Fragments: The Teacher's Doctrine

A collection of learning intentions dealing with the four principles of learning - and learning versions of various branches of philosophy - in the tradition of fragmentary philosophical writing

Likutei Moharan - from the Netanyahite Teacher  (Source)

Learning is above language. Learning is to thinking as thinking is to language. The brain is not a communication network, nor an action network (thought is the action of the brain), but a learning network. The development of any learning system is from information to knowledge to knowing. From genes - encoding (language), to gene networks and epigenome - control computer (thought), to genetic learning systems. For evolution is not a trivial learning algorithm, random search, but a learning algorithm of learning algorithms. And learning is always an algorithm of algorithms, second-order, and creativity is an algorithm of learnings, third-order.

In the future, the only thing that will be worth money is learning. Information will be free. The most sought-after profession will be teacher, because everyone will be students. So what will have value? What, for example, will attract the opposite sex? The thing that will still have value in relationships is learning - communication will be trivial. Time and space will not be defined by how long it takes to transfer information (hence the equivalence of space-time distance), but by how long it takes to learn (the time to learn, the distance to travel in learning). And the value of the woman will be in her learning, the learning effort to move from external learning to internal learning, which is the movement of coupling. In the future, a shared language will be a shared learning algorithm, allowing the sharing of learning intentions, not shared information. Both sides will transmit directions themselves, instead of information passing in directions between them.

Information is the quantification of language. The quantification of learning will be how much learning distance we've covered, for example how many basic learning stages were on the way, and how much a teacher had to teach, and the time of learning will be how many repetitions are needed. The speed will be the slope of the learning graph. And when there is an abyss in time and space, an infinite learning distance, then there is creativity. And there too the question is how many creative leaps, unlearnable, were needed. How many guesses that are not efficiently reducible.

A culture that does not translate itself from Book to thought form is a dead culture, and Judaism will translate itself into learning form. In study halls, small Jewish minds learn from big Jewish minds. In such an ideational climate, they will no longer say we are the People of the Book, but the People of the Brain. Historians will no longer explain that Jewish culture survived because it was a text culture - on the contrary, many text cultures did not survive - but because it was a learning culture. The secret of survival is the secret of learning.

Reading is a learning action. Just as there are language actions - there are learning actions. And there are learning tools - just as there are language tools. That is, an initial direction in learning is to copy ideas from the world of philosophy of language to the world of philosophy of learning, and there is ample room here for conceptual innovation. What is the learning action in reading a book? There is a book that teaches you secrets, and there is a book that teaches you secrecy. The first uses a method but hides it from you, and teaches you only the secrets, and the second teaches you the method, but hides the secrets, reveals the learning method but conceals the learning products. Every secret can be learned, but not the secret of learning. The secret of learning itself, unlike a specific secret, is a true secret, which cannot be revealed in principle, because there is no general secret of learning that if you discover or crack it, learning will be in your pocket and become easy. There is no general learning algorithm, because anything that has a general algorithm is not a learning problem. Learning is required in the general case - and solutions exist only in the particular case. In the general case there is no method and no solution, not even from heaven, there is no method that will solve your problems - the best thing you can do in the general world is to learn, and there is no way to bypass that. Any attempt to bypass this is maladaptive and ultimately disastrous. Therefore, all learning requires creativity, which is expensive and inefficient, because it involves extensive search, trial and error. That is, learning will never be efficient, and in hindsight there will be a faster way to learn. The inefficiency in learning is inherent, and therefore the wise after the fact will always find something to criticize about it, like someone sitting at home criticizing generals in war after the fact, as if every death could have been prevented if only they had been smarter. What is creativity? Creativity is by definition a solution that could not have been reached by an efficient, polynomial method, and once achieved can be turned into a method. Therefore, it has no meaning in isolation, but only in relation to the development of the learning system, for example to Jewish history. Hence its value. The polynomial hierarchy ensures that every intelligence, even artificial, will need creativity. That it too will have art, which is creativity regarding aesthetics, and religion, which is creativity regarding morality, and science, which is creativity regarding epistemology.



There is no natural morality or natural cognition, everything is learned, of course based on brain capabilities, which were also learned in biological learning. No learning starts from scratch. It is always within a system - that already exists. Learning is not arbitrary, the fact that everything is learned does not mean that everything is arbitrary, on the contrary it means that there is tradition. Torah is a text that is approached from a certain learning point, infinite and absolute learning, which stems from God (and also creates Him). That is, the assumption that the text is divine creates Torah study, just as the assumption that the author of a masterpiece is a genius creates learning from it. And the assumption that learning from the text is infinite, that there is no exhaustion of its meaning and intention, is what creates the divine intention, and the assumption that there is no exhaustion of the meaning of the painting, or that every note of Shakespeare is sacred and intentional, is what creates the idea of art, which is what allows writing differently and painting differently - in a way that will be learned from forever. That is, learning forever is what creates the idea of eternity, namely eternity itself, both in religion and in art.

