Amount of information about AI, artificial intelligence, neural networks and various, chatting GPT and drawing MidJourney in various sources, and domestic, and foreign, over the last six months. Following the ozone holes, climate catastrophe and pandemic is the new, disturbing the masses attack.
What is it? True or not? What do they eat with? How to use? Harm or benefit? Or even that, and more together? This is worse than the atomic bomb and everyone will die, Or will I just be out of a job, but I will live? The intrigues of enemies or we again lagged behind a hundred years?
Let's try to figure it out and start with a retrospective. After all, what happens to humanity, has its own history. Something doesn't come from nothing.
"All things in nature happen mathematically"
In science, a priori information consists of the axioms (provisions, considered indisputable truths) and hypotheses (assumptions, subject to experimental verification). The axioms of mathematics were first formulated in Euclid's Elements.. There is an axiomatic of elementary geometry and many other sections. The basis of the deductive science of logic is the axioms of Aristotle. And further,, and further.
Life within the axioms is simple and clear, and it is not customary to doubt their inviolability. Mankind, as it were, is confidently moving along the high road of absolute truths to the complete victory of Reason..
But for Nature, the ultimate truth does not exist.. The history of knowledge is a story about the constant incompleteness of our knowledge of an infinitely complex reality and the inevitability of that, that some of them are wrong. It's about this 7 September 1930 year at the Mathematical Congress in Königsberg, 24-year-old Kurt Gödel said by saying, that any system of basic assertions (axiom), accepted as true without evidence, starting from a certain level of complexity or internally inconsistent, or incomplete, and 1931 published proofs of his incompleteness theorems.
Interest in the topic is growing.
In the mid-1930s, the work of mathematicians A. Church and A. Turing, which have become fundamental to computer computing, and the Church–Turing thesis, together with the theory of computation, forms the foundation of algorithmic problem solving and is directly related to the creation of artificial intelligence.
In Nazi Germany, engineer Konrad Zuse, applying the principles of binary logic, assembled real working electromechanical programmable computers from Z-series telephone relays, using perforated 35 mm film in command transmission. The Aerodynamics Research Institute of the Third Reich even managed to use Z to calculate the parameters of aircraft wings and flights of V-2 guided missiles.
AT 1943 year in America, neuroscientist Warren McCulloch and mathematician Walter Pitts publish an article, in which the computational model is described, based on the concept of an artificial neuron.
Then - the work of K. Shannon on information theory and the extraordinary Frank Rosenblatt, created the first neurocomputer "Mark-1" on the perceptron and neural network, and who gave a course of lectures under the disturbing title "The Theory of Brain Mechanisms".
Not only the West, but Russia also shone with minds. AND A. Visegradsky, AA. Lyapunov, L.T. Kuzin, like many other of our great compatriots, made a huge contribution to the development of the topic of AI.
Interesting, that everything described happened in the first half of the 20th century. Then there were periods of fading interest in this topic., which were replaced by the progress of technical developments in microelectronics and data storage systems.
In the end, in the mid-2000s, the "stars aligned". Scientists have received adequate mathematical models, learned how to create very powerful computers with huge databases and crazy speed.
Proud of your achievements, people decide for some reason, what a tool, working according to the established algorithm, easily solve all problems, difficulties and contradictions in our lives. 'Cause he's acting so fast, he knows everything, he is everywhere! Here she is, point of singularity! We will finally gain the power of the gods! He will do everything for us. You can lay on the sofa and relax.
Relaxed and literally for some 30 years from all the cracks climbed the consequences and problems. Strain and think or press a button? What solution?
The question of comparing the work of human consciousness and the machine created by him has been discussed in scientific circles for more than a decade.. How to form a machine understanding of reality? Can AI plan and learn?, communicate in natural language and have willpower, if it does not have an objective picture of the world around? And AI must think, As a person? Can he have his own thoughts?!
We are not the first to ask these questions..
The Soviet public on the pages of Literaturnaya Gazeta for five months 1962 led a discussion between "lyricists" and "physicists". The exchange of views ended with the publication of an article by the famous director of the Institute of Cybernetics of the Academy of Sciences of the Ukrainian SSR V.M.. Hlushkova. He spoke, but didn't answer. The search for truth has ceased, but precise and unifying definitions of “what is intelligence?' and 'what is consciousness?" still no.
