future. We like to build guesses about where we will be brought by evolution. We try to present what will be our equipment and technologies in decades and centuries. And we dream how we will meet reasonable alien beings who promoted much further us. Recently there is something it seems merges these reflections and imaginations. Speaking about the evolution, many futurologists predict singularity (singularity, originality). They declare: soon there will come time when computers become so powerful that will be able to reproduce human consciousness, or completely to absorb it. In parallel some seers believe that reasonable life which we will meet in the Universe, will be machine, instead of existing in humanoid bags with meat what we are.
These reflections can propose to us the possible solution of paradox of Fermi concerning which for a long time there are disputes. It is a question of seeming lack of reasonable alien life, despite that fact that such life is possible. If Artificial intelligence is an inevitable terminal point of technologies and biology, perhaps, aliens — it is exclusively developed cars — so perfect and advanced, so moved away from acquaintances to us biological forms that we do not recognize them even if we will see. In the same way we can imagine that interstellar communication between cars is so optimized and so well ciphered that we cannot distinguish it from noise. From such point of view, seeming lack of reasonable life in space can be the illusion caused by our own backwardness.
In forecasts of futurologists there is also deeper meaning. Our ideas of emergence of reasonable cars expose our imaginations (often unexpressed) that such perfection: not soft and biological, as we, and something firm, digital and for certain the incredibly powerful. For some people such future is the hope and greatness future. For others this future of fear and enslavement. In both cases we recognize that cars — it is deification of evolution of consciousness.
At first sight, the logic of such assumptions of space machine reason seems very solid. Extrapolation of a trajectory of our technical evolution says that in process of improvement and complication of computer technologies abilities and possibilities of our biological minds and bodies can become all less attractive. At some point we will want to move to the new cover constructed on order according to our imagination. In the same way, this technological trajectory can bring us there where we will create Artificial intelligence which or will be indifferent in relation to us, or will overtake and will subordinate us (or perhaps will simply crush).
Supporters of such arguments claim that the biology is not able to provide existence of all extraterrestrial civilizations and a human civilization in the far future. The difficulties interfaced to development of a space, are enormous, both in spatial, and in the provisional plan. To become an interstellar look, to us strong and reliable cars, instead of weak proteinaceous organisms with a pathetic expiration date in the fragility can be demanded. The car can live eternally and ideally reproduce itself, without suffering from the natural evolution inclined to mistakes, different flexibility and variability. The forms of life projecting, can adapt themselves for features of environment also. During life of one generation they can adapt for huge interstellar periods and spaces, and also for surrounding conditions of the extraterrestrial worlds.
If to connect all this together, there can quite be an impression that mankind — it simply instant, a transient phase. People perceive these ideas rather seriously, and such influential figures as Elon Musk (Elon Musk) and Stephen Hawking (Stephen Hawking), about dangers all-consuming Artificial intelligence openly warns. At the same time kibernetik Kurtsveyl Raymond did (Ray Kurzweil) a lot of noise by the works and conferences where questions of coming nearer singularity are considered. But whether really live organisms seek to become cleverer and are more durable? And whether really biological reason is the general deadlock, doomed to concede superiority to domination of cars?
fashionable descriptions of inevitable triumph of machine reason contain a lot of very important prejudices and parcels which do not allow to turn them into reality. Absolutely not clearly, whether modern computer technologies to singularity or by the grandiose moment of our greatest superiority as look conduct us. And all the same, the future can be tremendous.
Some of these extravagant ideas arose on the basis of amazing assumptions John fon Neyman (John von Neumann) about the self-replicating machine guns stated in the book "Theory of Self-replicating Machine Guns" published after his death (1966) This work helped to approve the concept about cars which create new cars, and there is it in the progression conducting to uncontrollable explosion, capable it is simple to destroy other forms of life if they follow ways of cars. John fon Neyman also considered a question of how such cars will be able to model and reproduce part of functions and actions of human nervous cages.
Since then development electronic engineers and communications definitely made huge impact on everyday life of the person, and even on how we solve problems and we think of new questions and problems. Who from us in the modern world did not google Internet in a network in search of the answer to a question, without having tried even to think of the answer independently or to ask other person? The part of our collective knowledge is loaded today into an ubiquitous cloud of data. Importance and the importance of individual knowledge and outlook gradually decreases. It is possible even that in the course of it value of the personal experience relating to specialization decreases.
it is unclear to what it will lead us. If want, it is not excluded that we in respect of consciousness move to a condition of a beehive, the collective organism more similar to a colony of termites or on mouse pack in a hole. Instead of developments the intelligence and consciousness, we, seemingly, block a way to the raw entrance data, and we learn to be more and more passive. The pessimist would see in it a picture of how our reason becomes deaf, turning into a component of the pack which have become reserved, instead of into group of constantly being improved ingenious persons.
the History teachesus that it is almost impossible to foresee long-term consequences of revolutionary technologies. Here it is possible to give one example. When at the end of the 1700th the steam-engine with piston back and forth motion was invented, it changed mankind life. Anybody could not predict it. In the same way, nobody could predict emergence of an internal combustion engine and electricity which later turned 150 years a steam-engine into an outdated relict. Equally, nobody could foresee that burning of hydrocarbons can seriously do much harm to our look, having changed structure of the terrestrial atmosphere.
