You may already know imgur, an image host you can use freely for whatever you please.
An interesting fact about them, however, is that each picture given a name consisting of only 5 alphanumerical characters. This leaves for roughly 916 million different names. As it turns out, there appears to be around 90 million uploaded images, which means that if we randomly guess a name for a file, we have an about 1 in 10 chance of actually finding an image.
Those odds are pretty good, so I wrote a small application which does just that.
-  Rough estimate based on empirical observations. ↩
Quines are a wonderful thing. A quine is a program, that when run produces its own code as output. Now, in most interpreted languages you can read your own code through means of I/O — I’d consider that cheating: We can do much better than that. We’ll do it without anything but a clever theoretical result. (And maybe a little bit of code to make it work in practice.)