This seems peculiar. I can't understand why, if this relatively straightforward modification in the production of EPSP conferred such a major advantage to the organism, that it didn't develop as a result of natural selection.
There has to be a downside that this study didn't pick up. Perhaps the change makes the plant more attractive to pests?
A simple reason could be that the gene just never happened to appear in the DNA; natural selection cannot "develop" new genes, in can only "select" which of the existing (combination of) genes are better.
There are several ways how a part of genome can suddently be duplicated with all its genes. That's how every species have different number of chromosomes.
I might guess that competitive advantage only apply where weeds are bad: in fields. Before fields there were no advantage to this.
Right, but that's not natural selection, that's mutation. Natural selection can pick up on the changes after they've taken place.
e.g., There's presumably a gene out there that makes humans twice as intelligent at absolutely zero expense, and although if it were to appear it would be selected for, it it clearly hasn't happened yet. Natural selection can't drive that to occur.
There has to be a downside that this study didn't pick up. Perhaps the change makes the plant more attractive to pests?