We also cannot give strong evidences of that our kind of reason it something bigger, than strange result of evolution of billions years, and that in space scales such evolution is optimum. (To be fair it is necessary to tell that there are no powerful proofs of the opposite also — the data specifying that we a strange quirk of evolution.) The positive moment here consists that when we try to extrapolate our feelings from consciousness and intelligence to offer a certain reality about a condition of extraterrestrial reason and its motives, it is extremely difficult to us to make it.
Such argument sounds extremely painfully. It turns out that we grow stupid, we cannot predict own future, and we have no concept, what reasonable beings exist in space (and whether exist in general). However I claim that there are also light spots because these actions of forces of introspection force us to look in the face certain severe, but to surprising realities of our culture and our technologies.
of One of such realities is the energetizm question (the doctrine considering a matter as power education). This subject was lifted by John fon Neyman, but in the conversations futurologists often ignored. In computer design a key factor is the ratio of computing capacity and power consumption. Sometimes this ratio measure in calculations on joule. Microprocessors today all become more difficult, and silicon architecture less and less (now it is measured by tens nanometers). And nevertheless, efficiency indicators still improve. As a result the ratio of calculations steadily increases every year by joule.
However, the increase is more and more slowed down every year by. Some researchers even claim that "wall" gets in the way of processor architecture somewhere in the future energy efficiencies, a component approximately the 10th gigabyte calculations on joule on such actions as simple multiplication.
This serious obstacle for the present Artificial intelligence and for the equipment of loading of a brain. If to estimate necessary for overcoming of this wall computing power (measured by the speed and complexity of operations), it turns out that energy efficiency has to exceed wall indicators in billion times.
can be stated It and differently. Our brain consumes the 20 watts energy. If you want to load yourselves entirely into the car, using today's computer technologies, you need a power source approximately the same power what is developed by the Chinese hydroelectric power station "Three gorges", being the largest in the world. To transfer to a computer form all 7,3 billion live minds, the 140 thousand energy petavatt is required. It approximately in 800 most times the solar energy getting to the top layers of the terrestrial atmosphere. Yes, before transition of mankind to a transcendental state to us while it is far.
One of possible decisions — to address to neyromorfny architecture. Such silicon designs which imitate some aspects of real biological neurons and their communications. The researcher Khasler Jennifer (Jennifer Hasler) from Institute of technology State of Georgia and her adherents believe that if the neyromorfny system is made correctly, it can reduce needs for energy at artificial system similar to a brain at least by four orders. Unfortunately, even after such big jump there will be a gaping hole in energy efficiencies about 100 thousand, and only after its overcoming it will be possible to come to level of a human brain.
it is unconditional, in the history of computer technologies it is full allegedly of absolute obstacles which fall year after year. Therefore to lose optimism yet does not follow. However that nothing from this can be considered as a reality is very important. It can quite appear that silicon and its colleagues simply is not able to imprint all complexity, depth and extraordinary efficiency of a modern human brain as them neither form and nor put.
Techno optimists like to offeralternative, speaking about possibility of Quantum calculations in which coinciding quantum conditions of atoms or systems instead of traditional computer transistors are used. Supporters of such alternative say that the magnificent computing capacity provided with imposing of states, can solve energy and speed problems, having put us on a way of creation of superreason.
the Universal Quantum computer Turing can possess in fact a boundless computing resource, at least, on paper. The British physicist David Eliezer Doych (David Deutsch) brilliant and a little crafty formulated this idea in the work "The quantum theory, Chercha-Turing's principle and the universal Quantum computer" (1985) . That is remarkable, he suggested the reader to find out details of how to solve this problem.
Rather universal computer theoretically can imitate with any demanded accuracy any final physical system, including, a brain — and others Quantum computer, time to that went. In a quantum field modeling can be carried out in parallel in large volumes, and to carry out probabilistic tests with a magnificent speed. But despite enormous laboratory and theoretical progress of the last years, practical implementation of such concepts — extremely complex challenge. Though some examples of practical application of offered Quantum calculations exist, say, kontekstualizirovanny search which fully fits into a framework of cognitive calculations, (it is the best example of numerous efforts on creation Artificial intelligence), to the real reasonable Artificial intelligence is still very far.
A actively conduct still today debate about, whether Artificial intelligence, created by analogy to the human will be able to work in general.
Here the problem energy efficiencies too raises the head. To manipulate the main unit of calculations what is Kubit (Kubit) — be it atom or any other quantum object — not enough energy is necessary. But it is very heavy to connect together all components Quantum computer (having delicately kept all these thin quantum states). For this purpose numerous systems of providing and engineering development which will greedy absorb energy will be necessary. We even approximately do not know what in the real world will be a ratio between calculations and energy at Quantum calculations.
there are also other factors causing not smaller alarm. Quantum computer on "n" Kubit for one cycle can carry out 2n calculations, but for ensuring these calculations it is required to solve the most difficult problem with a data flow. For modeling of all our Universe consisting of 1089 particles and photons, by some calculations can suffice 296 Kubit.