About modern discussion, submitted for public discussion, in which specialists in human physiology would participate, artificial intelligence and neural networks, linguists and philosophers, united by the idea of patriotism and understanding of solving the problem for the Russian State, so far there is no word. There is no single approach, no single methodology, no philosophical basis under discussion.
Yes, there is a National Strategy for the Development of Artificial Intelligence for the period up to 2030 of the year with the application program of the Government of the Russian Federation "The concept of regulation of relations in the field of artificial intelligence technologies and robotics up to 2024 of the year" which notes: «Lack of unambiguous understanding of the content of the terms "artificial intelligence", "robot", "smart robot", "robotics", "intelligent agent" leads to terminological problems in the formation of regulation».
There are GOSTs: R 59277–2020. "Artificial Intelligence Systems. Classification of the artificial intelligence system” and Р59276–2020. "Artificial Intelligence Systems. Ways to build trust. General Provisions». These documents use phrases: "Human Cognitive Functions", "human intelligence", verb "imitate". But what is it? And what does "imitate" mean?? And if you imitate, to what extent? And when imitation becomes reality?
The military defense of Russia is not distracted by such trifles and is working hard on the "intellectualization of weapons", and the head of the Department for the Development of Artificial Intelligence Technologies of the Main Directorate for Innovative Development of the Russian Ministry of Defense Vasily Elistratov declares, that management simultaneously accompanies for subsequent implementation from above 500 projects, and employees of military-industrial complex enterprises have mastered the modern concepts of "artificial intelligence", "self-learning neural networks", "end-to-end technologies", "technical vision", "simple integrated solutions".
That is, specialists from the Ministry of Defense and the defense industry, without bothering, cheerfully operate with the expression "human intelligence" along with the concept of "artificial intelligence". Not everything is one?
I guess, what needs to be sorted out.
Is it really something else, much stronger and faster, than our consciousness or just our complicated algorithm? Tool, created for our benefit? the, which will speed up complex calculations by billions of times? Where is the limit?, not to be crossed?
Let's dig deeper.
We are not machines. We are harder.
AT 1950 The famous Alan Turing in the scientific journal Mind publishes an article entitled "Computer Science and Intelligence (Computing Machinery and Intelligence)», which in five years, at 1956 year, will be reprinted in 4 volume of "The World of Mathematics" Already. R. Newman under the famous title "Can a machine think?»
31 August 1955 g. group of scientists as part of John McCarthy (Dartmouth College), M. L. Minsk (Harvard University), n. Rochester (I.B.M. Corporation) and C.E. Shannon (Bell Telephone Laboratories) made a proposal to hold 1956 year at Dartmouth College Summer Research Project on Artificial Intelligence.
The financial support of the project was carried out, as usual, by the Rockefeller Foundation., always keeping his nose to the wind.
Issues for discussion:
1. Automatic computers.
2. How can a computer be programmed to use a language?
3. Neural networks.
4. Computation size theory.
5. Self improvement.
7. Randomness and creativity.
The discussion proceeded along two lines.:
1) application of information theory concepts to computers and brain models;
2) consistent approach to environment and brain modeling for automata.
Seminar agenda, as for me, has not lost its relevance even today.
The organizers of the project have, what: «All research should be based on a hypothesis, that every aspect of learning, or any other characteristic of intelligence, can in principle be described so accurately, that you can create a machine to simulate them».
John McCarthy himself as the author of the term "artificial intelligence" (AI), did not give him that human essence, brought by the Russian-speaking community.
Word «intelligence», applied to the English context, in which it is used, means "the ability to reason intelligently", not "intelligence". There is a separate word for this - "intellect".
The important thing. It is this formulation of the hypothesis that has become the goal setting for the vast majority of research in this area., defining tasks for many decades to come, that have been determined to achieve it..
To simplify, let's skip a small time interval, in which the theory, and practice evolved to test this hypothesis.
AT 1989 Sir Roger Penrose, British physicist, mathematician and philosopher, published the book "The New Mind of the King", where he considered the issues of quantum consciousness and the theory of strong and weak AI. Justifying the failure of the implementation of such a form, how strong AI, to prove your thesis, he used, among other things, Gödel's theorem as an argument for the existence of a fundamental difference between our brain and a computer.
