It's certainly not a definitive proof on the matter, but definitely a strong point in favor of introspection on why various methods might or might not be appropriate to understanding the brain. The processor example is a funny one, because you can just say: hey we have a theoretical model about how this processor works, oh, here are some quick ways to prove the registers and ALUs and etc are here here and here. But when you don't even know that there is any sort of structured model and it's not all just one giant interconnected mess, you don't even know where to apply the exacto-knife to narrow down the number of tests and points you need to probe. The TA from the class I took this semester was telling me his prediction that a lot of the advances in neuroscience won't come from the people doing the hardcore-est electrophysiology or biochemistry, but the ones who look between animals that either do or do not have various cognitive features. It's a fun thought.