It’s not all that hard, actually – just picture two dots on a rubber sheet, dot A and B, then expand that sheet uniformly (i.e. pull on all sides): they’ll ‘move’ apart. Now, picture another point, C, twice as far away from A on the same line that connects A and B, and continue to expand the sheet – C will move away from a twice as fast as B does, because in the time it takes to increase the distance between A and B, the distance between B and C has increased by the same factor, the expansion being uniform and all. Then, picture a third point, a fourth, fifth, etc., each receding from A twice as fast as the one before, and at some point, any given velocity will be exceeded.
So the whole deal looks kinda like this, if you take snapshots of the expansion at fixed intervals:
[-A–B–C-]
[-A----B----C-]
[-A--------B--------C-]
[-A----------------B----------------C-]
And so on.
On preview I see I’ve been beaten to the explanation, however, this:
isn’t quite correct: We do observe, right now, galaxies that are receding faster than the speed of light from us. This seems somewhat paradoxical at first, however, the answer lies directly with the finiteness of the speed of light: if a photon is emitted in a region of space that is receding from us at a velocity greater than c, its total velocity will be directed away from us. Now, let’s call the boundary beyond which things recede from us faster than c the Hubble sphere, because that’s what it’s called. That boundary increases, because it is inversely proportional to the Hubble constant, which is itself inversely proportional to time, if we forget the question of decelrating/accelerating expansion for a moment. Thus, photons originally outside of the Hubble sphere, i.e. photons emitted by objects receding faster than light, can actually enter the Hubble sphere, which means they then are in a region receding with less than c from us, and, themselves still having velocity c in directed towards us, thus are approaching us.
Also, of course, we can see things receding faster than c from us now, because they weren’t receding at superluminal speed at the time they emitted the light we’re seeing now.