Net Neutrality is a weird one
While the principle that a company, certainly one that is the originator of a lot of content, can, or should, pay more for better delivery of their content, is fair, it's potentially problematic.
Any company can arrange for a high speed link to the Internet. Throw more money at it and you will get better a better link - lower latency, higher throughput, burst capacity provisions and so on. If a business relies on getting content onto the Internet, then it should have a high level of redundancy internally and should also have redundant (physical) Internet connections, therefore multiple arrangements would be made with different ISPs. This redundancy could be used a a fail-over or even to load balance content, or anything in between.
Few would argue that the above is a bad thing: The company is paying more to provide the amount of content it delivers, which is fair.
In order to improve efficiency, a company can pay even more money to get closer to the core of an ISP's networking setup. Cutting out a few network hops here and there may not seem that important however every network hop adds latency and slows things down overall and when you talk about a high volume of content, this adds to a lot of potential loss of overall speed and a customer's perception of quality is often dictated by speed, or more accurately the lack of (which makes it outstanding that many set top box manufacturer's still push cruddy, low spec, badly programmed kit). An additional point in favour of this arrangement is that by bypassing network hops that the content provider shares with other companies, it will help to optimise throughput and should improve the experience for these other companies.
Again, few would argue that the above is a bad thing. The company is paying even more for its connection and optimising the route is sensible on a lot of fronts.
All this is wonderful if you happen to be a customer that uses the same ISP, or one of the ISPs as the content provider. The delivery of content is optimal and the customer gets a great service. However there are many ISPs and if you're not with one that the content provider uses then you will get a worse service than the content provider would like to deliver.
A natural solution to improve this situation is to enhance the peering infrastructure that ISPs already have between each other, and it's at this point where it starts to get murky and less than ideal. ISPs are not equal, in size or capacities and the choice of ISP is usually constrained by physical location. On the one hand, it is arguably good that a heavy content delivery company would pay for additional peering from their core ISPs to other ISPs - after all they are using a huge amount of bandwidth. On the other hand, this starts to get into the problem of selecting the ISPs to peer with, which will usually disfavour the smaller or regional ISPs, and the actual implementation of the additional peering... if the content provider pays for additional peering bandwidth, then there are few that would argue that this is bad, however if the content provider starts taking a higher share (or priority) of existing bandwidth then there are definite downsides.
In the end, it's all down to implementation and control. I believe that a company should be able to pay for better delivery of their services but that it should be in a controlled and regulated manner, should not disfavour smaller or regional network providers and should not impact other content provider's share of existing bandwidth.