The real star is their ability to model molecular physics like this. I recall doing some protein folding work back in the early 2000s and wonder if this would make it substantially easier.