If we're talking about the original article, side projects were mentioned in the context of "The instinct to experiment first". There might be some correlation there, but if you limit the focus to programming side projects you've just chosen a slightly bigger box to experiment in.
Many advances in science come by drawing knowledge from a completely different field and applying it to a familiar one. Deep knowledge of programming is necessary but not sufficient for creative programming solutions. Design patterns in software, to take one familiar example, are based on Christopher Alexander's theories about physical architecture. The article itself recommends amusement park rides to foster greater risk-taking. Non-programming hobbies can have great benefits for programmers.