Brain computer confusion

There is a never ending discussion, which very concisely can be summarized in this tweet below:

And frankly any time I see similar exchanges (and I see a lot of them) I get mildly irritated. Let me get to the essence. 

Computer analogy

Computers have been undoubtably the shaping invention of the recent century and hence they have became a strong theme in our culture. Since the theory on which computers have been built is a branch of mathematics, by definition an abstract discipline, computers have also had a major impact on philosophy. We learned for example that everything we can write an equation for can be in principle calculated on a computer. This leads to somewhat profound philosophical consequences summarized as follows:

  1. Stuff we can write equations for is in principle computable
  2. We can write equations for physical interactions of molecules
  3. Everything is made of molecules
  4. Hence everything is computable
  5. Hence in principle we could simulate an entire brain in a computer
  6. And since we can in principle simulate a Turing machine in a brain, hence brains and computers have to be equivalent
  7. Furthermore, in principle we could simulate entire Universe
  8. Hence universe must be a computer too 

When formulated in this form the argument seems to be very hard to argue. And yet there is something disconcerting here. Let's start by doing a bit of a language mumbo jumbo to see where this argumentation has holes. Let's start with replacing word "computer" with an "abacus". After all, an abacus is a very simple computer. So anything computable is in principe computable on a sufficiently large abacus. And so jumping a few steps we conclude that brain is just an abacus. Which obviously doesn't make much sense. But what exactly is wrong?

Computable universe

Under more scrutiny there is a potential problem a lot earlier, with step 2 of this reasoning. We take for granted that we can write equations for molecules, yet this isn't really the case. We can write equations for "approximations" of molecules, ignoring some of the details. OK well, we can literally write the quantum field equations down to the most nitty gritty details of standard model of physics, down in indivisible primary particles and their interactions. After all, all this is governed by quantum mechanics i.e. Schrodinger equation and so it's all computable and we are good. 

Only we are not good. Since wave functions in principle span entire universe and require infinite representation of state (I'm not even gonna go into minor details such as why wave functions collapse into their eigenstates or whether such "representation" should require all the possible quantum outcomes etc.). We're entering here a very philosophically dangerous grounds: any "finite" approximation of universe we can think of seems in principle computable. But that does not mean that the universe itself is computable! Even with our own theory of computation, we have no idea what to think about a Turing machine with uncountably many states! And furthermore, there is no guarantee whatsoever, that even if the Universe is fully described by a set of equations, that this set of equations is finite!!! When digging deeper and deeper into the structure we may find that the total number of equations and constants necessary to describe reality in full detail is actually increasing without a limit!!! This entire mental exercise that tries to fit the entire Universe into a giant Turing machine is fundamentally flawed!!!

Emergence of analogy

People love analogies because it helps us understand things more easily. Whenever we have two things in some ways similar we tend to build comparisons and extrapolations. And human intelligence/brain is regularly compared to the most complex device we have available at hand. History is full of analogies made in the past in which we try to explain the unknown with the things that we do know. Hence a human can be though of as a steam machine. After all, we consume energy rich molecules and oxygen, exhale carbon dioxide (and some steam!) and produce heat and energy. Seems like a great analogy, and much like many analogies it is useful to a degree. But of course we know today we are much more sophisticated "mechanisms" than a steam engine. This analogy can only get us so far, and most people today understand it's limited applicability. 

But the computational analogy is more extreme. It's shoved down our throats by the crowd of people I'd like to call "computationalists" (computational fundamentalists) as some sort of fundamental truth about the world of some semi religious profoundness. And the straightforward consequence of this flawed philosophy is that brain is just yet another computer, and since we build faster computers every day, it's just a matter of time when we build one as sophisticated as the brain. That symbolic day is marked as some sort or religiously significant "singularity" after which nothing will be like it is. I'd argue it's equally pointless to wait for such "singularity" as it is to await a giant man made bird to fly out of a Boeing factory. 

Computational brain

So let's get things straight here. Brains and computers are in some ways similar:

  • Both apparently rely on electric signaling (although in principle computers could work on light, or game of life, or mechanical gears, or wooden blocks ...)
  • Both seem to exhibit what could be called a memory
  • Both have what could be considered input and output
  • Both need energy to work
  • Both are capable of what could be called computation, i.e. manipulation of abstract symbols, though computers seem to be vastly better at that
  • Both are capable of what could be called perception, i.e. breaking down representation of physical reality into a set of useful categories and symbols, though brains seems to be vastly better at that

but these two are also separated by a vast number of differences:

  • Computers are built by humans while brains emerge from biological goo that somehow self-organizes
  • Computers are hopeless outside of their intended role in reality which is currently entirely determined by humans who build and deploy them
  • Brains on the other hand spontaneously seek new roles and opportunities in the complex world to facilitate energy extraction and reproduction of their hosts
  • Computers very much separate their abstract calculation from their underlying physics. In result they tend to be very strict and formal, but also incredibly wasteful of energy, since all this error correction costs a lot of energy
  • Brains seem to be incredibly energy efficient and leverage their underlying physics to advance their operation
  • Computers are extremely fragile, little spark, drop of water, small short circuit and they are dead
  • Brains are robust, even soaking them with alcohol which changes the dynamics of pretty much every neurotransmitter in every synapse results with impaired, but still functioning brain
  • Computers have to be given a program, brains (some of them at least) write programs

And so on and so on. So the brain-computer analogy isn't really more useful than bird-plane analogy. if you reject the fundamental philosophical point that the entire Universe has to be a computer (which I dismissed in the paragraph above), there remains nothing sacred in the brain-computer analogy. Like any analogy it has its limits and it even becomes a hindrance beyond certain point, just like it's a hindrance to put feathers on a wing of a plane. There is just a point at which brains and computers are different and dragging this analogy to the limit is silly, counter productive and hard to defend outside of of the semi-religious "pan-computationalism". 

Conclusion

My personal bet is that if we ever figure out how to build artificial brains (or have them emerge in some process of self organization), they will look so unlike anything similar to a computer, that entire analogy will be perceived as preposterous. Just like today we may laugh at people who though they could fly by slapping some feathers on their arms, people in the future will be laughing at "Kurzweilianizm", "pan-computationalism", digital universe and other such nonsense. 

If you found an error, highlight it and press Shift + Enter or click here to inform us.

Comments

comments