The essence of his reasoning: «The computer acts strictly logically. He is unable to determine, true or false statement "A", if it goes beyond the previously established axioms. It happens inevitably. Human, facing the same statement "A", logically unprovable or irrefutable, will always determine whether it is true or false, based on your daily experience. And this is the superiority of man over the machine., connected by ideal logical circuits. We are able to understand the truth of Godel's theorem, the computer is not. And because our brain is anything, not just a car».
"Remember, cattle, that you are people!»
Benjamin Libet's famous experiment, first held in 1973 year, showed, that the human brain generates signals for solving a problem long before, How does he realize his essence?. That is, it reacts, the machine. But the decision itself is made only after, How does our mind make a choice?, that is, we act like people.
So that we can physically interact with the world around us, the brain must very quickly independently distinguish the body's own actions from external influences on it. He does it, creating special (efferent) copies of each command, that sends to the muscles.
We shake our heads left and right, but at the same time we realize, what is swinging is not the world around us. This is because, that the efferent copy makes it possible for consciousness to determine, that our brain itself gave commands to these movements, although the signals, entering it through the optic canal, create a different picture. That is, inside ourselves we tirelessly compare the ideal algorithms of the machine of our body with the awareness of reality by our consciousness..
The mind in the head unites the conscious and the unconscious. Our thoughts and actions are the result of both automatic, unconscious processes, and conscious choice of mind.
But humanity does not give up and the computers we have created are becoming more and more complex.. We are trying to refute the theorems of the brilliant Kurt. Bring matter closer to spirit. Become creators, bringing them together. "We are speeding up the process! Give cyborgs!», - demanded by the masses.
For starters, Neuralink Mask is launching a great project to implant chips in the brains of pigs. People are getting a little uncomfortable.: "Look, it's like a Fitbit in your skull with little wires. Healthy and happy pig with implant, which has been in her head for two months, and everything works fine!», Elon says smiling sweetly.
Then at the World Economic Forum in Davos in January 2023 Duke University professor of legal ethics and futurist Nita Farahani delivers a speech “Are You Ready for Brain Transparency”?» Everyone is happy and applauding in unison.
In May, the US Food and Drug Administration (FDA) authorizes Neuralink to conduct human clinical trials. Really?
Why is everyone freaking out? What's wrong? This is for medicine and healthy lifestyle. Right here, near west of Belgorod, for thirty-odd years, they reprogrammed the brains of the population of an entire country and without microprocessors in their heads. yes so good quality, that despite the acquisition of certain death in the Donetsk expanses, in all video messages to his chief hetman, Ukrainians whine and complain only about the lack of opportunities to die trained and well-equipped, according to NATO standards. That's if we are shod, will feed, fill their pockets with cartridges and grenades, then they are ready to burn alive inside the German panzers for the ideals of overseas friends! After excellent results on animals, it is possible to chip without clinical trials without exception of all ukrvoyak. underworked. White masters were pressed for time, so they drove the herds of aborigines to slaughter without "iron" in the head.
I will be objected, that the actions of each individual fighter of the tactical level are honed to instincts, bringing skills to automatism. May be. But still for military affairs, and especially the operational-tactical and strategic level, conscious choice and decision is crucial.
Explain by comparison: a pilot in the cockpit of a combat fighter and an operator with a UAV control joystick in his hand. they, who was at the helm, feel the difference immediately, despite the efforts and the pilot, and the operator are transmitted to the controls remotely through a system of communication channels, amplifiers and drives. Moving in space inside the aircraft and remote control of the drone from a chair in front of the screen are very different.
The feeling of using weapons is also radically different..
And if tomorrow there is no pilot in the cockpit, and there will be no operator in the chair in front of the screen? Is there a possibility of a semblance of consciousness appearing with the help of a neural network to control weapons without a person? And the right to decide to open fire to kill? Is it possible to create a system of fair self-selection for machines to destroy a target? Is it even possible to entrust this to a machine and give AI the right to kill people, who created it? After all, we, people, other. We are absolutely not similar to the computer we created, but all the time we try, if not to make it more perfect than our loved ones, then at least level with yourself. What for?
But is it moral to ask these questions, when those who systematically appear among us, who consciously decides to act in the name of absolute evil? Instances like Hitler and his pack, who created the Nazi state of "racial superiority", the victims of which were millions of industrially killed in concentration camps and destroyed by punitive detachments.
