When it comes to complexity, I think it's a miracle that spinning hard disk drives actually work.
Though I guess I can sort of understand how that works. What I can't understand is that I can understand anything... I simply can't fathom how the human brain works...
Look for some popularized science articles (e.g. from arstechnica.com/science) and books (e.g. from Ray Kurzweil) that mention neural networks, fMRI, connectome, cellular automata, etc. There are competing theories, but there's enough accessible information out there to dispel some of the mystery.
Until recently, there were two basic camps on "how we know stuff." The Rationalists and the Empiricists.
The Rationalists thought that knowledge was came from mental abstractions, not from reality. For example, Plato believed that we could all access the "ideal" world of "forms" by just introspecting. The medieval scholastic philosophers tended to be very rationalistic - they made arguments about abstractions not connected to real experience, like how many angels could fit on the head of a pin. Decartes was a rationalist - "I think, therefore I am" is rooted in the idea that we look inward to perceive reality.
The Empiricists, which are somewhat more modern, were a reaction to Rationalism, and rejected all of that. They said that all we can really "know" or "trust" is sensory experience. We can't build up complex mental models, and we can't have principles - we just have to be pragmatic all the time and do what seems best, because we can't really "know."
Kant tried to improve on both of these approaches by saying that we can't really know anything about "true" reality, because reality has to be filtered through the senses. There is only subjective "reality."
Ayn Rand, of whom I am a big fan, went in the other direction from Kant, but also rejected Rationalism and Empiricism. She said that you can build complex models of reality, but they have to be based on actual observations of reality (i.e., sense data). To do this, she proposed a new theory of concept formation (i.e., how to form abstractions based on observation of reality). Her approach is very Aristotelian, which is interesting because Aristotle was the main guy who disagreed with Plato at the very beginning. I'd highly recommend "Introduction to Objectivist Epistemology" if you're interested in this approach (although you may be better served by starting at a less advanced level... still, the book is short and very accessible, albeit kind of a mind trip).
Now if we dig down into the details both look seemingly impossible. For hard drives we have read heads that must sweep so close to the spinning platter that they use the wing ground effect to float just over the surface. We no longer can read data with the write coil (not enough gain) so instead we use spin polarization sensors that pass currents of electrons with only one intrinsic spin and then measure the effect the magnetic field of the platter had on the spin of the electrons in the current.
In the case of solid state memory, we are able to pattern silicon wafers with a minimum feature size of 22nm and falling. A single chip is a seemingly miraculous network of chemically deposited thin films and optically patterned cutouts all at a size that beats the detraction limit of light. The mask (imagine an overhead projector's transparency) can not be shaped the same as the intended end-shape pattern of silicon due to very small scale refraction but never fear! We have figured out how to calculate what our mask must look like to beat the diffraction limit. Its all amazing stuff.
My point really was that one is to this day made of a disk coated in magnetic material (with the read/write head and electronics getting more sophisticated over the years) while the other uses billions upon billions of transistors that are "etched" out of silicon wafers using an ever more complicated process.