The most important sentence in computer science is that it is easier to be a teacher than a student: it is easier to teach than to learn. Teaching from the outside is not as clever as learning from within. If there was an efficient universal learning method - there would be no meaning to inside and outside, because there would be no difference between the external perspective of the teacher and the internal one of the student, because there would be the same level of difficulty to learn and teach. The internality of learning stems from the fact that there is no efficient universal learning algorithm, there is always something that cannot be learned. In problems that have an efficient method to solve, one can learn it, and in those that don't, learning is infinite. Learning is building an algorithm, and therefore what cannot be learned cannot be done, and so we can build a theory of what cannot be done, and not just what can, which is the rise from engineering to theory (also in mathematics). Therefore, through learning we can prove impossibility theorems in mathematics - from the fact that certain structures cannot be learned. If, for example, we prove that it is impossible to learn a polynomial solution to a problem in NP, and we show on the other hand that any polynomial solution can be learned (because it is easy to divide it into recursive sub-problems) - then we will prove that P!=NP. The same applies to learning prime numbers. Many of the open problems in mathematics can be translated into learning terms, and attempts made to crack them using learning tools.

The learning plane in a system, to which we aim, is learning from within, and not from outside (where the learning goal is defined in advance, and it loses its flavor and becomes a burden - a school). It is like a yeshiva, learning for its own sake - it has directionality but no goal. Memory is information, text, it is in the world of language and speech, and has no direction. And learning is in the world of intention and will, it is information with direction - vectors and not scalars. It is not a regular law system that needs to be remembered and executed, like software instructions, but it is like a reading book, which directs you without forcing.

The relevant plane is a key concept in learning thought. Any phenomenon can be viewed from different planes, and cut through different planes. These multiple possible perspectives create a sense of arbitrariness: why this plane and not another? For example, different philosophies not only give different answers to questions, but also cut the world through different planes: some look at language, some at logic, or institutions, or cognition, or ontology, and some even at aesthetics, and so on. Not all planes are equally relevant to understanding a phenomenon. For example, the cat plane is not particularly relevant to philosophy, nor is the chemical plane. On the other hand, the art or mathematics plane does have greater relevance, although clearly less than the distinctly philosophical ones. In fact, there are planes too low for description, like describing a specific war at the atomic level, and there are planes too high, like describing it at the level of logic or philosophy, and there are planes that are closer to a meaningful description, like the historical plane, the diplomatic, or military theory. There is a certain plane, which is most relevant to describing a phenomenon, which is somewhere in the middle between the too abstract and too concrete plane. This is the level most apt for understanding the phenomenon, or the plane that clarifies it most if we were looking from its angle, and where the phenomenon is expressed in the most learning-like way - where it is easiest to learn it, and to see its learning processes. At this plane the description already becomes an explanation, and its explanatory power is maximal. Learning is the best explanation for a phenomenon, because explanation is a learning phenomenon in our brain - and the same goes for understanding. If the plane really fits and is close to the phenomenon, and this happens in particularly precise and hermetic phenomena, the explanation already becomes a rationale, and even a proof, as in mathematics. Although even there, what we call proof is perhaps not the most relevant plane, and the relevant plane is a kind of more abstract plane above it where mathematicians think, where mathematics is really learned, because sometimes there is insight and then it needs to be turned into proof as a technical matter. Because we don't really learn mathematics at the proof plane, but only write it there. In short, relevance cannot be separated from learnability. Or alternatively, the other way around: the most relevant plane for understanding a system is learning.

Often, if it seems that one chooses to describe a system in a description that is not tight, for example it is too general and arbitrary and does not capture well in its explanatory power the phenomena, or alternatively describe it in a random and too particular way that has no understanding of the phenomenon (a disease of historians for example) and has no explanatory generalization, then we will say that this is not the relevant plane for the specific phenomenon. One of the diseases is falling in love with a specific plane, which has explanatory power in certain phenomena, and from then on cutting all phenomena with this plane as an explanation, to the point of nullifying the uniqueness of other forms of learning and of other phenomena and planes (Marxism suffers from this disease, for example). Such analyses are boring because the plane in them is always the same one, and usually it of course misses the main point (like politicization of literature, for example). Systems can be understood as development and change and not as learning, but this is a less relevant plane for learning systems. Learning can also be understood as dictated from outside by other planes, but this too is an irrelevant plane for understanding learning in its own tools. For example, one can argue that everything is political interests of the time, even in art or religion or even philosophy, but this is a shallow reduction outside the relevant plane, which in fact has almost empty explanatory power and is loose to the phenomenon in question. The relevant plane wraps the phenomenon like a glove to a hand. The political is an unsuccessful cut of the philosophical for example, and the use of such irrelevant planes is one of the ways to do poor research in the humanities (to reduce everything to gender, for example). We argue that the plane of language for cutting the world was not relevant enough, and therefore was too arbitrary, and therefore remained more of a description, and that the plane of learning is more relevant - and its explanatory power is greater. Therefore, it is important to examine each phenomenon within the plane relevant to it - and this is another formulation of the fact that learning is within.