Can this be justified, that then there were no such supercomputing capabilities?
And the lack of power of Ukraine, understanding the complete futility of further existence and continuing to send people to certain death for the sake of personal enrichment and the fulfillment of other people's goals? crests have computers of the wrong model? Or maybe, curators from across the ocean have data centers loaded with analysis of other content? The answer suggests itself.
On the other side of the ocean are well aware, what are they doing. And such opportunities will be provided to AI. Work is in full swing. And our delay in this direction is just as dangerous, as in the case of the atomic bomb of the Manhattan project.
the, that the former “partners” who have no doubts about anything will try to arrange a massacre worse than Hiroshima and Nagasaki for the population of Russia, by brewing biological weapons in one broth along with a massive strike with conventional weapons, spiced on top with AI-controlled satellite weapons, no doubt. It's right there in their national security concept papers..
The process launched
After analyzing various forecasts of such think tank, as RAND Corporation, it becomes clear: major changes in the nature of future wars, forms and methods of using the armed forces will be based on the achievements of science and industry.
Breakthrough technologies in armaments and military equipment are considered in a single alliance with new principles and methods of command and control.
New Strength and Power Requires New Global Control Capabilities, and AI as a tool for amplification, an exoskeleton for the brain is just very handy.
Big Data и machine learning, in which computer vision is of particular interest to the military, symbolic computing and logic programming, work with natural languages and modeling of thought processes, nonlinear control in nonlinear systems, what is our human society, - these are just some parts of the giant array of new science under the new global brand "AI".
The previously mentioned “National Strategy for the Development of Artificial Intelligence for the period up to 2030 years" defines AI as: «Artificial intelligence - a set of technological solutions, allowing to imitate human cognitive functions (including self-learning and search for solutions without a predetermined algorithm) and get results when performing specific tasks, comparable at least with the results of human intellectual activity».
So the “Artificial Intelligence Strategy” as part of the American “National Defense Strategy” states, that AI will change society itself, and the nature of future wars.
For the US Congress in November 2020 year, a report was prepared on the topic "Artificial Intelligence and National Security" in which it is directly written, what "the main reason for the creation of various military systems, with AI, is the need for rapid processing of structured and unstructured data of huge volumes (so-called big data), due to the constant expansion of the number, nomenclature and technical capabilities of modern means of obtaining information.
This information can be presented in various forms., including photo- , video- and radar images, as well as audio- and text messages in different languages, data, received from cyberspace, and so on. d.».
The document indicates, that AI will spread in all types of combat and support activities of the US armed forces in the near future.
Words don't match deeds, and the UK Department of Defense conducts an exchange of advanced technologies within the framework of the AUKUS block under the Pillar program 2 (cooperation in high-tech areas, including cyber capabilities, Artificial Intelligence, quantum computing, hypersound, as well as systems to counter hypersonic weapons).
The event was initiated by the British Defense Science and Technology Laboratory (DSTL), which held "live retraining of drone models during the flight and established the exchange of AI between AUKUS participants». Her teams then developed AI models and controlled each other's air and ground systems., tasked with identifying targets.
What Australian Intelligent Drone Swarms Are Ready For? It is believed, that they could be the deciding factor in the Taiwanese scenario.
Back in May 2022 a report on UAV swarms was published, communicating with each other using a distributed "mesh" data exchange network using lasers within line of sight. The instantaneous exchange of information between individual drones made the group virtually completely autonomous.. Thinking swarm can work in conjunction with manned stealth aircraft, increasing the range of their detection and control systems in complete electronic silence. With their large number, they are able to load enemy radar detection channels with decoys., forcing useless expenditure of missiles and shells, while manned aircraft destroy exposed targets.
Machine learning and artificial intelligence will also allow drone swarms to study targets from different angles., cross-check various target designation data streams and offer the most optimal attack option, which will allow you to strike at the most vulnerable points on warships (rocket launchers, conning tower, radar antenna arrays).
Pentagon, inspired by the progress of autonomization technologies, proved their effectiveness in the military conflict in Ukraine, launched its own program "Autonomous Multi-Domain Adaptive Swarms-of-Swarms"(AMASS) to create swarms of drones, which can be launched from the sea, air and land.
