[lillecarl]

YouTubes contentID system is quite good at detecting just a couple seconds of sound to determine if something is copyrighted or not. It should be quite easy to match copies of content and just get rid of it straight away, meaning they don't have to see duplicates. I would assume they already do this. Which leads me to believe there's actually that much content being produced, I imagine most of it is not produced in the West though.

[larrywright]

There is a national database of child porn content that most big companies that host video and photos use (Facebook, Twitter, YouTube, et al). It allows them to filter the known content. I don’t know how fast that gets updated though. The larger issue is the new content. Facebook makes it very easy to take and share video, and makes producing and sharing this content easier.