It would definitely work more like a torrent application than anything else. I'm investigating various P2P systems to extract the bits I like. It will ultimately come down to what's best for what we're trying to do. It will definitely be a hybrid of existing technologies because A) they're proven and B) it's faster to develop.
I have been thinking about it a lot the last couple days, and you're right if enough people in a niche don't grab hold it probably wouldn't be successful, and there is a pretty large gap from one niche to another. However, even given that I think we can make it work, for a few reasons.
It's appealing on a budget level because people will download more if they're not paying anything.
It's appealing on a laziness level because I think people will share more if they don't have to do anything.
We wouldn't be replacing current practices, just augmenting and hopefully enhancing them.
That said, there's still an additional dimension we could take this to which is cloud storage. What I'm thinking on that front is, a premium level membership which is by no means meant to make money, but to split the cost of the cloud storage space/bandwidth and eliminate each participating person's cost for membership to sites like RapidShare & MegaUpload, etc. At the same time keeping the convenience that exists for people that pay for those memberships. Remember that those membership sites charge enough to both make money and use to their advantage the fact that some people are more active and cost them more than others. Some people just feed them money for nothing. This app would be more appealing if you were in the latter group obviously, but would eliminate the profit overhead.
In that scenario, the cloud would hold on to the last x number of items posted and premium members would have access to the cloud as kind of a proxy torrent seed. In that paradigm it could be successful with a small number of participants. Ultimately there's a lot to consider with that kind of approach because you don't want to alienate people and I certainly don't want to provide a platform for elitism.
When you get right down to it though, it's meant to do two things:
Enhance your experience: Notify you when new content of your favorites are posted, downloading favorites with a single click/automatically, enhanced searching capabilities, etc.
Pool Resources: Time, money, work, bandwidth, storage, etc.
I can make something useful work with just two interested and active participants, it's just that the design changes dramatically based on the number of people involved, and how those people plan to use it. That's why I wanted to get a feel first.
Like I said originally though, I can make this application do anything. So picture whatever you'd like to see happen and we can start from there.