I find the whole “ten hours of noise” YouTube genre both amusing and disturbing, mostly because of the compression and bandwidth implications. The Shannon entropy of such a stream is huge, while its Kolmogorov complexity is (hopefully!) tiny: the whole thing seems like a highly perverse use of streaming data. But, it would make a great venue for steganography.
For a site like lobste.rs I would rather have the submitter do some extra research and find the paper/post/article that discusses how the filters matched this as copyrighted content. There is one throwaway line in the article about how other copyright holders had uploaded white noise streams of their own. But there is no explanation about why they did it, what hypotheses there are that white noise against white noise would pass the filters etc.
I find the whole “ten hours of noise” YouTube genre both amusing and disturbing, mostly because of the compression and bandwidth implications. The Shannon entropy of such a stream is huge, while its Kolmogorov complexity is (hopefully!) tiny: the whole thing seems like a highly perverse use of streaming data. But, it would make a great venue for steganography.
For a site like lobste.rs I would rather have the submitter do some extra research and find the paper/post/article that discusses how the filters matched this as copyrighted content. There is one throwaway line in the article about how other copyright holders had uploaded white noise streams of their own. But there is no explanation about why they did it, what hypotheses there are that white noise against white noise would pass the filters etc.