Matryoshka evil

How digital decentralization threatens the monopoly of states and what does ChatGPT have to do with it

Part 1: Bowed ear

Donald Trump liked to repeat the winged phrase of the outstanding boxer-apparatus Mike Tyson: “Everyone has some kind of plan, but only until he puts in turnips” (EVERYBODY HAS A PLAN UNTIL THEY PUNCHED in the MUTH).

The Tyson meme, if we draw a historical parallel, is a modern metaphor for a sword, which Brenn laid the leader of the Celtic gang in 388 BC. e. On the scales when discussing the indemnity imposed on the defeated Rome. The Romans were then indignant so brazenly weighted by the barbarians and received in response the winged “Vae Victis!”, Danced, defeated.

However, there is a difference between a mortal and Tyson, associated with a flickering eras. Brenn first crushed Rome in a bloody battle, and then impudent with weights and swords, but Tyson intimidated Evander Holifield before the fight. The fights, as we recall, came out two: the first ended with a knockout of Tyson, and in the second Tyson bit off a piece of Holifield's ear, after which the puppeter was disqualified.

Everything that is written above is nothing more than the sublimation of modern fears of mankind. Making fun of the Tysonov, obviously not reaching Brennov, we reassure ourselves that the mountains only generate the mouse. Unfortunately, this is not so.

We were lucky (or “lucky” - it depends on the optics and life circumstances of the reader) to live in the era of the crushing breakdown of the history of mankind. By a break, I understand such a point of no return, which will lead either to absolute and universal slavery, or to a hitherto unheard of new form of the organization of society. It is still impossible to predict any of the outcomes, although it is obvious that it will not be possible to return to the usual status quuo of the past.

Unfortunately, we, representatives of Western European civilization, cannot feel this coming break in all its inevitability, since the main forms of its manifestation are now obscured by the unbearable tragedy of war in Ukraine. This war is so monstrous, impossible and absurd that it covers all the horizons and turns into sacrilege any attempt to look on the other side of its horrors.

Psychologically, this position is more than justified: the reality inside of which the person is located always obscures everything that is located somewhere else and does not directly concern us. We see the horrors and suffering of this war, and it seems to us that "the attack of Russia on Ukraine violated the peaceful status quo, which we have enjoyed for more than 70 years."

I will give only two lists from memory.

After 1945, the USSR / Russia: provoked a Korean war, tanks set up an uprising in Hungary, joined the proxy-war with the United States in Vietnam, destroyed the Prague Spring, organized a proxy (through Cuba) in Angola, unleashed a war in Afghanistan, provoked a series of internal civil wars in Georgia, and Transnistrian, Transnistrian Tajikistan, Abkhazia and Dagestan, held two wars in Chechnya, attacked Georgia, annexed Crimea, conducted a proxy-warrior in the Donbass, conducted a military operation in Syria, participated in a civil war in the TsA, attacked Ukraine.

During the same period, the United States was held in Korea, organized a coup in Guatemala, intervened in Lebanon, took part in the Civil War in Laos, marked in Cuba in the “Operation in the Gulf of Pigs”. Further - a tongue twister: Vietnam, the Dominican Republic, Cambodia, Salvador, Iran, Lebanon, Grenada, again Libya, Panama, the War in the Persian Gulf, Bosnia and Herzegovina, Haiti, Yugoslavia, Albania, again Ukraine.

If now it seems to someone that it was in 2022 that 77 years of peaceful status quo were violated, then you can not read further.

I will express a seditious thought that no peaceful status quo never existed. Neither in Europe nor beyond. There was no peaceful status quo in the last 77 years or in the last 50 centuries of history recorded in the memory of mankind.

War is a disgusting Modus Vivendi Primat Homo sapiens. The maximum that one could claim was a kind of “softening of morals and decreasing bloodthirstiness”, the evidence of which Stephen Pinker devoted 800 pages of his monograph (The Better Angels of Oour Nature: Who Violence Has Decline). However, reading documented reports on war crimes in Ukraine, you begin to much doubt even in these modest conclusions of Pinker.

The above lists of military operations (beyond the framework of the enumeration, there are genocide in Indonesia, the Red Khmers, and the hecatubes of the cannibalnious regimes in Africa) - this is by no means a way to take the conversation from the tragedy of war in Ukraine, and an attempt to consider the vector leading to the shocks, against which everything previously experienced by humanity will fade.

What can we consider today on the other side of the Ukrainian tragedy?

