I've already paid most of the cost of running our own: there isn't much left to pay (unless something goes horribly wrong, I suppose). It's definitely a lot of wasted effort on everybody's part if we switch to bzr or git, and that doesn't thrill me so much.

It's more about the positives I think we can get by using a collaborative service like this. I agree about post-launch, that's a good compromise between "get things done" and "get where we want to go."