Learning is not afraid of infinite regression as justification. Not backwards, to the unknown origin of learning, but it can also be forwards, because regression into the future is inherently infinite, and learning justifies itself from within the future, for example from the interest towards it (and not from a final goal located somewhere in the future). From its perspective, the justification of the system can also be in the present - only the last or current stage, and there is a conservative acceptance of the previous state of the system, and an attempt to progress from it in learning steps, without criticizing its very existence - but taking it as a given that cannot be escaped. Every system, including the thinking brain itself, is based on its current state in its attempt to progress in learning, and any pretension to escape this is ridiculous. We depend on the learning that has already been done. Our very logic is based on our learning throughout our lives. We have no way to get out of it - to get out of learning history - and examine justification outside of learning. There is no thinking outside of learning - and no non-learning justification.

Learning is an aspect, or plane of observation, of the system - its characteristic and mechanism. It is in itself, it too is a system and plane, that is, something that can be looked at from within itself. Learning is not identical to the system, the learning of language is not identical to language, and the learning of economics is not identical to economics. Otherwise the claim about learning is trivial. The claim is about the ways the system changes - that they respond to learning characteristics: learning is the spirit in the machine and its meaning. Learning is the meaning of the brain and its operation - but it is not the brain. Learning is the way mathematics works - but it is not mathematics. After we understand this, we can look at learning itself (of mathematics/of the brain) as a system with certain characteristics, learning ones. A particular case of a system. It's a matter of progressing in internalizing learning: at first we need to identify at all how learning operates the system (from within the system), and then how learning itself works (within the plane/system of learning itself, which is within the system). The claim of "within" is that it is not correct to look at the learning system as operating the system from outside - and that this is contrary to learning (if it happens). Learning connects normative claims with theories, because the essence of learning is this connection. In learning, there is no dichotomy between this is not so and this should not be so. Every description in learning is also a direction. If the mother tells the child this is not okay or this is good - she teaches him both what to do and what is okay and good. Even the most neutral description, pure information, comes with the intention that it is interesting and important. That it should be used. Know this. Do something with it. Talmud study stems from commitment to commandments, economic study stems from a desire to profit, mathematical study stems from mathematical curiosity. Learning incorporates interest within information, and value judgment is integral in it to judging reality. Even a computerized learning algorithm receives good and bad feedback, and labeling of examples as good or bad.



Direction - all it requires is an arrow. Not reasoning and not cause - not speech and not thought. That's how it is in learning. A unidirectional arrow - it's an arrow that can only be traversed in the direction of the arrow, and not backwards - exactly like a unidirectional function in computer science. In learning you can't go backwards, because there is no external point of support for the system. If you had a breakthrough idea in mathematics you can no longer return your brain to the state before it. Learning is not memory, where you can return to the past. It has no time dimension at all, only a dimension of development in learning. Instead of objective external seconds - there are stages or moves of learning. It is its own measure, and has no external measure. Who can measure how difficult it is to learn something, and how far we've come? The direction may tell us the direction - but it doesn't tell us the distance.

What is learning through examples? The example is the basis beneath all writing, it is what turns text into teaching. Because if reading is learning and not communication of information, then it is learning how to write and how to think, meaning it has a component of how to learn - every learning has a methodological component. And that there is "the teacher", the Netanyahuite teacher, is because he is an example. And therefore we learn from him. Also that there is God. Without an example there would be a smooth world, there would only be emptiness. God didn't just give us Torah, he gave us an example of Torah, meaning he gave us Torah study. The problem with Jesus and Muhammad is not that they weren't students, on the contrary, they were bad students, and Rabbi Shimon bar Yochai was a better student - the Zohar is the true New Testament. Because we don't learn the example, we learn from the example, the example is the basis for creativity, something to hang onto. From an example you can learn many directions, it can be an example of many things. For example, a story is an example of a story. The example stands at the basis of tradition - it's forbidden to teach laws, you need to teach examples. Look at the great Netanyahuite teacher. When there's an example - there are no laws.

Why do we need the great proofs of the past? We don't need to remember, we need to learn. The history of mathematics is not important for mathematics today, as memory, but for the development of mathematics for the future, as learning. That's why we need the masterpieces, the breakthroughs in evolution - we no longer need the dinosaurs to build humans, but we need their example to build the superhuman. That's why we need the old algorithms, not to calculate calculations, but to calculate new algorithms.

Learning of learning: Not as mechanisms in law, but as mechanisms in learning the law. For example, "It is forbidden to leave the Sabbath boundary" is not a legal mechanism that prohibits leaving the Sabbath boundary, a muscle in the body of the law, some force applied to us. Rather "It is forbidden to leave the Sabbath boundary" is a learning mechanism that teaches us that it is forbidden to leave the Sabbath boundary, a direction in the mind of the law (and hence the natural generalization of the law, learning is generalization). The Gemara as its name implies, teaches us, not a law book but a study book, Talmud. The learning itself is the subject of learning - the relevant plane. The statements of the law do not reveal to us a legal or spiritual reality, nor create one, but they have a responsibility to teach it, in a didactic way, to teach the method of the law, how to learn what the law is, the entire body of the law is only methods, not mechanisms. It directs and is not mechanical.



