Some of what people are writing seems to imply more than attacks against specific parameters but rather some theoretical breakthrough that would render the scheme completely unusable. (e.g. factoring in polynomial time to key length)
People have been moving to longer keys over time and thus simply being able to attack a given key length isn't the same as saying the entire scheme is broken. Today a 768 bit key is very difficult to attack. Is RSA with 64k-bit keys likely to be broken within 5 years?
I guess in one sense our ability to work with much larger RSA keys has progressed faster than the theoretical advances in factoring. Until the time someone has some practical and provably strong cryptosystem we're kind of always at risk. ECC could also be broken so why should I feel more comfortable using ECC over RSA with 64k bit keys?