[sokoloff]

I’m no expert in the field and the only remotely relevant experience I have is a tiny amount of genetic algorithm programming at a rank amateur level.

I assume that making perfect replicas is the default way for a given type of life to dominate. Making “random copies” (an oxymoron) is not likely to result in domination of any individual type (of building block, single cell, or multi-cellular life). Making perfect copies allows a build-up of “winners”.

If that implies a very strong selection process for perfect copies, then after that a selection process for increased (but still small in aggregate) mutation rate would seem globally helpful.