But how, devil take it, to enter all these 1089 initial states? Even more difficultly another. How to choose the correct decisions from quantum modellings? It will be a little simpler to model a human brain, but all the same in this case it is necessary to present in a quantitative form and to initiate at least 1014 neural communications (them approximately so much at us in the head) to begin process of calculations. Allegedly we need a quantum brain with very high capacity, and the touch interface with world around with high-quality reproduction. This one more unknown, and is possible also, an insuperable barrier.
Frankly speaking, I simplify those methods and technological tricks which can be used. My ideas of the future can be too limited. Nevertheless, I think that there are bases for more reserved reaction to optimistic forecasts about Artificial intelligence human level. We have to recognize: though to create the equipment on maintenance of reason, comparable or even surpassing human, probably, it can not provide that exponential growth of calculations about which so often speak.
In other words, mathematical bases of exponential being improved machine intelligence can be sensible and thorough, and here practical barriers — is insuperable the high.
to understand to what it can lead, I (a little hypocritically) will pull out a leaf from the book of futurologists and I will carry out mad extrapolation. I want to understand that happens if we synthesize idea of slow developments machine intelligence and Fermi's paradox. To make it it would be interesting and very informative.
Let's assume that the certain developed space reason successfully turns into a machine form, or it is overtaken by super-clever machine creations which, however, do not surpass it in a geometrical progression. What will occur further?
As these cars are limited to efficiency indicators, there is an opportunity that they as a result will start looking back at the past in search of new ways and tricks for advance forward. They will know (as it is known to us) that the biology works, and that it works very well. According to estimates of some scientists, the brain of the modern person has computing limits, but for alteration of such difficult body the car which is only slightly cleverer than it can be demanded. In other words, there can be more optimum trajectory which will take away us away from the car back to the biology possessing surprising energy efficiency.
Besides, there is no guarantee that the machine reason will be absolutely rational, and will be able to be that. To interact from the difficult Universe where the mathematics contains unprovable theorems, a little bit of irrationality can be necessary. Today we usually build assumptions that the future of our reason in any other form, silicon, or perhaps quantum as we consider that these forms surpass a flesh. Perhaps, the same performance is played with any other reason. The car can want to become again biological creation for practical reasons of an energetizm, or for other reasons which we are not able to present and understand.
If life is widespread and regularly gets reasonable forms, probably, we live in the Universe where the reason was in the past, and will arise in the future. Universe 13,8 billion years, and our galaxy almost same ancient. Stars and planets are formed the last 13 billions years. There are no convincing bases to think that space did not make anything interesting for those eight billions years that passed till the birth of our Solar system. We can sometime decide that for future reason on Earth the biology, instead of machine calculations is necessary. Probably, the incalculable set of reasons already passed billions years ago through this transition period.
These early reasons very long time ago could be at such end of the resources where made the decision to return from a machine form to the biological. If this is so, comes back also Fermi's paradox: where now all these aliens? The simple answer consists that they could be fenced off as interstellar flights are exclusively difficult, especially for physical, biological beings. Probably, the reason exists, but having returned to a biological form, it paid for it, having appeared in isolation.
These early reasons could create once mega-designs and start fruits of the space engineering thought from a star to a star. Perhaps, something else remained there, and is not excluded that we just about will find it by means of the constantly upgraded astronomical devices. The recent agiotage concerning KIC 8462852 star which brightness changes in the way inexplicable for known natural mechanisms, is based on recognition of that our devices are rather sensitive today to investigate such opportunities. Perhaps, alien civilizations receded, having passed into a lonely biological state, and the remains of designs of their mechanical era collapsed under the influence of space radiation, evaporations and blowing-up star dirt.
it is possible, our present state it to space measures a short interval between the first generation of machine reason and the following generation. The machine intelligence or transcendental state in other parts of a galaxy could be short-lived as interstellar force; the previous could disappear already, and the following did not appear yet. Probably, they had no time for visit to modern mankind. Perhaps, the machine reason dreams to become again biological, having returned to the isolated state on huge open spaces of interstellar space. Our technological future can be such — refusal of machine imaginations and return to quieter, but also more effective organic existence.
Should not be ashamed, recognizing exclusively hypothetical nature of these ideas. Also there is something special in those questions which generate these ideas. We investigate the possible future for ourselves. It is possible to imagine that the Universe already prompts to us what we have options. Such actions on research themselves are not similar on anything in human activity, and even because of one it it is worth paying attention to them.
Caleb Sharf — the research supervisor on the astrobiology, working at Kolumbiysky university in New York. He is the author of the book Extrasolar Planets and Astrobiology (Extra solar planets and astrobiology) and the winner of an award of Shambli for works on astronomy.
Source: INOSMI.RU
the Opinion of the author not always coincides with a position redaktsii.
