Instead of waiting for carriers to uniformly open up GPS triangulation to consumers through their APIs, he encouraged developers to take advantage of the falling prices of GPS components. They should consider using a pocket bluetooth GPS to interact effectively with their cell-phone hosted location-aware UGC application.

What kind of applications?

1. Identify unknown images in a photograph using historic tagged metadata. Ajit describes a related thought experiment on his blog, but the gist of it is that you can search a photo website for previous photo tags at the same location to determine blurry features in your own picture.
2. Infer memorable events by looking for data thresholds. This is a bit of a thought experiment as well and certainly unnecessary given the invention of AM radio, but hear me out. I’m not intimately familiar with Flickr’s API, but imagine the following:
   • Let’s assume we live in a world where photos are instantaneously uploaded to a photo website from a wireless capable camera.
   • You setup a location-sensitive “trigger” to send you an SMS to your cell phone Twitter-style.
   • You setup a Flickr “rule” like “Send me a text message when Flickr receives 500 photos in a 10 second interval from AT&T Park (where the San Francisco Giants play baseball) over the next 10 days”

   You’ve just subscribed to be the first one to know when Barry Bonds beats Hank Aaron’s home run record! Furthermore, you could subscribe to the feed associated with those pictures.

   I believe that this concept is powerful and can be more generalized. I could subscribe to any Flickr photo feed when a certain threshold of pictures taken has been exceeded in a short time interval. Even if we remove that futuristic assumption that photos are immediately beamed to Flickr, the rule is still useful as long as the GPS coordinate and timestamp are preserved upon upload.

3. Integrate olfactory devices and GPS (I am half joking here). By mashing up smells and locations, one can broadcast the existence of restaurants, gas leaks, fires, etc. It sure smells like UGC

To summarize, by capturing aggregated user generated content in the form of tags, time, subject context and smell (!), we can infer potentially useful information.