I didn't get much out of this. Their traffic numbers are unremarkable. There's the standard "databases are hard to scale" meme, and their software stack is also pretty standard. Gnip as a business are interesting, though: there's a ton of money to be made in data integration, ETL and related services.
I'm not but my friend works for them and I get the concept - it's like feedburner for APIs. So instead of individual web apps polling twitter a few times an hour, for example, you subscribe to gnip and it pings you (gnip is ping spelled backwards) when there's an update. It turns your typical app's pull request into a push instead, basically.
Yes, if you've ever tried to create a mashup using data from Twitter, digg, delicious, etc., you know the problem is that you have to poll their entire data set (without exceeding their api request limits) to get at the data you are trying to track.
With Gnip, you get an abstraction layer so you can create a filter based on usernames, for instance. When these users do something, like tweet, we post that data back to you.
In this scenario, you no longer have to poll arbitrarily for needles in haystacks--we push the needles to you when they occur.
Hope that helps! If anyone is interested in utilizing data from the social graph to create their app (or if you are interested in publishing data you have for public use), we have a Group list: http://groups.google.com/group/gnip-community as well as documentation on our site, www.gnipcentral.com and a set of client libraries (we are currently evolving) to help facilitate such endeavors.
I'm have a CS degree and am currently building an API heavy application, and yet when I looked at GNIP I could not understand as well as your previous paragraph. Lol, you should e-mail your explanation of their company to them.
