The computer and the brain according to Asimov

Good scientists can also change their minds if the facts demand it. Asimov changed his approach in the field of artificial intelligence, when the personal computing revolution showed that the computer is nothing but another external aid device and not a substitute for human reason

Isaac Asimov is known to us as a science fiction writer. His science fiction books are populated by supercomputers with enormous databases and computing power, which in some cases have human-like qualities. His books also star humanoid robots, with seemingly human qualities, but programmed using the three laws of robotics, so that they can never harm a human being and are thus better than humans.

But Asimov was also a science writer. The hundreds of books he wrote include, besides the science fiction books, also books such as the genetic code and the encyclopedia of science and technology. He also wrote many articles that were published in various settings. We will quote here two of his articles comparing the computer and the human brain. These two articles express opposing views.

In the article "Our Intelligent Servants" Asimov argues for a principled comparison between the operation of the computer and the operation of the brain and predicts that the computers will catch up with us in intelligence and even surpass us. He also comes up with a wild idea that computers will take over us and the world. The idea expressed in this article is in line with Asimov's science fiction books and stories and also reflects speculative mindsets that also prevailed among some of the scientific community in the field of computers at the time the article was written.

In the article "Reflections on Personal Computers", Asimov emphasizes the distinction between the skills of the computer and the skills of the human brain. Has Asimov changed his mind? - Definitely yes. In science, politics, as well as in other fields, people change their opinions and positions according to the circumstances and the information they have and we should not call them fickle.

The article "Reflections on Personal Computers" reflects the limitations we know today in the applications of artificial intelligence, each of which represents a certain facet of human intelligence in a limited way. In other words: the applications of artificial intelligence are very useful, but from here to breathing a human spirit into them - the road is long. Between the lines of the article, the disappointment with the translation machine - a computer that translates from language to language - which the computer and language scientists worked hard on at the time and placed many hopes on it that were disappointed, probably also stands out.

It seems to us that the difference between a computer and a brain is only quantitative, a quantity above a certain threshold that we will call the threshold of consciousness, it becomes a quality. In recent years, we have been talking about supercomputers, expert systems, parallel processing, fuzzy logic, biological, molecular, quantum computers and more. But today's developments are still far from the threshold of consciousness. The difference seems to start at the level of the individual logical cell. While a logic cell contains one or two inputs, maybe three, a typical nerve cell has about a thousand inputs. Its output, like that of the logic cell, is zero or one, but this result is a complicated logical processing of all the thousand inputs, including variables of the timing of the signals at the different inputs or the rate of the signals at one input. One of the characteristics of the operation of the brain, as opposed to a computer, is chaos. Because of the huge collection of input factors (senses and prior memory), and also because of the multitude of inputs at the level of the single nerve cell - it is difficult to predict the output - the behavior, because a small change in the inputs may cause huge changes in the output. Indeed, there are facts in the field of brain research and psychiatry that indicate that certain mental disorders are characterized by the fact that brain activity is less chaotic and more orderly.

It is possible that the developments in the next ten to twenty years in the fields of electronics and computer science, together with progress in understanding the intricacies of the brain, will allow us to create truly intelligent computer systems. All that will be left for us is to find suitable applications for such systems. In any case, these will be computers of the type we are used to meeting in science fiction books and movies. Whereas Asimov, if he were to write another article on the subject, would return to his position expressed in the first article, namely: computers will indeed surpass the human mind.

Our intelligent servants

Isaac Asimov (Translator: Aryeh Seter)

In this article written about thirty-five years ago, the author claims that computers will reach the level of the human brain and even surpass it.

Robots don't need to be very intelligent. If a robot can follow simple commands and do the housework, or if it is able to operate simple machines in a sequence of repetitive actions - this will definitely satisfy us. It is not easy to build a robot, because we have to squeeze a compact computer into its skull - if we want the robot to have a human-like shape. And it is difficult to build a complex and complicated computer that will also be the size of a human brain.

Let's leave the robots thing. Why should we strive to make the computer smaller? Computer components are getting smaller and smaller - from vacuum tubes through transistors to silicon pieces of integrated circuits. What will happen if, in addition to reducing the components, we increase the structure of the computer?

If our brain were big, it would start to lose its efficiency, because the nervous attacks don't move that fast. The fastest nerves transmit attacks at a speed of only about a hundred meters per second. A seizure can flash from one end of the brain to the other in 440th of a second. If our brain, try to imagine, was 15 km in size, it would take the nervous attack 2.5 minutes to pass along it. The great complexity that would have been made possible by the enormous size of such a brain, would have failed in its function, due to the long wait for the transfer of the information and its processing.

In computers, on the other hand, the electric shocks make their way at a speed of 300,000 km per second. A huge computer with a width of 600 km, will transfer signals from end to end, despite its enormous size, in 500th of a second.

In this respect at least, even such a huge computer - like a large asteroid - would be able to process information as fast as the human brain.

Let's imagine that the computer components will get smaller and smaller, the connections between them will be more complicated and the computers themselves will be bigger - isn't it possible that the computers will then be able to do everything that the human mind is capable of? Is there any theoretical upper limit to a computer's intelligence? I have not heard of such a limit. It seems to me that whenever we manage to squeeze more into a given volume, the computer is able to do more. As we increase the size of the computer, when the level of complexity and density remains the same - even then the computer is able to do more.

If in the end we manage to create a computer that is sufficiently complex and large enough, why not achieve a human level of intelligence?

I am sure that many will not believe this and will say, for example: "But how can a computer write a good symphony, create an important work of art, or develop a new scientific theory?".

The winning answer I usually tend to give to this question is - "And you can?" Even if the questioner is an ordinary person and he really can't, there are special people who are geniuses and they can. But their genius comes only from the fact that atoms and molecules inside their brains are arranged in a certain order. There is nothing in their minds but atoms and molecules!

If we arrange atoms and molecules in some complicated order inside a computer - it too will be able to produce ingenious works. And if the computer components are not as small as the human brain - then it will be bigger.

There will be people who say: "But computers can only do what they were programmed to do"; The answer to this would be - "True, but even minds are only able to do what they were programmed for - by their genes".

Part of the programming of the brain is its ability to learn and this will also be part of the programming of the future computer. And basically, if it is possible to build a computer that will be as intelligent as a person, why not make it more intelligent than the person?

Well, why not? Maybe this will be the next step in evolution! For three billion years, in a ceaseless process of experiment and confusion, structures of atoms and molecules were developed, which finally created - after gradual improvement, a species intelligent enough to be the next step in evolution - a step in terms of hundreds of years, maybe only tens... and then things will really start to move!

And if the computers become more intelligent than us, maybe they will have to take our place? They will be able to be kind, just as they are intelligent and instead of destroying us, they will let us exhaust ourselves and become extinct. Maybe they will keep some of us as pets or for the purpose of preserving our race.

See for yourself what we are doing to ourselves at this moment - to us and to all living beings on the planet we are on. Maybe it's time for us to pass. Perhaps the real danger is that by the time computers are developed to a level where they can replace us, it will be too late.
Think about it!

Reflections on personal computers

Isaac Asimov (Translator: Aryeh Seter)

In this article written about twenty years ago, Asimov repeats his position in the previous article and claims that the computer will not reach the capabilities of the human brain.

Any progress in the development of problem-solving aids raises hidden concerns. One can imagine the ancient Egyptian architects looking suspiciously at the first measuring rods and saying with displeasure: "People will develop a dependence on these mechanical mechanisms. They will forget how to measure distances with their examining eye and will rely on their judgment on pieces of wood that are not alive. Instead of reaching a slow, calculated and correct decision, they will reach a hasty and mistaken decision." I dare say that the appearance of the abacus, the Arabic numerals, the slide rule and the mechanical calculator also caused similar claims. And today, it is the personal computer and its "intelligent" programs that "release us from responsibility and cause our thinking to dry up".

All those aids for solving problems present us with a tautological problem; The solution exists and is embedded within the data used as input to the problem. Even when we are looking for a solution to 9x7, we have the information that actually gives the only answer - 63. We can look at the multiplication table, or if we have time, arrange 9 piles of 7 pebbles each and count all the stones. Although the multiplication table provides the answer faster and better (since the counting of the stones is prone to errors), the result will be the same in both methods. If we have plenty of time and don't have anything more important to do and know how to count carefully - we can give up the multiplication table - but no one will do that, because speed and the absence of errors are important.

It is the same with the personal computer. For office workers, as for scientists, the mechanical aids gave way to the multi-use electronic tool. The personal computer can store entire libraries on its high capacity drives and it can calculate billions of times faster than us and without error. However, when we load the computer with data and instructions, the answer is already contained in the input. We (without the computer), can start with the same information, manipulate it and arrive at the one and only result, provided we have enough time and are careful enough to avoid error. But because of the long time needed to solve problems, this option is out of the question.

We insist on using a computer and have built such complicated computer systems that we can no longer do without them. But the shadow of "GIGO" (garbage in - garbage out) always hovers over us. If your name was entered incorrectly in an address database, or if you made a mistake entering the numbers in a calculator - the result will be wrong. The personal computer is prone to the same kind of errors as any other calculation aid and will also give you wrong results in such cases. Sometimes it is dangerous to trust the result of the computer, but equally the finger following a row or column in the multiplication table - may be wrong. As always - the "failure" lies in the human factor.

There are many ways in which you can check results, but the best way to separate the bar from the chaff is simply knowing when the answer looks wrong. Here the human brain is much more efficient than the computer. Computers have an amazing ability to solve problems, but the human brain has an ability that surpasses that; Call it intuition if you will.

You can enter all the dictionaries in the world into the computer, along with all the grammar rules and all the exceptions to those rules and all the expressions that go beyond the rules. The computer will then be able to write quite well and in some cases also literature and poetry, but the results will unfortunately be devoid of any inspiration. I myself know only the simplest rules of grammar and rarely use them consciously. Despite this, I instinctively write correctly and I am proud to think that my writing is interesting. I know when something sounds right and when it doesn't and I can distinguish between them without hesitation, even when I write at breakneck speed. How can i do it? - I have no idea and since I don't know how I do it, I can't program a computer that played me doing it.

The human mind is not a numbers machine. He handles numbers with appalling slowness - and without simple aids like pencil and paper, errors are inevitable. The brain's specialty is intuition, insight, fantasy, imagination and creativity. The human mind is able to discover and perceive situations it has never experienced, or on the basis of uncertain information and partial understanding, draw logical and correct conclusions. A computer cannot do this, just as a person cannot summarize complicated equations in nanoseconds, or instantly calculate the value of transactions in a particular month. It is of course conceivable that computers will one day be able to demonstrate all the skills of the human brain - but not in the foreseeable future.

The human brain has 10 billion neurons and close to XNUMX billion supporting cells and they are all connected in an astronomical tangle of connections. A single neuron by itself is a complex system of proteins and transitions.

Why should we build a computer that will do what we do satisfactorily? Would we train a person from birth to do what the simplest computer does so well? I think not!

The personal computer, during its development, will thus remain a tool, perhaps more complicated, perhaps with an improved ability to examine the instructions given to it. But as complicated as it may be, I am satisfied if the computer will ever reach the level of intuition and creative power of the human spirit. In my opinion, in the future we will see computers and people, each representing completely different forms of intelligence, working in cooperation, rather than in competition, and achieving together, more than they could achieve individually.