> The rules it follows can't possibly be that complex

Sure, but could it be that Feynman Integrals are one of the least complex measurable predictions/calculation of those simpler, underlying rules, potentially acting in dimensions we can't measure?