwidth usage enough in the market in question that we can afford to wait. Now we can wait a few months, maybe to 2015 and spend that money on other projects.

It's meant a little more solar and a lot more coal. Fear of the unknown has led to more pollution.

I'm speaking from experience here. I've seen the drop in peak bandwidth before and after deploying these. We spend hundreds of millions a year upgrading our network, so a free box that eases the pain is an easy call to make.
Netflix provides the appliance for free. The space and power costs are pathetically small compared to the benefits in increased quality and decreased expense and capital upgrades.
Also, they didn't offer this only to Comcast. They offer this to every ISP, see the link below.

https://www.netflix.com/openco...

What cost? Space and power? Installing a caching appliance will save thousands of dollars per year, maybe per month, for any ISP of any significant size. So, maybe an extra $10 a month on the electrical bill is balanced against better service to your customers and reduced network expenses, and the capital cost to upgrade routers and transport.
This is a no-brainer for any ISP concerned with providing service, rather than 'monetizing' their customers.

Comcast plays ball with Google because if they impair Google then Comcast's large enterprise customers will raise hell. They can get away with squeezing Netflix because nobody uses Netflix at work.

That's like saying physics is useless because there's no such thing as a perfect sphere with no friction. The Free Market alone can't fix everything, but neither can command economics.

That's news to me. I don't see why Netflix would care where the residential customer is located. They can geolocate IP address regardless of the peering point to block international traffic as needed. I'm certainly transporting traffic a lot farther than across IL to get to peering points in Chicago, Atlanta or Dallas. If you check peeringdb.com, Netflix doesn't have a peering point closer to Southern IL that Chicago anyway. They don't look to be in St Louis or Davenport.

There are two main answers to the second question.

Cable ISPs are originally video providers, so they have a financial incentive to impair an online video service. FTTH services are also big into Triple Play so that's why Verizon and AT&T are being jerks. If you look at traditional DSL companies like Frontier, Windstream or Centurylink they are not rattling their sabers about how Netflix is 'using' their network. These companies only sell video by bundling Satellite with their service, and they only do that to compete with Cable on Triple Play. God knows they are having trouble in the copper last mile, but they're at least trying to do the right thing by their customers on the peering side.

Also, online video is the largest source of bandwidth usage, and Netflix is the largest source of online video. Impairing that traffic, causing your customers to drop down from HD to SD resolutions, reduces your network load and lets you slow down upgrades elsewhere on the network. Choke at the peering point where network is cheap, save money in the regional transport where it is expensive. It doesn't even require special traffic shaping routers. You know everything on those peering links is video, just stop upgrading them.
Hulu is a lobotomized alternative pushed by the television networks. They dont have enough money to squeeze or enough traffic to be a problem.
Amazon and Google generate tons of business traffic as well through cloud services and of course the search engine so impairing those ASes would piss off the most valuable customers the ISP has.

So that's why Netflix is singled out. Hulu and Vimeo and so forth are small potatoes, Amazon and Google are not to be #$%&ed with.

What? I work in the industry too, our network has multiple dedicated 10GE peering ports with Netflix in every major IXP where we have a presence.
Netflix is easy to work with on peering because it's very much in their interest not to use Level3, Cogent, or other transit providers at all.
The point about Netflix using transit providers that are relatively more expensive to the ISP on the other end may be valid, but is any ISP with more than a few thousand customers should be peering directly with Netflix anyway.
https://www.netflix.com/openco...

Netflix does have a CDN program. They will provide a caching appliance free of charge to ISPs which will immediately reduce the load on that ISPs network. The only reason not to participate is if the goal is not to provide service and reduce costs, but to artificially choke back Netflix to make the ISPs own video product more competitive. The Open Connect appliance is actually a pretty cool design.
https://www.netflix.com/openco...

Replying to my own comment here, but Content Delivery Networks aka Caching is also a win-win for everyone. It keeps IP traffic local and cuts down on the amount of bandwidth that has to leave the ISPs network and burn up transport bandwidth and possibly also increased transit costs. The customer gets faster service, the ISP gets reduced costs, the Content Provider has a better product. This is also something we need more of for the Internet to continue to grow.

I agree, this is really about the ISPs actually providing the product that they've sold and there's no need to get into what 'should' happen or what people 'deserve'.
I wouldn't put too much weight on the article author's description of how the Internet works. He gets some of the concepts right, but the implications wrong.
Peering is a win-win for absolutely everyone. It's not preferential treatment, it's a way for two networks to reduce both of their IP transit monthly bills. We don't need less peering, we need more peering. The only traffic that should be hitting paid transit for an ISP are packets heading for smaller networks and the other side of the globe, which are not within reach to peer with.
The US network is built on direct peering, it wouldn't work at all without it. We are slowly catching up to the EU where peering fabrics are more popular. This means that an ISP can use one port to peer with dozens or hundreds of other networks.
Peering doesn't disadvantage smaller ISPs and content providers, because it's still more affordable for them than buying transit.

They targeted both right and left wing groups critical of the adminstration? Oh well that changes everyfuckingthing. Thanks for pointing this out.

Is God confusion? That would explain a lot.