Living in a digital world: how is computer technology being embedded in the brain?

2024 Author: Seth Attwood | [email protected]. Last modified: 2023-12-16 15:55

Our brain is adapted for life in a cave, and not for processing non-stop streams of information - studies show that it stopped in its evolutionary development 40-50 thousand years ago. Psychophysiologist Alexander Kaplan in his lecture "Contact with the brain: realities and fantasies" told how long a person will be able to cope with life in the conditions of huge highways, movements around the planet and endless incoming, and also how we ourselves can fix or spoil everything with the help of artificial intelligence …

Imagine a situation: a person comes to a store, chooses a croissant, gives it to the cashier. He shows it to another cashier and asks: "What is this?" He answers: "40265". Cashiers no longer care what the croissant is called, it is important that it is "40265", because the computer in the cash register perceives the numbers, not the names of the buns. Gradually, everything plunges into the digital world: we live next to computing technology, which understands physical objects as digital, and we are forced to adapt. The era of the Internet of Things is approaching, when all physical objects will be presented in digital form and the Internet will become the owner of our refrigerator. Everything will revolve through numbers. But the problem is that the intensity of information flows is already too great for our ears and eyes.

Recently, a method has been developed to accurately determine the number of nerve cells in the brain. Previously, it was believed that there are 100 billion of them, but this is a very approximate figure, because the measurements were carried out by a not entirely correct method: they took a tiny piece of the brain, under the microscope they counted the number of nerve cells in it, which was then multiplied by the total volume. In a new experiment, a homogeneous mass of the brain was stirred in a mixer and the nuclei of nerve cells were counted, and since this mass is homogeneous, the resulting amount can be multiplied by the total volume. It turned out 86 billion. According to these calculations, a mouse, for example, has 71 million nerve cells, and a rat has 200. Monkeys have about 8 billion nerve cells, that is, the difference with a man is 80 billion. Why was the movement in animals progressive, and why the break with man was so sharp? What can we do that monkeys cannot?

The most modern processor has two to three billion operating units. A person has 86 billion only nerve cells, which are not identical to an operational unit: each of them has 10-15 thousand contacts with other cells, and it is in these contacts that the issue of signal transmission is resolved, as in the operational units of transistors. If you multiply these 10-15 thousand by 86 billion, you get a million billion contacts - there are so many operational units in the human brain.

An elephant's brain weighs four kilograms (a human's one and a half at best) and contains 260 billion nerve cells. We are 80 billion apart from the monkey, and the elephant is twice as far away from us. It turns out that the number of cells does not correlate with intellectual development? Or have the elephants gone the other way, and we just don't understand them?

The fact is that the elephant is big, it has a lot of muscles. Muscles are made of fibers the size of a human or a mouse, and since an elephant is much larger than a human, it has more muscle fibers. Muscles are controlled by nerve cells: their processes fit to each muscle fiber. Accordingly, the elephant needs more nerve cells, because it has more muscle mass: out of 260 billion elephant nerve cells, 255 or 258 billion are responsible for muscle control. Almost all of its nerve cells are located in the cerebellum, which takes up almost half of the brain, because it is there that all these movements are calculated. In truth, 86 billion human nerve cells are also located in the cerebellum, but there are still significantly more of them on the cortex: not two or three billion, like an elephant, but 15, so our brains have immeasurably more contacts than elephants. In terms of the complexity of the neural network, humans have significantly overtaken animals. Man wins by combinatorial skills, this is the wealth of the brain matter.

The brain is very complex. For comparison: the human genome consists of three billion paired elements responsible for the encoding. But the codes in it are completely different, so the brain cannot be compared with the genome. Let's take the simplest creature - the amoeba. She needs 689 billion pairs of coding elements - nucleotides. There are 33 coding elements in Russian, but 16 thousand words of the Pushkin dictionary or several hundred thousand words of the language as a whole can be made of them. It all depends on how the information itself is put together, what the code is, how compact it is. Obviously, the amoeba did this extremely uneconomically, because it appeared at the dawn of evolution.

The problem with the brain is that it is a normal biological organ. It is evolutionarily created in order to adapt a living creature to its environment. In fact, the brain stopped in its evolutionary development 40-50 thousand years ago. Research shows that Cro-Magnon man already possessed the qualities that modern man has. All types of work were available to him: collecting materials, hunting, teaching youth, cutting and sewing. Consequently, he had all the basic functions - memory, attention, thinking. The brain had nowhere to evolve for a simple reason: man became so intelligent that he was able to adjust the environmental conditions to fit his body. The rest of the animals had to change their body for the environmental conditions, which takes hundreds of thousands and millions of years, but we completely changed the environment for ourselves in just 50 thousand.

