I'm not overlooking it. It sounds very reasonable but every example I've seen described shows the problems associated with tampering with the designer's codes.
Well,
the single example you cite does involve a duplication that results in a disease state - this appears to be a tactic of yours: to bring up one or two detrimental mutations as "proof" that some beneficial genetic mutation cannot occur.
For every "type" of mutation, whether it be base change, insertion, deletion, inversion, frame-shift, duplication; one can find a pile of detrimental mutations of that type.
That in
no way means that beneficial forms of these mutations have not occurred.
™ Version 4.2