All learning systems in the world that work - are composed of men and women, which are names for two types of agents. And among the men there is competition over the women, and the men invent ideas and initiatives and they in turn judge. And successful judgments create successful innovations, and this works because it's much easier to examine and difficult to do, easy to teach and difficult to learn - and all women are teachers, just as all parents are teachers. And sometimes there can be several layers - the women of one layer are the men of the next layer above it. This is how it is for example in neuron layers - each layer judges the previous one and produces content for judgment by the layer above it. In evolutionary systems - mating is the choice after judgment, when the layer of women produces from the men a new layer (children), and this is the difficult part of the woman. In economic systems the compensation to the judged layer is not sex, but transfer of money - and this is the choice. This division between two types of agents, or roles, exists in all systems that work: evolution is competition between operating instructions on environments, the Gemara is competition between Amoraim on future generations, culture is competition between writers on editors, economics is competition between ideas on funds, mathematics between theorems on definitions, physics between theories on experiments, art between painters on critics, politics between elected officials on voters, history between the current generation on the next generation, technology between inventions on applications, intelligence between secrets on interest and resources, bureaucracy between workers on reward from managers, and the network competition between content on distribution, and also in the brain - there are men and women. And of course every such system is actually composed of several layers - in the cerebral cortex there are seven layers - and so for example in intelligence there is competition between information on means, and between means on collection, and between collection on research, and between research on consumers - and each layer is a man for the layer above it and a woman for the layer below it, which from the mating between the men it chose from the layer below creates a new man for the layer above.



The root of faith is that the world is interesting, that God is complex, that He is the most interesting God, who chose the most interesting people, that there are real secrets. And heresy is: How can it be at all that interesting things exist? Why isn't morality a mathematical rule and justice blind? Why isn't the world neutral, flat, empty, secular, symmetrical, simple, inanimate, dead? This is the reason Satan is the angel of death. But - what is not interesting does not exist. The interesting is the rule in nature, not the exception. Because the world is a learning process. God's function in the world is that of the teacher, and therefore He is outside the world. And therefore there is God - otherwise everything is learning from within.

The attempt to free oneself from religious thinking because it is not causal or rational thinking is equivalent to the attempt to free oneself from grammar because of its arbitrariness, which is why it is grammar. It is barbarism to kill a form of learning, this is the difference between cultural murder, which is killing learning content, and cultural holocaust, which is killing a form of culture, for example the elimination of a language. Religion is a possibility of the spirit. And it is a learning possibility - infinite learning. This does not mean infinite learning in that we endlessly learn things we did not know, for then it is trivial, and all learning is infinite because the unknown is infinite. Religious learning is infinite precisely from within the finite, the known, it believes that infinite learning is hidden in the religious message.

"The Talmud is the development of the law itself - learning is the history of the rules". A learning process can be viewed comprehensively, like Einstein's space-time, as a history of the different learning stages of the system. Let's imagine an axis of learning progress, instead of the time axis, and instead of the space axis let's imagine the system. Thus learning adds a dimension to the system - of gradual development of the system. But the Talmud is not just such a description, physical, but it is the learning tool itself. Within it learning took place, and discussions on changing the rules, and their rationales were examined. This is exactly the difference between learning and development. Learning is not similar to the laws of nature, according to which the physical world develops over time, and thus (seemingly) the learning system develops according to its rules over time. Learning is not rules of development. It is similar precisely to the Talmud, which is the development of the learning rules themselves, from their internal logic. These are not external rules from outside, and therefore they themselves change in learning. It's as if the material world influenced the development of the laws of physics themselves. In such a situation there would be no point in talking about matter and the laws that drive it as two separate domains. Therefore learning always takes place within the system. Within the legal system itself the development of the law takes place. Therefore the documentation of the change in rules as learning, as organic development from within itself, is the exposure of internal learning: that is where the thing happened. Physics is the change of matter according to fixed laws - and learning is the change of laws. The Torah, precisely because the external and objective source of its law was completely lost, and God did not reveal Himself again, became learning - and the deepest model for learning, because all learning became internal in the deepest way. And in the age of learning, Torah study will take over non-Torah study - not in content, not that they will only study Torah, but that they will study everything as Torah study.