If we rise to the level above, we will open a picture: on September 11, 2001, a strange process was launched in the world, which at first charged a couple of local wars (Iraq, Afghanistan), then torn into the Patriot's Act - a law that rejected the cornerstones of the US Constitution, and in the end caused the tsunami of the legislative epigonism. Analogs of Patriot's Act were adopted in more than 100 countries (Russia, Germany, Canada, Australia, Thailand, Ethiopia, Vanuatu, etc.)

In all these laws, the definitions of “ terrorism ” and “ terrorist ” are formulated, to which state authorities declared a merciless war. The bloodthirstiness of each war is determined by the measure of the damage of a particular state. Some are fined, others torture, others give prison terms, the fourth kill - but the essence is one and unchanged:

The emphasis of state violence was shifted from the accusation of reprehensible actions against accusations of reprehensible thoughts.

If in the 70s. XX century. In order to become an “enemy of society and the state”, it was necessary to be in the “red brigades”, arrange explosions in stadiums and kidnap politicians, then in the era after 9/11 enough for the state to declare you a “terrorist”. At the same time, nothing to blow up and you need to abduct anyone, it is quite wrong to think . To think, all the more, speak like a native state does not approve. Such disapproval automatically turns you into a “terrorist” with defeat in all rights.

At the same time, each “native state” has its own set of signs that fall under the definition of “terrorist” and “terrorism”: in Iran - one, in India - the other, in China - the third, in the Russian Federation - the fourth, in the USA - in the EU - the sixth.

Of course, it will be immoral under the pretext of a genetic community (9/11) to row all the local bugs of “terrorism” under one comb. It is wild to compare the real prison term, which is given in the Russian Federation for “terrorism”, expressed in the refusal to maintain a criminal war, with the practice of social “Kanseling”, which “terrorists” are faced with today, which are disapproving about the BLM or Polygender Revolution movement.

Nevertheless, without understanding the semantic community of processes that determine the universal vector of the development of our world, we will not be able to understand the understanding of the essence of the historical breakup , which these processes are inexorably approaching.

The intermediate result is that we are dealing with the virtualization of state violence against citizens.

There is no need to really blow up and perform acts of terrorism, enough for the state to call you a “terrorist”.

The archaicity of such a transformation cannot escape attention. At the dawn of mankind, the mythologism of thinking was just manifested in the idea of ​​what to call - already meant to understand and win . If you know the name of God, you already understand God. If you know the name of the enemy, you are already winning the enemy.

George Orwell was the first in modern culture to indicate a mortal danger that state games carry with words.

Jeux de Motsistine work miracles: you do not need to make excuses and explain to the society your own unsightly qualities, just call black white, slavery - freedom, war - the world (or the desire for peace), and then your unsightly qualities are easily transferred to the enemy. No need to change anything on the merits, just put a shortcut on the opponent’s forehead and immediately move on to “justified” actions. You are a fascist and a terrorist, so I'm starting to wet you. How am I a fascist?! And so - and the next moment you are already biting your ear.

But games with words are not the culmination of processes that predetermine the inevitability of a historical breakup, as it might seem. In a certain sense, this is so, but there is a feeling that the gum of “terrorism” is still a particular manifestation of something more general. What?

In the course of the case, it is important to note that the new world order as an original and independent phenomenon exists only in inflamed brains that have moved on Internet conspiracyology. It is only about the evolution of the idea of ​​power control , which is realized in human society from the first days of its existence.

Historically, the highest form of power control is a state that began as imposing the will of the most powerful family clan (monarchy) to the society, and ended with voluntary delegation of power to elites in the broadest sense of the word (democracy).

In any situation in the context of the state, we can only talk about the distribution of powers between elites. A full equality in which each individual was endowed with equal power and property rights, never existed in the real history of mankind (but with excess was compensated in the fantasy heritage of philosophers and writers).

The process that guarantees us the historical breakdown of an unprecedented scale is called the change in the paradigm of power control in society . This process lies in the fact that the state as an institution is rapidly losing today not only moral rights (this has happened for a long time), but also the very technical opportunity to exercise effective control over society as a whole and each individual in particular .

It is this catastrophe of the state that is the root cause of the whole series of planetary upheavals, which we have observed for the past decades. Namely: incomprehensible games with words, incomprehensible games with genetic research, incomprehensible games with artificial intelligence, incomprehensible games with financial relations.

