Coming on the heels of the various revelations about the NSA's activities, this article in MIT News is fascinating reading:
When Médard, Duffy and their students used these alternate measures of entropy, they found that slight deviations from perfect uniformity in source files, which seemed trivial in the light of Shannon entropy, suddenly loomed much larger. The upshot is that a computer turned loose to simply guess correlations between the encrypted and unencrypted versions of a file would make headway much faster than previously expected.
The point here is that encryption is still secure. It's just not *as* secure as we might have assumed.