The AMASS concept is aimed at developing control systems for autonomous drones, working together to destroy enemy air defenses, artillery, missile launchers and command centers, and, likely, the result will be aimed at preventing a Chinese invasion of Taiwan.
article, published in November 2022 of the year for the Hudson Institute, Japanese military expert Koichiro Takagi writes, that AI can outperform humans in decision-making speed, while referring to the risk of a blitzkrieg, during which opposing AI systems will launch an uncontrollable chain reaction, which will lead to the start of a conflict or even to the launch of nuclear missiles.
He noted the tendency of people to trust artificial intelligence in the "heat of battle", and even if there is evidence that, that the AI decisions are wrong. Takagi points out, that throughout history, wars have not been won by superior technology and science, and thanks to human intelligence, which uses all these tools. «Future military conflicts will largely depend not on whether, who has the best artificial intelligence algorithms, but from innovative concepts, using it along with human intelligence and creativity», - this is his opinion.
A vivid illustration of the possible problems was the recent incident with the attack, described by Colonel Tucker "Cinco" Hamilton in a simulated test, in which the drone, using artificial intelligence, it was recommended to destroy enemy air defense systems, and he attacked anyone, who interfered with this order.
But China is not far behind.. "China AI Development Plan", which was accepted into 2017 year, defines, what to 2030 China is set to become a world leader in AI technology. The application of artificial intelligence is determined, as a prerequisite in the field of creation and management of military equipment, weapons and military operations. Plans don't stay on paper.
Wuhan University this year conducted an experiment involving AI and an artificial earth satellite (IZZ) for Earth remote sensing Qimingxing 1. The scientific team decided to optimize the operation of the satellite, as the analysis showed, that China's constellation of 260 devices for the study of the earth's surface is not working at full capacity and is wasting its potential to obtain data, having no practical application.
Satellites are expensive, their service life and resource is strictly limited, and therefore it is necessary to dispose of the heavenly herd in a businesslike way.
Who is our best and zealous owner? And let me order II, to maximize satellite capabilities with new orbital applications?
To ensure that the data received is most useful to customers, scientists created a trainable model similar to a large language ChatGPT and introduced the parameter of freedom of choice to study ground objects into the project.
The program used the algorithm, which correlated the relationship of events not only based on training on databases, but also on the example of direct observation of nature and human activity. Then the control of the satellite for one day was transferred to the control of the machine, which began to independently choose targets for observation. AND… surprise!
First of all, the neural network directed the cameras of the spacecraft to the north-east of India - to the city of Patna on the banks of the Ganges., where the Bihar Regiment is stationed. AT 2020 In 2018, this military unit was involved in a clash with PLA units in a disputed border area.
After that, priority was given to the observation of the Japanese port of Osaka., Where are US Navy ships based?.
So talk after that about impartiality.
Probably, the, who created the Chinese analogue of ChatGPT before connecting to satellite control, were by no means indifferent to China's national security issues, if the device was immediately sent by the system to check the situation in the area of its responsibility in the places of the past and, perhaps, upcoming military operations.
What is the scale of promising space battles? According to UN official data, is in orbit 10 911 artificial earth satellites, and the number of objects in orbit will only grow. Of this amount, 4 408 on 72 orbits are owned by Starlink, but the US National Space Intelligence Agency (National Reconnaissance Office, NRO) continues to work out the possibility of additional equipping of communication satellites of this type with systems of high-altitude ultra-clear sounding of the Earth's surface, simultaneously solving the problem of transmitter power.
Need an extra thousand heavenly eyes and ears, requiring simultaneous and coordinated management. That's not counting space junk., collisions to be avoided, calculating hundreds of thousands of trajectories of fragments and fragments. No TsUPs will cope. Only specially created self-learning algorithms.
And again the problem: choice, what to hear and see, the system will carry out? They want to solve the problem with the help of an operator, who seems to be watching, how AI will perform the tasks.
And confidence, that he will show everything to a person, absolute? And the human consciousness will always understand the model and logic of the behavior of a self-developing structure? But these questions are secondary.. Need to be first. “To move forward, you need to run twice as fast”. To become and remain the first, you need ultra-high speeds and unprecedented accuracy in processing a huge amount of data.. And the race is on.