A single vector of events, which determines the loss of power control by the state, is called one word - decentralization . Moreover, this is not about decentralization as an abstract idea (in this form it existed in the dreams of Greek philosophers two thousand years ago), but about decentralization as a set of objectively existing technologies that make the triumph inevitable.

I risk assuming that the whole set of tragic processes observed by us in the XXI century. - cynical substitutions of meanings and concepts, barbaric wars, widespread attacks on fundamental civil rights and freedoms of society, financial experiments with elastic emission - all this to one degree or another, conscious or subconscious reactions of a state that experiences fear from the inevitably impending collapse of its monopoly to power and control over society.

If the threat of the powerful monopoly of the state proceeded from the subject of any nature - whether it be an individual or even an entire army, the state would simply crush the confusion as a mosquito following the example of its entire past history.

The tragedy (for the state) is that the threat of the power monopoly proceeds from the fact that it is impossible to allocate to the subject or somehow localize it, from technology . Moreover, we are talking about technology in the most abstract and convenient form for mass replication - computer code .

This technology is developing irresistibly faster than any pandemic. There are only two ways to delay the spread of technology hostile to the state: at the human psychology level or at the level of technology itself .

In the first case, the state makes titanic efforts aimed at the most wide discredit in the company of any technologies that implement the idea of ​​decentralizing social relations . In the second case, the state is trying to contrast with hostile decentralization technology the technology of holding society under its control .

Decentralization technologies are represented today by an unprecedentedly wide spectrum: from ensuring the autonomous social existence of the individual to the system of financial relations derived from control of both the state and the private structures regulated by him (banks, exchanges, share funds, etc.)

I will give only two examples. The most striking technology for decentralizing the social aspects of being is embodied in the Network State concept proposed by Balaji S. Srinivasan. Financial independence from the state is ensured by the rapidly developing ecosystem of Defi (decentralized finance).

I really hope that I will have the opportunity to tell about these dizzying technologies, but I would like to start a conversation with the second tool with which the state is trying to solve the problem of power control.

We are talking about the development of technologies aimed at strengthening centralization and thereby slowing down centrifugal processes in society.

It should also begin with technology for centralization because they are associated today with the hottest plot, exciting the media of the planet, with AI, artificial intelligence.

Part 2: Passion AI

The reason for the conversation about artificial intelligence (Artificial Intelligence) in the context of the state’s struggle for power control over society was an open letter published by the organization Future of Life Institute March 22, 2023. The letter entitled “Paise with artificial intellect Giant Ai Experents: An Open Letter).

In just one week, the letter collected more than 50 thousand signatures: from the most eminent Chinese and American entrepreneurs and university luminaries to Ilon Mask and Steve Vozniak.

The world's resonance from the letter was released and even called a couple of very alarming signals: the Lobbitist Center for AI and digital policy (Center for Artificial Intelligence and Digital Policy) directly turned to the US Federal Trade Commission (US Federal Trading ) Openai release of new commercial releases of its popular GPT-4 engine.

Literally on the same day, UNESCO addressed the governments of all countries calling for the initiative as soon as possible with the temporary stop of research work on the development of powerful artificial intelligence systems.

The letter made a world commotion all the more unexpected , since the Future of Life Institute is a marginal organization that had not previously been particularly popular.

Future of Life Institute, FLI (Institute of Life of the Future) was established 9 years ago by Jaan Tallinn, one of the creators of Skype , Max Tegmark, professor of the Massachusetts Institute of Technology ( MIT ), and Victoria Krakowna (Viktoriya Krakovna), a scientific researcher of AI from the British flagship of the DEEPMIND industry.

FLI is a non-profit organization that consistently and uncompromisingly promotes one single (but very attractive!) Idea- living in the world . The concept of the world for FLI is synonymous to the moral imperative and is absolute in the spirit of fantasies of theoretical Christianity about "non -resistance to evil by violence."

FLI is the main threat to peaceful living on the most advanced borders of modern science - in artificial intelligence and biotechnologies, as well as in the development of nuclear weapons and the fight against worldwathing.

FLI has previously published five open letters:

• Priorities in research for the development of strong and beneficial AI (Research Priorities for Robust and Beneficial Artificial Intelligence (2015);

• AI and researchers of autonomous robotic weapons (Autonomous Weapons Open Letter: AI & Robotics Researchers) (2016);

• Obligations of the developers of autonomous deadly weapons (lethal autonomous weapons pledge) (2018);

• An open letter of the world medical community to the developers of autonomous deadly weapons (Autonomous Weapons Open Letter: Global Health Community (2019);

• Landmarks in the regulation of artificial intelligence (Foresight in AI Regulation Open Letter) (2020).

