Actually, if I go to the trouble of encrypting my data, I have nothing against someone storing it indefinitely. That's the whole point of encryption — it no longer matters who has the data, it only matters who has the key.
But, unfortunately, looking at the absolute pushback from congress and the white house over this whole thing, it's pretty clear the citizens have absolutely 0 say in the matter. After all, we're just ignorant sheep. They're the ones will all the knowledge and we should just trust them. We're not going to be able to change the NSA's policies or magically make everyone follow the constitution and the law to the T overnight. This is a compromise on our end.
The NSA may have a large amount of resources for storing and running attacks on my emails, but they don't have limitless resources. For every hour they spend bruteforcing my keys to recover a sappy "Just thinking of you <3" to my wife, the less time they're spending bruteforcing someone else's emails.
AES-256 takes 2^256 guesses to break by brute force and by using Grover's quantum algorithm that's only improved to 2^128. It is something, but I can bet that AES-256 will at some point become insufficient due to Moore's law, not due to quantum computing - a theory that has originated in the 1980-1981 and that has yet to yield actual machines that we can use. And of course, Moore's law also works in our favour, since with more power and specialized chips, we'll be able to use bigger keys that aren't likely to be breakable by Moore's law or Quantum computing in the next century or even millennium.
As long as we cannot reduce NP problems to P, then virtually unbreakable cryptography is possible. Also, cryptography as a (mathematical) field is pretty new.
Uhm – classical computer science originated sometime in the early 19th century (possibly earlier) in the form of analytical engines and the first actually usable machines[0] were created just a bit before World War II. General relativity originated before WWI and was first used well after WW II. Quantum mechanics was invented at the beginning of the last century and the first usable LASERs were built in the 1960s.
Really, 30 years is not that long in the context of ‘hard’ science. Even in the ‘fast-moving’[1] field of software engineering, people still use SMTP (1982), TCP/IP (1975) and FTP (1971).
[0] ‘Usable’ in the same sense that QC are not usable yet because people can perfectly factor 15 without Shor’s algorithm.
[1] Because it’s young and hence easy to discover new stuff.
I'm not saying that it won't happen eventually. I'm only saying that Moore's law is a more imminent threat to current encryption methods and that quantum computing still doesn't solve NP and as long as NP is not P, unbreakable encryption is a reality that quantum computing won't change.
edit: That was odd downvotes. Anyone care to explain?