One of the points I believe is sorely missing in these discussions is how important "improvements in algorithms" can be. In the narrowest sense, I agree with your statements - but I have also seen what elegant alternative approaches can do to systems that were presumed to be vulnerable only to brute force, and I've also seen how nicely they may be placed into custom hardware.
When you are talking "heat death of the universe" time lengths, improvement is algorithms don't really add up to all that much time. In the real world (outside of Academentia) we have different threat models that we need Crypto for. To keep a credit card safe, we need only to make sure that a given undesired decrypt be more expensive than it's worth--and the encrypted credit card string has to last what? Three years? before it's worthless anyway. I'll take the risk that someone will improve factoring by what? 6 or 7 orders of magnitude? (that makes 1,000,000,000 years into 1000 years. I think my card will be expired by then). Other sorts of banking operations have an even short life--from minutes to months. They could take almost 9 orders of magnitude(unless I don't understand this order of magnitude thing)--does it really matter if a banking transaction falls to a break in 10 years? One would think that a bank would be wise enough to expire it's keys more regularly than that. Or military secrets--because of the nature of the military, keys can be expired even more rapidly 3 to 5 years ought to be plenty. And hey, if we do get a break through in factoring speed, it seems cheap enough to double our key size. Quantum computers are a different story--and may (may) make a shambles of our current crypto schemes--but as near as we can tell no one is close to a working system.
Call it threat analysis - I think it is reasonable to assume they know a few tricks that aren't public yet. And any trick related to factoring or Feistel networks is sufficient to obsolete those "age of universe" extrapolations.
There is a wide difference between "age of universe" and "age of man". The point of the whole "heat death of the universe" thing is that even if a given brute force decrypt can be made 1000 times faster, it's still going to take a *LONG* time. -- A quote from Petro's Archives: ********************************************** Sometimes it is said that man can not be trusted with the government of himself. Can he, then, be trusted with the government of others? Or have we found angels in the forms of kings to govern him? Let history answer this question. -- Thomas Jefferson, 1st Inaugural