The mathematical foundation of encryption has been shaken. The assumption of the security of encryption relies on the source file being absolutely uniform. However, this is often not the case - for example, in in the English language, the incidence of the letter 'e' is a lot higher than the incidence of the letter 'z'. Hence, decryption schemes which first use pattern recognition to guess at the encryption key based on probabilistic concepts have been used for years.
http://web.mit.edu/newsoffice/2013/encryption-is-less-secure-than-we-thought-0814.html
It is still exponentially hard, but it is exponentially easier than we thought.
http://web.mit.edu/newsoffice/2013/encryption-is-less-secure-than-we-thought-0814.html