The Last Battle for the Singularity
Among futurologists and transhumanists, they talk and write a lot about the Technological Singularity. It is believed that this is such a point in development when technological development becomes in principle uncontrollable and irreversible, which gives rise to radical changes in the nature of human civilization. Usually, its onset is associated with the emergence of a technology of strong artificial intelligence, which is able to perfect itself - and ultimately can enter into an “unrestrained reaction” of self-improvement cycles, ultimately creating a superintelligence that surpasses the intelligence of all mankind (wiki).
It is also believed that we are on the verge of a singularity. Well, just about, it remains to wait a couple of decades, and it will happen. However, there is no strong AI yet. Various new technologies are emerging, but relatively slowly. Forecasts of futurologists are not in a hurry to come true, or even turn out to be pipe dreams. And everything that happens looks as if nothing special will happen - neither in the next decade, nor in the next century... Are our hopes in vain? And can we do something to really bring the Singularity closer?
Note: There are no practical solutions or suggestions in this article; no answers - there are only questions. We can say that this article is not independent, it is a kind of designation of the problem and at the same time an introduction to a series of articles or even a book. But the main thing is an invitation to reflection and discussion.
The battlefield is human
One of the key problems is that as we move forward, the person himself, the human brain, his ability to receive, process and issue information become more and more a bottleneck. Between the capabilities of the human brain and the tasks that need to be addressed for the movement of progress, the gap is rapidly growing. Man is the smartest creature on the planet, but to solve modern scientific super tasks, we increasingly rest on the bottleneck of the human mind: the speed of receiving, processing and issuing information is too low.
A person receives information from the outside world with at least 6 senses (vision, hearing, taste, smell, touch, vestibular apparatus). Of these organs, only two are involved in obtaining abstract information and exchanging abstract information - vision and hearing (and partially touch - for example, Braille for the blind).
The widest channel for obtaining information is vision. It is believed that with the help of the eyes we get about 80% of all information. According to rough estimates, the resolution of the retina of each eye is approximately 120-140 megapixels, the frequency of perception is about 25 frames per second. From the point of view of modern technologies, these are extremely high indicators (you can estimate the video stream as about 6 gigapixels per second). You will not find cameras with such permission on sale - if such cameras exist, then in single copies, they cost a lot of money and are used exclusively in research problems.
Hearing provides us with about 16% of all information. The human ear is capable of perceiving signals with frequencies from 16 to 20,000 Hz and in a rather large range of amplitudes - from 0 to the “pain threshold” of 120.140dB. Also, sound pulses replacing each other with a frequency of more than 16 Hz we perceive as continuous sound. However, in contrast to vision, hearing is almost “one-dimensional”, therefore the amount of information entering the brain from the organs of hearing is somewhat less than from the organs of vision.
The brain is made up of about 100 billion neurons. For comparison, the latest processors have about 2 billion transistors, and a transistor is a much simpler structure than a neuron. However, the transmission rate of nerve impulses between neurons is low: from 0.5 to 120 meters per second. On average, 10 pulses per second pass through one synapse, i.e. operating frequency of the order of 10Hz. This is extremely small compared to the GHz frequencies of modern processors.
The incredible capabilities of the brain are provided only by unprecedented parallelism - unlike a single processor with only a few nuclei, a huge number of neurons work in the brain at the same time. Up to 5,000 synapses can be located on the body of one neuron, and taking into account the total number of neurons, even though they are not all involved, huge computing power is obtained.
According to scientists, the amount of memory in our brain is about 1 petabyte of information (for example, the Google search engine processes about 24 petabytes of data daily).
At the same time, the human brain consumes only 25 watts of energy, according to this parameter it can be called the most efficient computing device on Earth.
However, no matter how powerful the brain, all this power is literally locked in the skull. The first problem is how fast people are able to exchange abstract (mainly textual) information with the outside world.
Let's start with the view
The average person’s reading speed is from 200 to 250 words per minute (for English), 128-180 words per minute (for Russian), depending on the language (data for non-hieroglyphic languages).
In characters, the average speed is 1500 characters per minute. If we arbitrarily take one character for one byte (in the scientific and technical literature that interests us mainly, there are usually more different characters than in the fiction, but nevertheless in non-hieroglyphic languages we can fit roughly in bytes), we get a channel of 200 bits/sec. This is slower than the very first modems!
Speed reading technologies do not significantly increase this speed. For example, US President John F. Kennedy could read at a speed of approximately 1,200 words per minute, which corresponds to approximately 800 bps.
The speed of perception of human speech
According to studies, the optimal rate of speech for reading audio books in English corresponds to 150-160 words per minute. For a personal conversation - 190 words per minute. As you can see, the speed roughly corresponds to the reading speed, so the use of the audio channel does not add any special advantages.
And if a person can always just watch and listen (if he does not sleep), then focused work with textual information is a more difficult task. In fact, can you read constantly, especially if it is not an exciting adventure novel, but difficult to understand scientific or technical literature?
Information output speed
Things are even worse with the channel for abstract information output from the brain to the outside world. The most common way in the modern world is typing on a computer keyboard. The average typing speed is 200 characters per minute, which corresponds to only 27 bps! Perhaps, speech recognition programs would have helped a bit - because the speed of speech on average corresponds to the speed of speech perception, which is more than the speed of typing.
All this is very good for simple everyday life, similar to the life of our evolutionary ancestors - primates, and is quite suitable for a quality life of many generations of people; but not so good for achieving Singularity.
The second significant problem is the fundamental inability of the human mind to work with complex abstract information.
When processing "abstract" information, the brain does not operate with such huge streams of incoming visual information that provides us with vision, but is limited to miserable hundreds of bits per second. And this concerns not only the speed of obtaining information, but also the speed of understanding it.
The amount of brain memory is huge, but only about seven objects can be contained in a short-term human memory.
Why is this happening? Biologically, man is still an animal. The human brain is still designed to ensure the survival of the individual and species, mainly in the wild savannah. The human brain was not originally intended for reading, typing, working with complex abstractions.Our eyesight is capable of processing huge amounts of natural visual information very quickly - but this is necessary in order to quickly notice predators lurking in the jungle and other dangers, as well as to find a variety of food; but the speed of our reading is slower than the very first modems, released in the 80s of the 20th century! And the speed of issuing information is even less.
Just as people are upright, but the human skeleton is still not fully adapted to upright posture (hence many diseases of the musculoskeletal system) - similarly, people are intelligent, but still not fully designed for scientific thinking.
And this is the narrow neck of our civilization, blocking our path to the Singularity.
Imperfection of the biological nature of man
The biological nature of man also leaves its mark.
Fatigue, procrastination, and exhaustion are characteristic of man. The human brain cannot solve complex scientific problems 24 hours a day. People need rest, distraction, a change of scenery. Our eyes are also subject to fatigue; we cannot read too fast or too much. People periodically want to sleep, they may have a headache, they are prone to diseases...
The peak of our development falls on the age of 20-40 years; before this - a period of childhood, when the brain is flexible, but knowledge in it is not enough; after - a period of aging, when there is less and less strength and health every day. And most are immersed in family affairs and problems, science and progress are no longer of interest to them.
Battlefield - Civilization
Huge amount of knowledge
There is too much information. We, as a civilization, have enormous knowledge, but our human resources are not enough, even to simply get acquainted with all this knowledge; the current amount of information is so huge that it’s physically impossible for a particular person to cover it, even for a lifetime; do not read, do not listen, do not view, do not realize...
Our libraries are full of books, but can we read these books? There are 130 million books in the world, but you cannot read even a hundredth of a percent of that number! A human life is simply not enough for this.
There is a huge amount of knowledge in the world that is simply inaccessible - because they are buried somewhere, in all this heap of information. As a result, there is duplication and a lot of unnecessary work that people often do, without even knowing about each other.
One of the most important components of the scientific and technological revolution - the computer revolution, has brought computers and the global Internet into our lives. Without exaggeration, the Internet is a breakthrough on the same scale as the emergence of writing once. It is believed that humanity has accumulated 2 trillion gigabytes of data, and the annual volume of Internet traffic is about 3 zettabytes. 4.54 billion users have access to the Network, each of which on average spends about 7 hours a day on the network. These are fantastic numbers, but what do we have in reality?
Most of the information does not represent any value for the achievement of the Singularity. Not only that - this information is of a closed, unstructured nature, for it basically only one operation is available - copying and forwarding. Most of the traffic - video, audio and images - are human-oriented blobs with which the computer basically can not do anything meaningful.
But even simple textual information remains obscure for computers. Search engines are the most complex systems, in fact only primitive crutches adapted for primitive human text queries.
The problem is that almost all information in the world is people-oriented. For those very people with all their human defects, which we examined above.