1983 year - an indicator in 1 gigaflops achieved by NEC SX-2 supercomputer. 1996 year - milestone 1 trillion flops per second, i.e 1 Tflops passed by ASCI Red. 2008 year - plank in 1 quadrillion flops, i.e 1 Petaflops, overcome by the IBM Roadrunner Frontie. This supercomputer has achieved speeds of 1 exaflops (1 quintillion operations per second). Frontier Hewlett Packard Enterprise CRAY EX235A Supercomputer, at 2022 year passed the milestone in 1 exaflops (1 quintillion operations per second). In the work with AI in the first place, almost reaching the speed of 7 ExaFLOPS. Installed at Oak Ridge National Laboratory, USA. OceanLight and Tianhe-3 computers put into operation in China, gaining speed in the LinPack test 1.05 and 1.3 ExaFLOPS.
Are there Russian cars in the tops? Yes, in 25th place since November 2022 year in the ranking appeared the car "Yandex" - "Chervonenkis". Mainly used for working with neural networks. Inside our machine, all the filling is imported - NVIDIA A100 GPUs along with AMD EPYC 7702.
Two more Russian cars from Yandex are among the world leaders: 16-petaflop "Galushkin" (44 a place) and 12.8 petaflop Lyapunov (47 a place). AT 2019 and 2021 year, two products appeared at Sberbank of Russia: Christopher (87 a place) and Christopher Neo (50 a place).
World leaders and in terms of power, and by the number of machines are China (about 170 machines) and the US (nearly 150), behind — Japan, Germany, France and the Netherlands.
Part of the systems in the rating is not taken into account, as they are used by countries in the interests of national security. In Russia, such a machine is in the National Defense Control Center. known, what performance is 16 Pflop/s. and that means, the machine of the Ministry of Defense of the Russian Federation is one of the fifty most powerful in the world.
Americans in the military-computer field are not far behind. According to Datacenter Dynamics, US Department of Defense Supercomputing Resource Center, in April of this year, after completion of testing, received a new HPC - a system called Nautilus: located at the John Stennis Space Center. Besides, The Navy DSRC has upgraded another of its supercomputers, the Narwhal system.. Updated monster operates 308 480 cores with peak performance in 13,6 Pflops.
And that's just part of the unveiled High Performance Computing Modernization Program. (HPCMP) US Department of Defense for Climate Simulation, performing research in the field of hydrodynamics, chemistry and other branches of knowledge extremely useful for mankind.
What did you think about the war?
The modern computer was created, as a tool for automating settlement work. Today, the car is used to "revive" various models., as existing physically, as well as virtual. At the same time, they try to minimize the direct use of.
Software development is based on mathematical principles, what creates problems for problem solving, not related to pure mathematics. Everyone is waiting for revolutionary changes in the existing paradigm and the removal of artificial restrictions, as it were. (address space, command execution sequence, bit depth and more).
But it has not yet been possible to develop intelligence in silico. A quantum computer is being tried for this role.. Yes, it is designed to solve a specific range of tasks, different from those, that solve standard machines, and until recently, external laser systems were used to project photons, which were too bulky for use outside the laboratory.
The problem is trying to eliminate due to light sources, integrated directly into the microcircuit. This solution not only takes up a minimum of space, but also provides better stability.
example: Dutch company QuiX Quantum is still in 2021 sold her 8- and 64-qubit processors to the German Air and Space Center. Last year, a joint team of specialists from Leibniz University Hannover and QuiX Quantum created a prototype device with a fully integrated light source., which independently generates entangled photons for quantum computing. stated, that the new technology has reduced the size of the source by more than 1 000 times with the necessary stability for scaling up and mass production. For this, a hybrid technology was used., which combined indium phosphide laser on one platform, silicon nitride filter and resonator.
When solving the problem of microminiaturization, the problems of supersensitivity of qubits to noise were also overcome., what can be considered (if it's really true, not an advertisement) a major step towards the quantum advantage of the photon chip.
Unlike Google research, which currently uses supercold qubits in cryogenic systems, quantum advantage with microplatform photonic systems can be achieved at room temperature.
What are the prospects for such a system, assuming, that achieved the desired result? This solves problems with power consumption, and sizes, and speed, and complexity of possible calculations. In the same amount of drone body space, it will be possible to place a computer that is hundreds of thousands of times more efficient and simultaneously performs a greater number of different tasks.. There will be a new generation of a community of people who can reason intelligently. Man made. fast, autonomous, deadly, insensitive.
For the war of people space, air, land and water require new solutions. And the former "partners" left us no choice.