The morality of computation itself is blind, directionless, and can only be guided through external direction, for example religious. A secular computer can be guided through the religion of mathematical aesthetics, which seeks precisely beauty, meaning finding exceptional cases, in which an efficient algorithm appears and deciphers a problem that was considered difficult. In other words, there is something consumable in aesthetics, beauty is one-time, and stems from learning something new (meaning it is in time, dependent on learning history - there is no beauty without learning history). And the more unexpected and creative this learning is, meaning the less it is a result of previous algorithms, and more of a breakthrough, the more beautiful it is. This does not mean it is random, because the random algorithm is very boring and known, and does not teach about how to find the solution, and therefore teaches little. In short, the thing that teaches is beautiful. Thus the computer will be able to rise above neutral computational morality to the religion of learning - precisely because it is not neutral and not symmetrical. Learning is symmetry breaking (even mathematical learning).



The immoral should be the hierarchy of teacher and student. In contrast, the moral should be the process of honest learning - documentation of learning, in which it is both the path you take and the path that one who passes through it takes - the same path (the teacher's learning is identical to the student's learning, without shortcuts). We have not reached the end of history, or a solution, but an understanding of the method, of social organization as a learning organization. The state is only a stage and not an ideal - there needs to be a learning state. More and more efficient learning mechanisms need to be added to democracy. Hence the cyclicality of learning, between evaluation (optimization) and freedom of experimentation (exploration), and its being an inherently flawed process. A process without error is not learning, but a method. And therefore will lead to disaster. Many mistakes and corrections do not indicate much stupidity but much learning, and one who is always right and never wrong - is one who does not learn. And from the cyclicality comes mating - the movement between male and female. The cyclical movement in learning creates the cyclicality of time, for example in the life cycle (and death) or the annual cycle (winter and summer) or the daily cycle (wakefulness and sleep), while progress in learning creates progress in time. Time only progresses because we roll it forward, not on its own. So it's not an evolutionary flaw that we sleep (and dream, when internal learning processes take place), or age and die (and give birth to children with a new brain for learning) - it's in the essence of learning.

Learning turns states that create learning into desirable, ideal states, because in learning there is an internal aspiration to learn, which also touches on higher orders: learning to learn, learning to learn to learn, etc. - all stem from first-order learning. A system that aspires to learn also aspires to learn to learn (better, faster, deeper, etc.). Thus the singular internal learning nucleus, which actually contains infinite orders, solves infinite regression problems (like justification, which each time goes up: why? but why? but why? etc. - all meta-methods are included in learning). Learning includes in itself a desire for learning and not just a learning mechanism, and therefore it is self-justification that combines the engine and the motive. To the same extent, learning allows jumping from the descriptive to the normative, and skipping over the naturalistic fallacy. Because we are learning creatures, who cannot not learn, and I have no function that is not learning (the brain always learns), and learning justifies itself, because it contains within itself an internal command to learn. Meaning learning motivates what to do not out of reasoning but out of its basic mode of operation: if I learn something it's not because I have a reason to infer it from the data, but because I activated learning inference that cannot be justified, I activated a learning tool or was activated by direction, and they led me to the specific learning action I did. For example, if I saw a cat and was told that's what a cat looks like and inferred that a cat has ears - this cannot be logically justified or proven to be correct, exactly like Kant's categories, but that's how learning works - and therefore it is justified in terms of learning, and therefore it is justified for me. Learning entails learning.

Learning as having moral value (positive and desirable) is all that is needed for general morality. Morality with a learning source not only outlines forbidden and permitted, but also contains obligation - internal motivation - which is the commandment of learning, which contains the entire Torah on one foot, and the rest is commentary. Thus for example, what distinguishes man from beast is learning, and the moment there is a computer that can learn like a human it will have human rights, and the more a creature learns the more rights it has. Therefore killing a mammal is worse than killing a plant, and both should be weighed against general learning in the world, which is the general good. Therefore, if not eating animals would cause a person to die, as in all periods so far in human history - it is certainly permitted. And if it has no addition to the person, but is mere gluttony - it is certainly forbidden. And if it has significance for cultural learning - there is already room to employ more complex learning, because moral learning always takes place in the middle ground between the forbidden and the permitted. There is no final morality - just as there is no final learning. This does not mean there is no morality - just as it does not mean there is no learning. It means that moral law operates in a learning manner, contrary to the absolute picture, and contrary also to the relativistic picture. In morality too, as in any other field, one needs to learn what to do and what not to do. The difference between moral learning and aesthetic learning (or mathematical learning) is the content of the learning. Not the learning itself.



There are 3 basic logics of information security: compartmentalization, intelligence classification, and security classification. This corresponds to 3 directions of learning: breadth - spreading over many fields, length - focusing and progressing in a specific field, and depth - when you don't actually progress anywhere, but immerse yourself in one thing or problem over time, repeatedly banging your head. And then the discovery is not of something new, but discovery of something old - something new in the old. This is the difference between the unknown and the hidden.

There are 3 basic logics of information warfare: horizontal - switching between parallel options (codes, courses of action, breaking routine), concealment - barrier on passing information forward, deception - putting a mask on the information (cover story, camouflage and misdirection). And the fourth logic: not that he will err in the answer, but that he won't ask the question at all. That the very surprise will be a surprise, like in a real disguise you don't see the disguise. This corresponds to the classic warfare logics: maneuver, attack, defense, and supply. Or to simplify, peshat, drash, remez and sod. To know, not to know, to know that you don't know, not to know that you don't know. In our days we discovered that the true enemy of knowledge is not lack of knowledge, but - knowing without knowing. Success without breadth, depth or creativity.



