You should have some text on the linked site that says how long the video is. A lot of people are going to look briefly, see no apparent description of what you are doing other than the video, and move on. Telling them the video is only 33 second makes it more likely they'll look.<p>On the other hand, there is nothing in the imagery in the video that actually helps understand what you are doing. All the useful information in the video is in the audio. It might be worth spending the two or three minutes it would take to make a transcript and put that on the site, too.
Honestly I'm skeptical. The idea of auto generating groups from pre-existing data isn't particularly new and I'm confident that with enough work it can be done reasonably well.<p>But once you've done that, then every time I post a status update (or tweet, or whatever you end up calling it I'll be forced to consider which circles(s) I want to share with. This significantly raises the cognitive burden of an update.<p>Unless you've got a really clever solution to this problem. In which case you've got something very interesting indeed and you should definitely keep going.
You asked if you should continue working on this. Based on your site, and video, it's hard to say. I really have no actual idea how the site itself is going to function, or what it's actually going to do.
Yes, you should. It's much needed. How you do it will determine your success. I've got my own ideas around this, and played around with it minimally last weekend. The first person that gets this right and can execute will win big.
I think everyone's relying on clever implementations that require people to "just try it out" and give it a fair shake to be convinced that it's better technology. No one will give it a fair shake, why should they when Facebook works fine for most people? There's also no attempt at breaching the chicken-and-egg problem of joining a new social network while one's friends are all still on the old one. I'm not saying I have a solution, but I think most people are looking at the wrong problems.
You could really make waves if you made a way to better organize your friends visually.<p>Option A: People's icons as little boxes that I can drag around. Make a venn diagram for groups. When you friend people, you select their group(s) by dragging their icon to the right circle/shape. A video/slideshare by google on here a while ago showed how friend groups and cliques were more varied than just one big group. Play off that concept and let people create their own boxes.<p>Option B: Create a pseudo-tree of connections. Play to people's egos and show them at the center of the tree. Make the neighboring vertices be groups and let friends branch off the group. The pseudo part comes in when people are repeated within groups: either repeat them as vertices on the tree or have more than one edge coming off. I prefer the later, as it would make for a cooler tree.<p>But really, whatever you make will be better than facebook.
Also, probably should never let other people see a person's groups. It could cause severe problems socially.
I, along with the rest of my team, created the same concept back in 1998 and by the year 2000, we were averaging between 14 million and 20 million pageviews per day which, even in this day and age, is still a decent amount of pageviews!<p>But that's the internet -- everything old is new again. The old. Then new. Then...
I've had privacy circles in Appleseed since about 2006. You might want to check it out:<p><a href="http://opensource.appleseedproject.org" rel="nofollow">http://opensource.appleseedproject.org</a><p>Although the circles aren't auto-generated, the user creates them.
Without a data importer, there seems to be an initial conditions problem, not just from a user acquisition standpoint, but from an algorithmic standpoint. How can it use network analysis to infer cliques when starting with such a small/sparse/nonexistent network?<p>Maybe you could add this to Diaspora. It may be burgeoning, but it's open source and does have users. Also, how does this differ from Diaspora's aspects?<p>That said, yes! There is definitely a need for it.
I've also started working on a similar concept. My approach to this problem is to use context based LISTSERVs. The goal is to help make social sharing relevant by targeting both interests and "social roles".