Please provide a citation for this "fact". There is an attempt underway to calculate discrete logs on a 130-bit elliptic curve over a prime order field. Without some massive algorithmic improvements we're not going to have any chance of attacking 256-bit curves in eight years. I seem to recall that there is some speculation that humankind will never be able to count up to 2^128 let alone perform an attack with such a work factor.
ByteCoin
I read a paper on it, though I checked my saved documents and it doesn't appear that I had saved it and I don't remember the author. It was also written more in the sense about symmetric cryptography and how long do you need your data to be secure and such, and backing it up with information regarding moore's law and other factors. 128bit might have been 2030 too, I'm just going from memory which is why I said "or so." 256-bit security is the magical number that would be impossible to count to, though 128 is still pretty significant. But DSAs and SHAs are more prone to vulnerabilities than symmetric cryptography, so how long 128 bits will be secure remains to be seen.