Or it could happen by some freak accident 5 minutes from now.
Your response was aimed at strengthening the computationally expensive argument.
Why are we using a 160 bit hash instead of something that is more resistant to collisions? Will we be able to move to a 224/256 bit hash if and when the need arises, even if it's 100 years from now?
I'm not a programmer, I don't work on bitcoin code, but I do understand how both of those things work. I'm going to say "yes". It might take everyone moving to Bitcoind v2.1.744 in the year 2121; but it really doesn't sound THAT difficult.