"Way more" is an exaggeration. I'd estimate it's a 2:1 ratio (of profs and postdocs, not techs). Second, I am not disputing that most theorists use code from time to time. Lord knows I do. The claim is that this is nothing like 80% of their job.