I would assume editing In Situ would require a retrovirus or something similar. Seems incredibly risk, but then again, so was open-heart surgery until we figured out all the things that will go wrong.
There are more cells in a body than stars in the universe; I'm under the impression that the first CRISPR therapies will be on early-stage embryos where the number of cells is still manageable and we can feasibly edit all of them. Would be happy to hear more from someone who knows more about the latest in gene therapy though, my training here is getting a little dated.
1 x 10^24 stars vs 3.2 x 10^13 cells.
Estimates from a quick google.
Darn, internet info foiled by more internet info...