The news article reporting on this is too general to really figure out exactly what the hell is going on, the paper it references to is far too technical and mathematic for me to figure out what the hell is going on. The only thing I gather from this is that user generated non-random passwords are probably indeed easier to guess than we previously thought, I am not convinced this has any implications for actual encryption but I cannot figure it out either way. Yay looks like Bruce Schneier has something to say on it: