There is a well known phenomenon of evolution. If you get an island, or a bunch of islands, and you stick some species on them, you'll tend to get fairly rapid divergence. You get an island group full of white grizzly bears, for instance. You get the Galapagos.
As I understand it there are basically two things going on: small populations lead to a lot of inbreeding which brings out recessive traits; and the separation from other populations causes the recessive traits so brought out to stabilize. With a bunch of islands and a bunch of small populations, you get different families of traits brought out and stabilized. Larger populations in large environments tend to produce homogenization. If you take a whole bunch of dogs, it doesn't matter what breeds or mixes, and let them run wild for a few generations you get a 40 pound animal, yellow-ish brown, medium length coat, with a moderate length tail that curves up toward the head. Every single time. If you get a small population of three or four dogs and stick them on an island, you're gonna get something else. Maybe border collies, maybe no dogs at all in a couple generations. Probably not the generic yellow dog, though.
Let's think about social norming and pictures for a minute. If you have a huge population of ideas about pictures, let's say flickr or 500px, they're going to converge on the generic 40 pound yellow dog pretty fast. We've seen it happen. The same little family of extremely pretty and emotionally sterile pictures starts popping up and then the taste of the population as a whole is defined, and we're pretty much done. Ugh.
Let's think about the curation problem, too. Digging out good work from flickr is a nightmare. It's buried under snapshots of lattes and purple sunsets, by the billion. flickr is too much like a giant pile of pictures.
So here's a free idea for a startup, that addresses both of these problems, creates a potentially interesting new paradigm for social networking, and has a lot of revenue potential, as these things go.
The cellular social network is built around the idea of a group of islands. Each island is, well, let's call it a mini-network. A mini-network is a standard social network, but with limited membership. Within the network all the usual things apply, one can follow people, like, +1, whatever. Perhaps you can share photos, or comment on one another's walls, or share video clips, or broadcast snarky little messages of 141 characters or less. If you're smart, you'll build the infrastructure to support anything at all. There's a cellular social network infrastructure, and you can build photo-sharing, video-sharing, twitter-alikes, facebook-alikes, whatever you like on top of the cellular structure. You gotta have a story for mobile, it can't just be the web, blah blah blah. The usual.
The conceit is that the mini-networks have limited membership. Perhaps 100, perhaps 1000. A person, an identity on the larger system, can belong to a limited number of mini-networks, perhaps 2, perhaps 50. So, there's some cross-fertilization, but most of the activity takes place inside the mini-networks. Identities have limited visibility to mini-networks to which they do not belong, perhaps they can browse "public" content, but cannot contribute. Perhaps they can only see a random sample of content. Whatever. Make that configurable too.
There needs to be ways to explore mini-networks that you do not belong to. There needs to be a way to enter and leave a mini-network. You might want the ability to join a waiting list for a mini-network. Mini-networks might have a "fission" operation in which they split in two, with some rationale for assigning members to the new mini-networks.
Most mini-networks will simply die on the vine. That's ok. It's a few rows in a database somewhere, so what. The point is that the ones that work out will, ideally, evolve unique identities. A photograph sharing cellular social network would, ideally, generate mini-networks with distinctive artistic visions. The social norming within each would not, ideally, stabilize on the same half dozen pictures, but might stabilize on a variety of different things. By browsing, new members could find mini-networks with compatible visions, and could contribute in interesting ways. Because the populations are small, individuals have the power to move the social norms by more than infinitesimals.
Recessive memes, in the original sense of the word "meme" not in the sense of a picture with a stupid caption, can be expressed, nurtured, and stabilized within a mini-network in ways that the cannot in a giant singular social network.
How does this monetize? What do you tell Y-Combinator? The mini-networks form affinity groups that can be targeted for marketing, obviously. A cellular social network for creatives (photographers, etc) could sell access to talent scouts, giving them deeper browsing and search capabilities. User-generated content produced by highly specialized affinity groups is obviously a lot more valuable, and easier to find, than user-generated content in a giant monolithic network. Connections between affinity groups can be mapped (the BMW-lovers mini-network has a lot of members who belong to the Bi-Rite ice cream lovers mini-network... interesting!). And so on and so forth.
This model, applied to photography in particular and art in general, helps to render the curation problem more tractable. 1000 mini-networks with 100,000 pictures in each is a lot more tractable than a simple pile of 100,000,000 pictures. This self-organization contributes both to monetization and to curation.
Ok, so, someone. Go out and build this thing.