Printed and electronic books - text in human language. Images, audio and video are generally oriented towards the human organs of vision and hearing, and for a computer in the general case they are just a jumble of bits. All this information is focused on the human brain that has inherited a different type of perception.
This is a huge problem. The computer operates with terabytes of information and has no idea about them at all.Yes, computers make it easier for us to work with special information - such as circuit diagrams, drawings or source codes for programs, but they cannot work with general information. Basically, they can only entertain us.
There is still a huge gap between man and computer.
Avalanche of difficulty
If we consider only scientific information, then here we are faced with another problem: the huge, incredible complexity of modern scientific problems. The times when genius singles could make scientific discoveries in home laboratories are long gone. Modern science requires professional laboratories, sophisticated and ultra-expensive equipment - such as the Large Hadron Collider or the LIGO interferometer, and the coordinated work of a large number of professionals. This is not possible without the support of states or major transnational corporations.
Speaking about the Singularity and the most important tasks of Transhumanism - the victory over old age and death and gaining control over the biological nature of man, one cannot but mention the enormous complexity of biological systems.
One DNA molecule on average consists of 100 billion atoms. A biological cell contains an average of about 100 trillion atoms. The length of all DNA molecules of a double set of chromosomes in one human cell is approximately 2 m. The number of conformational variants of a single molecule of a typical protein may be greater than the number of atoms in the Universe (source: Introduction to General, Organic and Biochemistry).
Some examples of complexity
Here is how Professor Michael Panteleev describes the biological complexity of
The transformation of one substance into another under the action of a single enzyme can often be described by a fairly simple formula with good accuracy, although there are many exceptions. The two simplest enzymes, acting on each other, already form a system with non-trivial behavior: stabilization, the presence of thresholds and switching, and autowaves can be observed in it. Without a detailed mathematical analysis, it’s hard to say how she will behave. In a system of three enzymes, we already get a complete “bouquet” of dynamic modes, including chaos and strange attractors.(source: Defeating old age, cancer and heart attacks; Mikhail Panteleev about biological complexity )
These are just three enzymes, but the simplest of the real biological systems of the cell is a network of hundreds of different enzymes, cofactors, metabolites, which affect each other in thousands of ways.
In 2005, employees of the Los Alamos National Laboratory managed to create a dynamic model of the work of the ribosome synthesizing a protein molecule. For this, 768 microprocessors were required, which worked for 260 days. During this time, it was possible to “shoot” 20 million frames, reflecting only 2 nanoseconds from the life of the ribosome.
( source )
In 2011, Chinese scientists created a simulation of the H1N1 virus at the atomic level. A Mole-8.5 GPU-based system with over 2,200 NVIDIA Tesla GPUs is capable of simulating 770 picoseconds per day with an integration time step of 1 femtosecond for 300 million atoms or radicals.
Of course, full atomic modeling is far from always advisable; in most practical cases, one can restrict oneself to some approximations. However, the enormous complexity of biological systems is manifested at higher levels.
For example, in 2013, it took 40 minutes to simulate 1 second of 1% of the human brain (1.73 billion nerve cells and 10.4 trillion synapses) on a cluster of 82,944 10-petaflops K computer processors.
( source )
And in the project brain modeling in 2018, a modern supercomputer consisting of a million ARM9 cores capable of processing 200 trillion operations in second, can simulate in real time only one percent of the total number of neurons (1 billion, not 100 billion).
Yes, there is progress; and I have no doubt that pretty soon it will be possible to model a full-fledged bacterium at the atomic level, and a multicellular organism, and the entire human brain, and much more; and if there is not enough processing power of ordinary supercomputers, quantum ones will help. Rather, the problem is different: in the transition of quantity into quality. In building a bridge from the level of human understanding (which will not become smarter in principle) to those inconceivably huge volumes of information that are mechanically processed by modern supercomputers - without any understanding on their part.
This article does not discover anything new; in fact, this is only an introduction to the problem and a look at the achievement of the Singularity from a different angle.
Most likely, the technological singularity will not occur on its own, as a simple result of a spontaneous transition of quantity into quality, huge processing power into some kind of superintelligence, capable of infinitely self-improvement. It seems that in order to achieve the Singularity, we need to make some focused efforts . Which ones? This is a separate topic, very large and complex. I deliberately did not include here any mention of possible actions (although of course I have certain thoughts on this matter) - otherwise it would not be an article, but a book :) But I hope this article will serve as a kind of introduction and start for further research.