For all the irreducible complexity fans.
Anyone can sell their soul to get results out of the ordinary; this is nonsense.
Next this person is going to start using statistics, a "science" long since disproven. Just ignore them.
Perhaps someone will someday explain how to use ping lists.
Now that I've read the paper I can comment. It's rather nice. They emulate DNA functionality better than my simulations. I liked the was that sub-optimal forms are generated naturally and that one may have to
retirer pour mieux sauter (I don't remember where the diacriticals go). I suggested some time ago (in analogy with the Metropolis algorithm) that one might acquire several neutral or even slightly disadvantageous mutations before the combination gave a great leap forward. The paper suggest that this happens.
I did not check the dynamics of their simulation. What I found in mine was that randomly driven (Brownian motion) mutations causes the genotypes to vary at a Sqrt(N) rate for N time steps but the selection worked proportional to N. (I should develop a real study of this after I retire and get some time.)
Thanks for the article. It was fun to read.