The brain was imprisoned for life in a cave. Is he prepared for modern palaces and information flows? Unlikely. Nevertheless, nature is economical, it sharpens the animal for the habitat in which it exists. A person's environment, of course, changed, but its essence varied little. Despite the dramatic changes that have occurred since antiquity, the mechanics of the environment in the routine sense has remained the same. How has the activity of designers making a rocket instead of a Zhiguli changed? Of course, there is a difference, but the meaning of the work is the same. Now the environment has changed fundamentally: huge highways, endless phone calls, and all this happened in just 15–35 years. How will a cave-polished brain cope with this environment? Multimedia, huge, inadequate speeds of information flow, a new situation with movements around the planet. Is there a danger that the brain can no longer withstand such loads?

There is a study of the incidence of people from 1989 to 2011. Over the past 20 years, mortality from cardiovascular and oncological diseases has decreased, but the number of neurological disorders (memory problems, anxiety) is sharply increasing over the same time. Neurological diseases can still be explained by behavioral problems, but the number of psychological diseases is growing just as quickly, and at the same time they become chronic. These statistics are a signal that the brain can no longer cope. Perhaps this does not apply to everyone: someone goes to lectures, reads books, someone is interested in everything. But we are born different, so someone's brain is better prepared due to genetic variation. The proportion of people with neurological diseases is becoming very significant, and this suggests that the process has gone in a bad direction. The third millennium challenges us. We entered the zone when the brain began to send signals that the environment we created was not useful for it. It has become more complex than what the brain can provide us in terms of adaptation. The stock of tools sharpened for the cave began to be depleted.

One of the man-made factors pressing on the human brain is that many decisions are now associated with the likelihood of a serious error, and this greatly complicates the calculations. Previously, everything we learned was easily automated: we learned to ride a bike once, and then the brain did not worry about it. Now there are processes that are not automated: they must be constantly monitored. That is, we need to either call an ambulance or return to the caves.

What more progressive ways of solving this problem do we have? Perhaps it is worth combining with artificial intelligence, which will refine the flow: reduce the speed where it is too high, exclude information that is unnecessary at the moment from the field of view. Automatic controllers who can prepare information for us are akin to primary cooking techniques: they chew it so that it can be consumed without wasting much energy. When the man began to cook food on the fire, there was a very big breakthrough. The jaws became smaller, and there was room for the brains in the head. Perhaps the moment has come to dissect the information around us. But who will do it? How to combine artificial intelligence and natural intelligence? And this is where such a concept as a neural interface appears. It provides direct contact of the brain with the computing system and becomes an analogue of cooking food on fire for this stage of evolution. In such a trio, we will be able to exist for another 100-200 years.

How to implement this? Artificial intelligence in its usual sense hardly exists. A highly intelligent game of chess, in which a person will never beat a computer, is akin to a weight lifting competition with an excavator, and it's not about transistors, but about the program written for this. That is, programmers simply wrote an algorithm that provides for a specific answer to a specific move: there is no artificial intelligence that knows what to do on its own. Chess is a game with a finite number of scenarios that can be enumerated. But there are ten meaningful positions on the chessboard to the 120th degree. This is more than the number of atoms in the universe (ten in the 80th). Chess programs are exhaustive. That is, they memorize all championship and grandmaster games, and these are already very small numbers for enumeration. A person makes a move, the computer selects all games with this move in seconds and monitors them. With information about the games already played, you can always play an optimal game, and this is pure scam. In no championship a chess player will not be allowed to take a laptop with him in order to see which game was played by who and how. And the machine has 517 laptops.

There are games with incomplete information. For example, poker is a bluff-based psychological game. How will a machine play against a person in a situation that cannot be fully calculated? However, recently they wrote a program that copes with this perfectly. The secret is too much. The machine plays with itself. In 70 days, she has played several billion games and accumulated experience far exceeding that of any player. With this kind of baggage, you can predict the results of your moves. Now the cars have hit 57%, which is quite enough to win in almost any case. A person is lucky about once in a thousand games.