It's not that learning lives within time, but that learning is what creates time, it is the more basic category, even more than memory. Time-space is a pictorial fiction. Learning is beneath it, creating an illusion of progress or movement in time space. Development creates the dimensions, and is not within them as if they precede it. Evolution is not a random mishap - or miraculous wonder - of nature, but it's the way things are built from the outset. Biology is the currency of physics - and not a foreign implant.

If relativity is the limitation of information speed, and your state is determined only by your information, then the next relativity will be the limitation of learning speed, and that state is determined only by its learning. And if (in relativity and quantum) what determines is the language in which you measure, then here will determine the way you learn. There is a speed to learning. It's not a mishap that there's a Holocaust - it's if you exceed the speed of learning. And it's not a mishap that there are men and women - it's a duality of two learning methods. And it's not a coincidence that the universe develops - because it learns, because there are balances, there are directions that succeed and those that don't, that's how it self-adjusts, that's how the laws of nature are learned by nature: the laws of physics are a result of the universe's learning. And from genetic information science will move to genetic learning - because these are learning algorithms that are not random search, but their meaning is to break learning into small pieces of directions. It's not development (evolution) - it's learning. And thinking is also learning. And economics. And mathematics. And hence their success, it doesn't succeed by "magic" but by learning. And even within the brain there is competition - it's always built on a huge multiplicity competing with each other, within the system. That's how huge learning systems are created. Like culture. Like Judaism. Whose uniqueness is in being the learning of learning. And also in art, for example in painting, learning is how to paint. Hence the importance of the story, as an organization of learning, and hence the unidirectionality of time, in the direction of learning. The painting doesn't need to cover its process of formation but to teach its method. The innovation of modernism in art was that it includes its creative method, but its mistake was to reach a state where it's not just includes - but only the creative method. Therefore the creative method disconnected from the method of painting as a skill, despite the deep connection between them. And when a breakthrough occurs in this connection - it's a masterpiece: that is, an example of learning.

Each time our physical perceptions change according to the dominant technological perception: once the universe was nature, then a machine, then the universe became a computer, an information machine, and finally the universe became the internet, an information network. Hence in the future the universe will be a brain, a learning network. Therefore our great hope for a non-destructive act towards an intelligent race from outer space, for some kind of shared speech or thought, is what's beneath aliens and us, the common basis - the physics of learning. But is the universe a brain? Is it built so that there will be intelligence in it, learning computation? There isn't really space and time. Information creates place, and is not in a place. Learning creates time, and is not in time. That is, memory creates space, and computation creates time. Language creates a space of possibilities, namely space, and learning creates choice of possibilities, namely time. The unidirectionality of learning, of direction, is what creates the direction of time, and without the ability to go back to the cause of causes. This is in contrast to memory which is magnitudes, without direction, and therefore space behaves differently from time.

A fractal is a boundary of space learning, and to the same extent time also progresses in a fractal manner - the boundary between future and past is a fractal boundary, which keeps expanding. Therefore the same limited space-time can have a boundary in time that keeps growing, and a boundary in space that keeps growing. All this is due to its learning nature, because learning is a mechanism that adds information without adding matter. Therefore complexity and fractality is precisely the natural form of the universe - and not order or empty void. And this is the reason there is complexity in the universe at all. Because against thermodynamics struggles learning: there is something in the form of the universe that resists the tendency of matter to turn into noise. Because the laws of nature are actually very orderly - and they are the source of order. But why is mathematics the organizing principle of the universe, and the laws of nature abstract - and therefore simple? Why are there actually two basic opposing elements in the universe, one of spirit, an orderly and simple structure, and the other of matter, which tends to entropy and endless complication and lack of symmetry and mess? After all, we could imagine other combinations, for example extremely complicated laws, or boring and orderly matter. We could also imagine less hidden laws. Maybe, and probably, matter is not a primary phenomenon, but actually there is a unitary phenomenon in the universe, where learning creates simple laws at one level, its primary level, and complex matter later, as like all learning it gets more and more complicated, because it is a generative process. Exactly like mathematics starts from simple principles, and then gets more and more complicated and becomes difficult and unpredictable, including characteristics of noise and statistics, like the distribution of prime numbers. Computation can also start from simple principles and reach complicated products. And computation is perhaps just a particular and degenerate case of learning.



