I'm not sure how to line up what you are saying above, and my observation that bioinformatics algorithms work better the further they are from the RM+NS assumpt

I'm not sure how to line up what you are saying above, and my observation that bioinformatics algorithms work better the further they are from the RM+NS assumption, plus all the other discrepancies I see between theory and practice in bioinformatics. Sure, Koonin and others can say RM+NS is primary, but I don't see this in the actual work people do that is quantitative and rigorous vs speculative stories. Also, I'm not talking about genetic engineering, although the fact we can do genetic engineering vs just randomly jamming nucleotides together in the hopes something happens is also fairly surprising if it was all just RM+NS.