The coolest game that cannot be taken by any brute force is go. If the number of possible positions in chess is ten to the 120th power, then there are ten of them in the 250th or 320th, depending on how you count. This is astronomical combinatorialism. That is why every new game in Go is unique: the variety is too great. It is impossible to repeat the game - even in general terms. The variability is so high that the game almost always follows a unique scenario. But in 2016, the Alpha Go program began to beat a person, having also previously played with itself. 1200 processors, 30 million memory positions, 160 thousand human batches. No living player has such experience, memory capacity and reaction speed.

Almost all experts believe that artificial intelligence is still a long way off. But they came up with such a concept as "weak artificial intelligence" - these are systems for automated intelligent decision-making. Some decisions for a person can now be made by a machine. They are similar to human ones, but they are accepted, just like in chess, not by intellectual labor. But how does our brain make intellectual decisions if the machine is much stronger in both memory and speed? The human brain is also made up of many elements that make decisions based on experience. That is, it turns out that there is no natural intelligence, that we are also walking computing systems, just that our program was written by itself?

Fermat's theorem has long been a conjecture. For 350 years, the most prominent mathematicians have tried to prove it analytically, that is, to compose a program that will ultimately prove, step by step, in a logical way, that this assumption is true. Perelman considered it his life's work to prove Poincaré's theorem. How were these theorems proven? Poincaré and Perelman had no analytical solutions in their heads, there were only assumptions. Which one is a genius? A genius can be considered the one who created the theorem: he proposed something to which he did not have any analytical approach. Where did he get this correct assumption? He didn’t get to him too much: Fermat had only a few options, as did Poincaré, while on a specific issue there was only one assumption. Physicist Richard Feynman concluded that in almost no case was a great discovery made analytically. How then? Feynman replies, "They guessed it."

What does "guess" mean? For existence, it is not enough for us to see what is and make decisions based on this information. It is necessary to put in memory something that will be useful later to refer to. But this stage is not enough to maneuver in a complex world. And if evolution selects individuals for ever more subtle adaptation to the environment, then more and more subtle mechanisms must be born in the brain in order to predict this environment, calculate the consequences. The specimen plays with the world. Gradually, a brain function arose that allows one to build dynamic models of external reality, mental models of the physical world. This function adjusted itself to evolutionary selection and began to be selected.

In the human brain, apparently, a very high-quality mental model of the environment has developed. She perfectly predicts the world even in places where we have not been. But since the world around us is integral and everything is interconnected in it, the model should pick up this interconnection and be able to predict what was not. Man acquired a completely unique opportunity that sharply distinguished him in the evolutionary series: he was able to reproduce the future in the neurons of his brain using models of the environment. You don't need to run after the mammoth, you need to figure out where it will run. To do this, in the head there is a model with the dynamic characteristics of a mammoth, landscape, animal habits. Cognitive psychology insists that we are working with models. This is where 80 billion neurons are spent: they contain them. The model of the world of mathematics, of the world of mathematical abstractions is very diverse, and it suggests how this or that gap should be filled, which has not yet been thought out. Conjecture comes from this model, as does intuition.

Why can't monkeys work on full-fledged models of the physical world? After all, they exist on Earth for hundreds of millions of years longer than humans. Monkeys are not able to collect information about the world around them. In what units will they describe it? Animals have not yet developed a method for compact and systematic modeling of external information in the brain with the ability to operate on it. A person has such a method, and taking into account the smallest details. It's a language. With the help of language, we have designated with concepts all the smallest grains of sand in this world. Thus, we transplanted the physical world into the mental one. These are names that circulate in the mental world without any mass. By writing out addresses using complex brain structures, as when programming in a computer, we gain experience of communicating with the world. Connections arise between concepts. Each concept has flags to which you can attach additional meanings. This is how a large system grows, which works associatively and cuts off unnecessary values using addresses. Such a mechanic must be supported by a very complex network structure.

Our thinking is based on guesswork. We do not need to count variations of chess pieces - we have a dynamic model of the chess game that tells where to move. This model is solid, it also has experience of championship games, but it is better because it predicts a little ahead of time. The machine remembers only what is, our model is dynamic, it can be started and played ahead of the curve.

So, is it possible to combine the brain and artificial intelligence, albeit diminished and reduced in rights, so that creative tasks remain with a person, and memory and speed - with a machine? There are nine million truckers in the United States. Right now, they can be replaced by automated decision-making systems: all tracks are very neatly marked, there are even pressure sensors on the track. But drivers are not being replaced by computers for social reasons, and this is the case in a variety of industries. There is also a danger that the system will act contrary to the interests of the person, putting economic benefits above. Such situations, of course, will be programmed, but it is impossible to foresee everything. People will sooner or later fall into the service, the machines will use them. Only a brain capable of creative solutions will remain of a person. And it doesn't have to be due to a conspiracy of machines. We ourselves can drive ourselves into a similar situation by programming the machines so that, fulfilling the tasks we have set, they will not take into account human interests.

