CIDR vs. VLSM

Sort of. VLSM gives you exactly what it says. You can have subnets of varying sizes (/25, 26, bunch of /30's etc) Once routing protocols started supporting VLSM, there was no need to keep everything at one mask. It was incredibly wasteful.

CIDR just did away with the concept of CLASS A, B, C networks. Think of it this way, before CIDR, what would have happened if you needed 500 addresses? You would have been assigned two Class C addresses. That's two entries in the routing table. But with CIDR, you could just represent it with one routing entry (/23 for example)

As far as I understand it they are both essentially the same thing, just semantics.