Oops. I was thinking of using Grover's which works for both. Point stands, the security of addresses isn't reduced more by quantum computers through the possibility of collisions as opposed to cracking the ECDSA directly. Finding collisions of RIPEMD-160 is still difficult enough, combined with the fact that there is SHA256 hashing prior to it and that you're finding a very small set of keys.

You might be able to find a collision of two addresses far in the future but it still probably wouldn't be worth the time