In major markets, historically it has been the networks who pay the stations an affiliation fee, not vice versa. The stations have something valuable (a broadcast license) and the networks want access to it so they can air their national ads.
Yes, of course the station also gets something valuable out of the affiliation. On balance, however, the money has tended to flow from network to station, not vice versa. For example, here in Boston, NBC pays Channel 7 an affiliation fee of $15 million a year (or did in 2000 and years afterward; I haven't kept up with it more recently).
See http://ocw.mit.edu/courses/electrical-engineering-and-comput..., section 2.2 (page 4).
I don't see how local affiliate fees themselves are directly at risk for the broadcast networks. The local OTA broadcasts of the public channels are generated by the local affiliates, who contribute local programming to the broadcast, such as news. The cable companies therefore have to pay the local affiliates, who in turn pay the national broadcast networks.
The local/affiliate fees are only the proximal revenue source for the broadcast networks - the original source is the cable fees.
I'm not sure what the point would be of a large number of independents... in 20 years there may only be PBS affiliates out there and empty spectrum. Where I live there's only enough independent viewers (mostly syndicated) to keep two quasi independents alive, so tossing another 4+ onto the fire might well result in all of them going out of business.
The entire premise behind affiliate television is that local station owners can license the rights to national programming, and then exercise a quasi-monopoly over distribution of that content to local markets. (If I'm the FOX affiliate in NYC, for instance, I pay FOX for the rights to its shows, and I have exclusive rights to air those shows in NYC).
This business model was threatened the minute the internet took hold, really, because the very idea of local exclusivity is anachronistic now. It's one of the two horse-and-buggy concepts of the 21st century content distribution business (the other being time-bound content viewership). It is threatened much more extensively by Netflix, by YouTube, by Hulu, by Microsoft, by iTunes, etc., than by any one startup. It is threatened by virtue of its irrelevancy in the modern, globally connected era.
The idea of moving television networks to paid subscription models has been bandied about for at least a decade now. IIRC, it was NBC who first proposed this idea in the early 2000s. Indeed, this may come to pass at some point: not because of the likes of Aereo, but because broadcast television starts to make less and less business sense as time goes on (whereas content ownership and distributorship is still viable). Eventually, we'll probably have nothing but content shops, who distribute their wares across all channels -- perhaps offering different tiers of the products at different prices in different channels, and/or extracting licensing fees to different distributors.
As a consumer, expect a generally upward trend in content pricing in the long run (i.e., free to paid). It'll just depend on whom you're paying. Advertising-subsizided, free content will still exist to the extent that distribution channels like YouTube continue to thrive.