Having spent the last 8 years building a Netflix competitor, I think you overestimate the commoditization of VoD tech. You need a lot more than AWS + Akamai. The costs of those are too high for the types of margins a Netflix-like service has. Also, you need apps everywhere, and those apps have to be good. Even though it's getting a lot easier every day, it's still way outside the reach of any individual show, they're just not structurally set up to build and manage the tech side. HBO can do it because they are just barely big enough, but individual shows will never jump that canyon.
What's going to happen is not that shows replaces channels, but that services become the new channels. Instead of having a few cable companies monopolizing the last mile, building and running all the distribution tech, and charging an arm and a leg for access, the new model is that there are no gatekeepers on distribution. But that is not to say distribution is a commodity or easy, in fact doing good distribution ala Netflix that cuts across devices, geography, and a long tail of content will become an essential value-add and it will be more competitive than it's ever been.
Having also spent the last decade or so building a similar service, the tech side is commoditized by at least a half dozen companies. The ingest workflow is still somewhat manual, and there's a lot of QA that has to go in to your metadata, but the tech side (transcoding, CDN delivery, etc) are commoditized -- especially if you're a network that already owns the rights to your content.
This is all headed towards a world where the big ISPs run CDN nodes and all the channels buy delivery from them not because they force them to, but because the big ISPs can run their CDNs with a lower cost base (they already have the interconnects and disk space is cheap). The gatekeepers will still exist because they own a market control point -- you didn't think it was going to be that easy, did you?
Apps are an issue, but they're not that cumbersome to create, even cross-platform. I was able to hack together a prototype iOS app that played back video from a CMS attached to a CDN reasonably well in about 4 hours -- and I'm not even a professional developer. For the long tail, there are always services like YouTube and Vimeo -- they're a better fit than a custom app anyway because you get access to a larger audience.
Your response is orthogonal to my main point—which is that services will be the new channels, and they will have much more power than they did under the cable regime.
The mega-services at Netflix/Hulu scale will definitely by CDN capacity direct from ISPs, but that in no way means that these ISPs are gatekeepers with anything near the power that is crumbling from the grasp of the cable companies today. That would only be the case if the ISPs defeat net neutrality and become as evil as possible about restricting outside content to reach their users, and frankly if they twirl their mustache that hard consumers will revolt and regulation will come down on them like a ton of bricks.
What's going to happen is not that shows replaces channels, but that services become the new channels. Instead of having a few cable companies monopolizing the last mile, building and running all the distribution tech, and charging an arm and a leg for access, the new model is that there are no gatekeepers on distribution. But that is not to say distribution is a commodity or easy, in fact doing good distribution ala Netflix that cuts across devices, geography, and a long tail of content will become an essential value-add and it will be more competitive than it's ever been.