> There is still potential for mass-fraud, but repeated seemingly uncoordinated reports improves credibility Hmm. That would be an interesting piece of fraud (

> There is still potential for mass-fraud, but repeated seemingly uncoordinated reports improves credibility Hmm. That would be an interesting piece of fraud (oops I mean public art) to architect. Make a Blender scene of Central Park, populate the scene with bystanders with phones (yes, this will take time to source realistic models and skeletal animations). Each phone gets to be a camera in the scene. Not all cameras start and stop at the same time, and a few of them will have to capture video before and after the event without the actual event included. Timestamp and tag the videos appropriately, including clock drift, then have fun with avconv/ffmpeg converting the formats to different video formats based on known phone sales and what default formats iOS/Android encodes to (ugh, I actually had to do this once back when zencoder/pandavideo). Create twitter accounts in advance, make them look real by copying other twitter users, and the day of the event just start uploading the youtube videos all at the same time and get your fake tweets flowing. Now, what could this pipeline be useful for? Suppose you know that the videos are fake, but also suppose you can be reasonably sure that others will accept this data as real news. What could you do with that? First idea is some killer marketing stunt..