Has some group of anonymous mathematicans in Maryland improved their algorithsm to the point where attacks on encrypted communications might become feasible? Again, probably not. But it is impossible to know—unless, of course, another set of more candid researchers come up with something similar in short order.
Charles Babbage of the Economist speculates on what the “NSA’s crypto ‘breakthrough’” might be.
Personally, I doubt they’ve actually found an algorithmic way to break crypto. If they have, it would be incredibly INCREDIBLY irresponsible to use that breakthrough to spy on us, rather than trying to protect us from the impending catastrophe that would happen if cryptography could be mathematically broken.
I think their “breakthrough” is far more likely to be a backdoor embedded in Intel’s random number generator.