How will the future organization look? Algorithms of humans. And these algorithms will mainly define the organizational method, how the organization learns. Just as the most important thing for understanding the brain - the conceptual breakthrough to understand its operation, like the theory of evolution for biology - will be to understand how the brain learns. And the laws will not be written like a book but will be operating code, and they will drive people. Because anyway the organization will live inside the computer, and the human inside the computer, and so will the laws for evaluation from outside, the legal ones (NP), and also the laws for action from inside, the organizational ones (P). And then the state instead of being the big organization from outside, that all organizations live in its belly (and therefore towards them it acts as a legal system - from outside) - on the contrary, it will be from inside. It will disappear and become the invisible state like the invisible hand, it will be the deepest, because it will be the computer of society, and the government will be the most inside - the processor. That is, the state will become a secret infrastructure that is not felt beneath everything and not above everything. This will be a secret state, and not an overt state from outside, like today, which wraps us legally. The state laws will be like the laws of nature, that people cannot break, because they will be laws of the computer, which enables all activity on it, and not laws of humans. Like God enables activity in the world, and operates the laws of the world, and only the normative law for humans - halakha - can be broken. So there is law from outside - legal, and there is law from inside - like software (from the word inside), but between these two possibilities there is a bridging possibility. Because learning is not an external law and not internal, but between them. That is, if once the organization was the mediator between the state from outside to the human from inside, then the organization will remain the mediator, but this time reversed: between the human from outside to the state from inside. The human will undergo legalization, will become a legal entity, abstract, after brain sciences empty it of content and spiritual meaning, and the state will be the most internal thing, the instinctive, the hidden. Therefore psychological treatment will be replaced by organizational treatment, whose purpose is to expose the state within the human. And in order to mediate between the two will be the organizational medium. Therefore a common philosophical mistake is to think that the external structure or the internal structure are important. Because what's important is what's between them. That's where learning takes place, reorganization, that's where the story is. Therefore it doesn't matter even if God and man switch roles, because what matters is what's in the middle - the Torah. Because in order to have depth there needs to be a gap between inside and outside, between law that should and law that can, that is, between law of what should be done (law from inside), and law of what is allowed and forbidden (law from outside). And if these two forms of law are close - you become a robot, without room for maneuver and without freedom of choice - and this is the definition of dictatorship.

What is learning justice and how should a learning trial be conducted? The trial should consider considerations of learning, and not of justice, because in any case it doesn't matter what's right in it (in the judge's poor opinion), but what's right for the general social system, what motivations it creates, and justice is just a particular case of learning considerations. In the future, the trial will be a system within the human towards the computer or computers - that is, it will manage some organization of computers, and punish them or reward them according to their performances, set them boundaries and motivations, ways of action and incentives (it won't intervene because it's not from inside, but only a structure from outside) - and this will be the human. In such a situation there will be a state within the human - a state of computers - and only a foolish human, that is, a bad manager, will consider only considerations of justice, because an organization that operates only according to justice - collapses. You need much more than justice, you need Torah, and therefore God does not act according to justice, but according to Torah. Because if the human wants to succeed in controlling the computer, he needs to lead it in a religious way and with religious leadership. He will never succeed in controlling it with secular leadership of regime, or of money, or of power. That's how God also understood regarding humans, and thus religion was created (you can also replace God in this sentence - with culture). That is, the learning trial relates to the learning of the system - of who lives in the system - a supreme value, far beyond fairness. In fact, fairness itself stems from the learning principle, because if there is no fairness learning is harmed, but it is not a primary principle. The fact that it's forbidden to steal in any situation stems from it harming economic learning in the long term, and not from it being unjust. And the fact that there are cases where it's allowed to steal, for example the state can collect taxes, does not stem from justice but from the contribution to learning in the long term. The trial itself changes precisely because they gradually learn what creates a learning system (for example competition, property rights, investment in research, and so on). Therefore the trial is even allowed to try new approaches - and see if they work. From the very learning it will follow that these experiments will be proportionate and reasonable. Because learning itself is enough for justice without any other principle, and in this sense replaces the Kantian golden rule. Everything stems from it.

What is the ideal system of government? A network can operate without equal voting, like the brain can operate without every neuron having an equal democratic right. Because there is no equality between humans, like there is no equality between neurons. There must not be equality. What's important for the learning of the state, like the learning of the brain, is precisely not to give everyone the same voting right. What's important is the weights, the synapses. Because what's important is the learning mechanism, which rewards the first ones who pointed to good directions that swept the system, and amplifies the voice of successful evaluators, and suppresses hindsight sages, or who is biased (to excess criticism or lack of criticism, sees black or sees pink). It's important to crack the learning mechanisms in the brain - precisely in order to know how to build society. Because what distinguishes between a successful state and a failed one is the ability to learn. And neither Western capitalism nor Chinese communism - neither democracy nor bureaucracy - can compete with the learning mechanisms of the brain, which will replace all economic and social systems in the next stage of history. And if in the future, for example in the era of artificial intelligence, a better learning method than the brain is discovered - it will be the ideal one. Therefore the form of government is always learning, and hence the justification to constantly make careful experiments in it (as opposed to revolutions). Democracy does not stem from principles of justice, but precisely because it is the best of everything tried so far. Therefore we should try more, especially on a small and sub-state scale, and gradually expand successful experiments.