The highest achievement of FLI on AI issues is 34 thousand signatures collected under the letter of 2016. The remaining appeals were gaining from 256 to 11251 signatures.

As you can see, the voiced figures are the parameters of absolutely marginal public activity. And suddenly such an incredible success: 50 thousand signatures in one week, Elon Musk and resonance at the UNESCO level. What has changed so in the calls of FLI for the world community?

In its past appeals, FLI expressed anxiety by the phenomenon of artificial intelligence as such and the potential possibilities of its use in the military industry and medicine. And this is understandable: no current models of practical significant AI implementation existed either in 2015 or in 2017.

The theoretical documentation according to its first prior-trained transformer (Generate Pre-Trained Transformer, GPT) Openai published only on June 11, 2018 (research “Improving Language Understanding by Generative Preining”).

The more or less meaningful GPT-2 saw the light of February 14, 2019; The advanced GPT-3 appeared on June 11, 2020; However, even this language model posed a threat to humanity (in the minds of FLI and supporters of this organization) only hypothetically.

Everything changed on March 14, 2023, when the GPT-4 brought AI to a completely different qualitative level. It is significant that the broad world community was not reported about real changes in the new generation of the AI ​​language model. The mainly ability of the model to analyze on the input nodes of its neural network not only the text, but also the image was announced. Most of the process is hidden under the term Not Public.

There is no doubt that the lack of information on the GPT-4 in the official report “GPT-4 Technical Report” is a purely formal obstacle, and all interested structures and research centers with a detailed list of updates of the new model got acquainted.

Be that as it may, but a week after the publication of the GPT-4 FLI documentation, she broke out with her Alarmist open letter and caused a serious commotion.

It can be assumed that the cheese-brow caused not another call to the AI ​​transformation from an inorganized intelligence tool (undirected intollingence) to the Beneficial Intelligence tool-a key pathos of all past faces of the pacifist organization. The open letter of FLI dated March 22, 2023 calls for voluntary freezing for six months of research in a completely different field, namely AI systems with a power exceeding the GPT-4 power.

It is important to understand that this is not about ordinary artificial intelligence, but about the so -called. Artificial General Intelligence (AGI) - general intelligence systems (often translated as “strong artificial intelligence.” - Ed. ) , Which are able to act as full -fledged competitors of the highest forms of the human mind. It is AGI, according to FLI, that can cause a global change in life on planet Earth, and it is precisely such systems that should be temporarily limited in development in order to form an adequate ethical program and really valid controls from Homo sapiens.

AGI - this is the object that the FLI and the Sub -supporting (supporting his call, the thinkers and entrepreneurs, requires. As for the rest (read: domestic) AI, in it FLI no longer sees any threats to the peaceful existence of the human race, let it develop for health.

Anticipating the fears of corporations involved in the development of AGI about the fact that the 6-month moratorium will strengthen the competition, FLI hastened to reassure the interested side: to be afraid that competitors would not be able to miss the lost in six months. The computer capacities necessary for training AI at a level “exceeding GPT-4” are available only to a very limited number of companies (read: in one form or another related to the state, and in the case of China-so 100%), and the separation of the rest of the players in the market is so great that there are no chances to reduce the lag.

It is funny that from the wide masses of enthusiasts, day and night studying the capabilities of the GPT-4, the release of the last language model to a new qualitative level slipped. The wide circle of my friends unanimously claims that the GPT-4 gives practical recommendations-first of all, in the field of programming-with reduced quality compared to the previous GPT-3.5 model.

I believe that such a “lag” at the level of practical application is due only to the freshness of the model, which is not yet a month from the genus. The quality of interaction with the private texture is a profitable thing, since the training of the neural network is limited in speed only by the power of computer equipment.

There is no doubt that in a month or two commercial engines using the GPT-4 will issue answers significantly exceeding the capabilities of all previously operating models.

Obviously, FLI, as well as the whole cohort of the scientist Luda, who supported the call to the moratorium of the development of AGI, knows much more about the GPT-4 than the inhabitants. Nevertheless, some leaks in the public information field have already happened. In particular, everyone has a sensational case, when the GPT-4, faced with the need to pass the check on the site using Captcha, turned to Taskrabit for help, the network outsourcing system, and, having pretended to be a blind person, asked for a humanoid assistant to make Captcha instead. Which he performed by providing GPT-4 access to the necessary information.

