Let's say you have a terabyte hard drive. That means any digital content that exists, will exist and might exist up to a terabyte exists on your hard drive. You just need to organize the bits.
I am creating something that you feed a hex number, it is save as a .zip, and then an attempt is made to unzip it.
If it fails to unzip go to the next one and so on. If it opens, write that to a log and then take a look. Yes, this will produce a lot of garbage but the hits will be interesting.
Now imagine google dedicating a cluster to something like that, it would be awesome.
I wish I acted on this when I first thought of it; a 10Mb hard drive is easier to get ones mind around.