
The human nervous system consists of about one hundred billion nerve cells. These one hundred billions have a great responsibility towards the body. We can say we still live thanks to the remarkable ability of our neurons to predict the future.
One of the best human ways to predict the state of the surrounding space is to model the world using a variety of simple rules that the world seems to obey. For example, such: the released stone flies down with acceleration. Or this: creatures with a number of properties - four legs, a long tail, a mustache, ears of a particular shape, and so on, are united by a common genus of felids and the signs accompanying this genus. Love for fresh meat for lunch, for example. Agree, it is very important and interesting if you want to survive a meeting with especially large representatives of this kind of creatures.
A long time ago, even in pre-Platonian times, people faced the question: what is the structure of this Mechanism of Modeling the World? How to describe any of the one hundred billion nerve cells (then they, however, have not yet been counted), and their interaction with each other? What is the nature of the mind? Since there have not yet been any sensible nonhumans, it was quite logical for the world’s modeling mechanism to assume that intelligence is a special, special human quality. There is a soul, inspired people, thanks to her, reasonable. Where there is no soul, no mind. Animals probably also have the rudiments of the soul - difficult to say, but the mechanism certainly cannot have a soul or a mind. Consequently, it is impossible to model the mind.
')
It must be said that this theory still works as a whole.
Unfortunately, the theory of the universe, in which there is an unmodeled soul, is no good. It’s rather stupid when studying the world to introduce various functions of God with implicit behavior into the models being discovered. You can save the situation by assuming that one or another mystery of nature is inaccessible to us because of our lack of understanding of some deep basic building blocks. Who knows how a human neuron works? Any of a hundred billion? Perhaps, inside it, in cellular organelles, a whole quantum computer is hidden, calculating how to react to stimulation. (A quantum computer is a very good thing. No worse than a soul. We also know nothing about it.)
It is quite another thing if we have a neuron model. Very simple. Here is the input, here is the output, here is the conversion function, here is the self-learning function. Give me the resources of one hundred billion of these, I will follow the architecture and make you a little man. One hundred billion, of course, quite a lot. You have a smaller transistor in your computer. But Moore's law is still working, and it gives hope that when you re-read this article after a few years, the power of your system will be enough. By the way, I do not rule out that the article will not be re-read by you, but by your computer.
But still, can the Sacrament occur deeply in the body of a neuron? May be. Maybe the Modeling Mechanism is an incomprehensible thing. Maybe it is too complicated, maybe the exponent of Moore's law will stop growing before we can get closer to his understanding close enough.
Or maybe a simple model of the nerve cell will be enough? Maybe we
already know how a neuron works, and the matter is only about the network architecture.
Fortunately, if this is true, then we are all very likely to understand this soon.
Because to recognize in creatures with four legs and a tail of cats a
sufficiently large neural network can already now.