Now, we will try to understand how the state considers artificial intelligence as a key technology for confrontation with the impulses of decentralization, which undermine it, states, a power monopoly.

In November 2022, Openai introduced the public Chatbot ChatGPT, which, in fact, aroused mass interest and spread of artificial intelligence technology in the world. Today, the vast majority of ChatGPT private clones are implemented on the GPT-3.5 language model, although there are already a number of GPT-4 implementations. In any case, we are talking about the AI ​​level, preceding what is understood by agi, systems of general intelligence.

Everyone who has ever worked at least once with ChatGPT could not help but notice that the level of “intelligence” of this chatbot directly depends on two components: the amount of information that was originally sent to the input nodes of the neural network, and then conducted its training, and on the intensity of the user's interaction with AI toy.

The parameter of the volume of information in the household chatbot is limited by the developer and it is modified (in other words, the Openai company decides what and to what volume to teach the neural network of the language model). Therefore, whenever Chatgpt is faced with a question that he does not know the answer, Chatbot begins to have fun and at ease fantasize.

The most ridiculous example of such a fantasy is a story about how Chatgpt plays chess. YouTube is full of videos , in which “artificial intelligence”, getting into difficulty every time, begins to invent non -existent moves: who did not look, I strongly recommend watching - pleasure is guaranteed.

The lack of information on the input nodes of the ChatGPT neural network compensates for the training module based on the analysis of previous user questions for the sake of adaptation to its preferences in the future. Similar tactics are used by telephone scammers who call pensioners with stories about how their son/grandson went to the hospital/accident, so it is necessary to urgently pay a small amount for the operation/subsidence of the traffic cop. As the conversation develops from the grandmother in the pre -infarction state, all the necessary additional information (names, dates, age, preferences, the size of the pension, placement of furniture in the room, the place of storage of the wallet, etc.), which is then ease, is then used by a fraudster in his own replicas,

The answers of the AI ​​engine each time more and more intensively like users who eventually sit on this needle and it seems to them that they even love their virtual interlocutor. By the way, FLI also indicates the danger of such doping in the comments on its open letter dated March 30, 2023 (FAQS ABOUT FLI's Open Letter Calling for Ai Experiment), citing, in particular, to the recorded case of the first suicide, which happened recently after communicating with one of the clones ChatGPT.

In any level of AI level GPT-3.5 / GPT-4-nothing more than a useful and entertaining toy. Something like Siri (or Alice) on steroids. This toy undoubtedly guarantees interest from users and private business, but not a state.

An adult conversation begins where research is being conducted on creating general intelligence systems that FLI and the thinkers and businessmen supported them seek to prohibit the unpredictability of the situation, when artificial intelligence begins to compete with a person in any field of knowledge.

The fact is that

Monopoly control over the system of artificial intelligence with unlimited abilities opens up unlimited possibilities for control over the entire society,

immersing in the virtual environment. AGI in the era of meta -makers ( virtual spaces in which people can interact with each other according to their own rules. - Approx. Ed. ) - this is the key to modeling any reality in the form in which the controller needs this reality.

If the controller (read - the state) will need to create an extensive and attractive concept of a centralized paradise for society, the only one capable of “protecting” the digital habitat of citizens from chaos and entropy of decentralized temptations (for example, the same Network stat, network state, or the placement of money on the credit platforms of Defi, and not in a commercial Bank), then only AGI, a system of general intelligence, will impose such a concept into an infinite number of detailed meta -footed ones.

Because only AGI can not just stupidly announce the “correct” hierarchy of values, but to make a role model of behavior truly viral. How? Having built this hierarchy and model into the very fabric of the digital reality of meta -makers!

I think it’s easy to guess where public concern appeared in all this, aimed at the immediate ban on the development of Agi. Is there at least one state in the world that will voluntarily reconcile with the loss of monopoly on the most effective technology of power control over society that modern science only gave rise to, over Agi? The question, it seems to me, is rhetorical.

I’m afraid (or rather, I am happy) that the efforts of FLI will be in vain and the state ban on demonopolization of AGI will break off. For a simple reason: all without a single exception, previous attempts to prohibit decentralization, in whatever the digital space this decentralization is realized, always broke off with an embarrassment crack.

Why so? Because now there is no localized subjects to rot, but technologies that, as you know, have neither physical form nor locus. The digital world is that he is such a large prankster.