I am a scientist and I am impartial if that is good enough :)

The major problem is the same as it has been for the last 30 years - delivery. You can make the most amazing molecular machinery for modifying genes, but if you can't get it into a very high percentage of adult cells then it is useless for most treatment purposes. We still don't have any good delivery systems that work on all cell types and doesn't come with the very real risk of death via immune hyper reaction.

This. Completely. Also, I still worry about off-target edits / changes, though that is less of a concern with the technologies available today compared to even 5 years back.

But delivery remains the biggest challenge.

IMHO, in some cases, cells can be converted into stem cells, repaired in lab conditions, and then delivered back into damaged organ or area.

Yes some haematological diseases can be treated this way, but not much else. We really need much better delivery systems that work in all tissue types.

From what I read in previous posts here in HN, blindly running find & replace on your code might cause some side effects.

This is the most hyped up biomed topic I have seen in the last 10 years.