Elon Musk came up with a move: a person will walk with a backpack with computing power, which the brain will turn to as needed. But in order to assign certain tasks to machines, direct contact with the brain is required. A cable will run from the brain to the backpack, or the car will be sewn under the skin. Then the person will be fully provided with transcendental memory and speed. This electronic device will not pretend to be a person in history, but for employers, a person will expand his capabilities. The trucker will be able to afford to sleep in the car: it will be driven by the intellect, which will wake up the brain at a critical moment.

How to connect to the brain? We have all the technical means. Moreover, hundreds of thousands of people are already walking with such electrodes for medical reasons. To detect the focus of an epileptic seizure and to stop it, devices are installed that record the electrical activity of the brain. As soon as the electrodes notice signs of an attack in the hippocampus, they stop it. In the USA there are laboratories in which such devices are implanted: a bone is opened, and a plate with electrodes is inserted into the cortex by one and a half millimeters, to its middle. Then another die is installed, a rod is brought close to it, a button is pressed, and it abruptly, with great acceleration, hits the die so that it enters the bark by one and a half millimeters. Then all unnecessary devices are removed, the bone is sutured, and only a small connector remains. A special manipulator, coding for the electronic activity of the brain, enables a person to control, for example, a robotic arm. But this is trained with great difficulty: it takes a person several years to learn how to control such objects.

Why are electrodes implanted into the motor cortex? If the motor cortex controls the hand, it means that you need to receive commands from there that control the manipulator. But these neurons are used to controlling the hand, the device of which is fundamentally different from the manipulator. Professor Richard Anderson came up with the idea of implanting electrodes in the area where the plan of action is born, but drivers for controlling the motion drives have not yet been developed. He implanted neurons in the parietal region, at the intersection of the auditory, visual and motor parts. Scientists even succeeded in two-way contact with the brain: a metal arm was developed on which sensors stimulating the brain were installed. The brain has learned to distinguish between stimulation of each finger separately.

Another way is a non-invasive connection, in which the electrodes are located on the surface of the head: what clinics call an electroencephalogram. A grid of electrodes is created, in which each electrode contains a microcircuit, an amplifier. The network can be wired or wireless; information goes directly to the computer. A person makes a mental effort, changes in the potentials of his brain are monitored, classified and deciphered. After recognition and classification, the information is fed to the appropriate devices - manipulators.

Another move is the socialization of patients with motor and speech disorders. In the Neurochat project, a matrix with letters is placed in front of the patient. Its columns and rows are highlighted, and if the selection falls on the line the person needs, the electroencephalogram reads a slightly different reaction. The same thing happens with the column, and the letter that the person needs is found at the intersection. The system reliability at the moment is 95%. It was necessary to make sure that the patient simply connected to the Internet and performed any tasks, so not only letters were added to the matrix, but also icons denoting certain commands. Recently, a bridge was built between Moscow and Los Angeles: patients from local clinics were able to establish contact through correspondence.

The latest development in the field of contacts with the brain is neurosymbiotic clusters, which are controlled not by letters, but by the memory cells of a machine. If we take eight cells, or one byte, then with such a contact we can select one of the cells and write a unit of information there. Thus, we communicate with the computer, writing down the same "40265" into it. The cells contain both the values that need to be operated on and the procedures that need to be applied to these cells. So - without invading the brain, but from its surface - you can operate a computer. Material scientists came up with a very thin wire, five microns, insulated along its entire length, and electrical potential sensors were placed in its nodes. The wire is very elastic: it can be thrown over an object with any relief and thus collect an electric field from any, the smallest surface. This mesh can be mixed with the gel, put the mixture into a syringe and injected into the mouse's head, where it will straighten out and sit between the lobes of the brain. But the mixture cannot get into the brain itself, so the new idea is to inject a mesh into the brain when it is just starting to form, in the embryonic stage. Then it will be in the mass of the brain, and cells will begin to grow through it. So we get an armored brain with a cable. Such a brain can quickly figure out in which area it is necessary to change the potential for the computer to perform certain tasks or write information to its cells, because it interacts with the electrodes from birth. And this is full contact.