While I will not argue about brute force any increased computer power would have to help some.
Super computers there are less than a dozen in the world, according to the web and their computing time is very expensive and most are allready booked into the next century. So has anybody even tried or just relying on the mathmatic assumptions.
I did say it was a crazy idea.
If we use the new Magic ransomware as an example:
It's using AES-256-CBC, which has a 256-bit key. This means that there are 2^256 possible keys. Expanded, there are 115792089237316195423570985008687907853269984665640564039457584007913129639936 possible keys. If you calculate .1% of these keys, which is probably the minimum probability that could help anyone, you would have 1157920000000000000000000000000000000000000000000000000000000000000000000000 keys. Even if you generate all of these keys, which you can't in any reasonable timeframe, you would need 144700000000000000000000000000000000000000000000000000000000 _petabytes_ to store them all. Not to mention the fact you would need a way to query and test against this massive amount of keys to see if yours is there.
Note that my numbers might be 100% accurate, but you cannot crack AES today, regardless of the accuracy of these numbers.
The math isn't theoretical (you could try it yourself, but you'd just be burning energy you could save). AES is used to protect network traffic, classified documents, etc because of how unbreakable it is.
When you consider that the Magic ransomware probably has only hit ~100 people, there is a 0.00000000000000000000000000000000000000000000000000000000000000000000000000086361686% chance that a key you generate is the key needed to decrypt someone's files.
(I get it's a crazy idea, but I'm here to tell you it can't work. Nobody would use AES if it did. )
Edited by iangcarroll, 09 February 2016 - 09:00 PM.