The argument that "it would still take them billions of years to try all the keys ..." is specious. The Naval Enigma machine that was so thoroughly cracked during WWII had 81,836,646,696,895,916,507,136 possible keys. [citation:
http://www.cromwell-intl.com/security/enigma.html] But it failed utterly. Encryption has often been compared to a strong steel door mounted in a wooden wall, or the strongest link in a chain. It's not what matters. You can bet that any encryption scheme will
not be attacked by brute-forcing. Banks recently were forced to admit that they routinely shipped
unencrypted backup tapes via FedEx. No matter how "secure" their on-line systems may have appeared to be, those tapes (some of which were inevitably lost) were the weak-link. The data was actually
unencrypted, not secure at all.
If I were able to get in front of your computer, then the first and most obvious thing that I could do would be to
steal it... thus, you wouldn't have access to your own data either!
Otherwise, I would know that the necessay keys will either be something that you
possess, like a USB key-token, or a password that you are able to type in and to remember easily. I would also possess several gigabytes of enciphered text, much of it with very predictable characteristics. (More valuable than it may seem: see
Probable Plaintext Cryptanalysis of the IP Security Protocols, http://www.securiteinfo.com/ebooks/palm/probtxt.pdf;
Practical Approaches to the Recovery of Encrypted Evidence (and citations),
http://www.ijde.org/archives/02_fall_art4.html;
et al.) I would know that the password is in there somewhere ... like a swapped-out page on the swap device, for instance. Or else in
Webster's Unabridged.