The approach they used which is base editing doesn’t actually insert or remove DNA, it actually uses an enzyme to convert one base to another, which is much safer as this doesn’t require a double strand break in DNA: https://blog.addgene.org/single-base-editing-with-crispr
That is interesting, I didn't catch the difference my first time through the article.
I do still question their claim of 100% precise results though. At least based on that high level description I can definitely see it being safer, but I question any scientific claim that is an absolute.
Specific to the editing vs insertion mechanism, I question how it doesn't run into similar constraints where the mechanics of targeting exact portions of the DNA can occasionally miss or impact the wrong segment of DNA entirely.
I haven't dug as deeply down the base pair conversion though, so I could absolutely be wrong!