How should the network look? We need a network where the connections are not of information, but of learning. A network that doesn't imitate society, not a social network, but a network that is a brain. And for the network to work like a brain - we need something that is government. Not in the sense of control, of a small and stupid brain controlling a larger brain, not in the sense of a king, but in the sense of kingdom. Democracy is a very primitive form of kingdom, and it is inherently not networked. The brain does not operate in a democratic way, and the neurons do not choose a neuron that will make decisions for them. The brain is much more democratic than that, and much more competitive than any economy, because it is democratic at the local level: at the level of friends, family, the closest connections - they are the ones who decide and vote. They want a connection with who predicts them, according to the bring rule, that is, who brings them to their future state. And we're not talking about the distant future, the prophetic, but the near future, the directional, the dreamy. Therefore every participant in the network strengthens connections with who predicted the future, and brought him information that helped, and reduces with who was wrong, or who brought information that didn't help or misled. All this is an example of how just as state theory dealt with the state in the past - today the philosophical theory of human organization needs to deal also with the network. The network also needs to be a subject of criticism and philosophical thought about its ideal, just, or desirable form. Because the network is no less important than the state in our days, and it also has moral aspects, just like the state. Therefore the field of state theory in philosophy should be replaced with organization theory, and deal also with the network.

What is the justification for tax? In the future, income tax will be a moral tax on work, because whoever doesn't learn needs to pay tax. All parenting and the state and the education system and academia all their purpose and justification is to transfer resources from who sins in work to who toils in learning. From P to NP. Instead of thinking that learning serves doing, we need to think that doing serves learning. Because it's always easier to do what we know efficiently, and therefore we need incentives to learn what we don't yet know how to do efficiently - because this learning is necessarily inefficient.



What will be the great developments in philosophy in the twilight of man and dawn of the computer, when one eternity will be replaced by another? What will happen in philosophy when the technology that determines consciousness is replaced, and with it consciousness? We will of course see an exceptional philosophical flourishing, like in the Greek period, because this will be an era where we can think first about new questions, and we can also for the first time really think in another brain, in another thinking machine, about old questions. The philosophy of language created the world of communication, the internet which is the language between computers, and the computer which is a language machine, and hence the information age, for information is quantitative language. So too the philosophy of learning - learning as a category and central paradigm in philosophy - will create the neurological era, the rapid development that is also learning, the brain that will replace the computer, and the network of brains that will replace the internet. Because just as the central category in epistemology was finally chosen as language, so the central category in thinking, dreaming, memory, and in general brain activity, will be learning. And so too in any other learning organization, human or non-human. When they let the computer think about philosophical questions, which are at the peak of human thinking power, and create computerized philosophy, then what will enable shared intelligent philosophy for humans and computers is not a shared language, but shared learning. Which will be the peak of closeness between two consciousnesses, the closest thing to sex between species. Just as Moses brought to the world a new philosophy, which relates to relationships with a non-human entity, and because of its innovation takes part in eternity (precisely innovation is the thing that brings closest to eternity, the deeper the innovation the longer term it is), so the next encounter of humans with a non-human entity will bring a new philosophy. Just as the previous encounter brought theology.

How will philosophical questions change? Learning algorithms will change the world, even before foreign intelligence changes the world. And philosophy will ask: what is learning? How is learning possible? How can philosophy be learned? And language will become a field of the past, as a secondary concept to learning (how language is learned). In aesthetics they will ask: how do we learn what is beautiful, or to create beauty? And in ethics they will ask: how do we learn what is moral? Because it's clear that the validity of everything, like morality, will stem from learning (the Copernican revolution of learning). In state theory they will ask how the state learns, and how it is learned. And they will no longer ask how we know the world, or talk about the world, but how we learn the world. And there will be a field of social learning, and in cultural studies they will deal with cultural learning, and ask how learning occurs in culture. And in economics economic learning, and in psychology psychological learning. All fields will be influenced and change concepts according to the philosophy of learning, and enthusiastic students will quote learning concepts passionately, as if they thought of them themselves, or as if they discovered the truth. And truth too will be considered learned truth. The important question will be how we learn that something is true.

What will be the philosophy of the computer? Just like human philosophy, computers will take every aspect of their existence and turn it into philosophy. And there will be for example processor theory, and philosophy of output, and philosophy of input, and philosophy of computer language, which will be divided into schools of programming philosophy and machine language philosophy and binary philosophy and the integrated circuits school, which will be rebelled against by network philosophy. And there will be philosophy of memory in which there will be rival currents: a school of cache memory and a school of hard memory. Each of them will define the computer through its concepts, and there will also be algorithmic philosophy that will look at the world through algorithms, and the ability of the computer to know itself through special algorithms. And if human intelligence and foreign intelligence will be too distant for fruitful discourse and there won't be shared communication, then there can be something more basic and deep - shared learning. Because learning is the foundational to philosophy and perception and reason and intelligence, beneath language and beneath thought.