Future
Behind the Most Disruptive Technology Ever
How Generative AI is going to accelerate the coming of AGI

The excitement around Generative AI
The entire Internet seems obsessed with ChatGPT, and there is no end in sight. It is difficult to imagine a single regular user of the World Wide Web who has not heard of it.
As always, in such cases, reliable information is densely mixed with rumors, speculation, exaggerated delights, and fears, and not everyone can separate one from the other. All this unnecessary fuss makes many people skeptical about this phenomenon, even to the point of refusing to find out what it is.
However, as they say, there is no smoke without fire. There is always skepticism about any promising technology, regardless of whether its potential is exaggerated or underestimated.
In the first case, skeptics will save time, nerves, and, possibly, money. In the second case, it will be too late for them to take advantage of the technology, which sometimes means equal losses. Thus, any interested user constantly faces the same question — how to get a reliable picture of all the pros and cons of what is on everyone’s lips in the least amount of time.
However, the value of emerging technologies is often impossible to measure in terms of banal practical benefits. Because it can also happen that in the development and spread of technology, its very purpose changes — from a means of satisfying human needs, it turns into a self-developing entity that influences society due to some inherent teleology.
I present in this article the point that ChatGPT is one of these phenomena, and behind its appearance, it is much more than a tool demanded by consumers and even more than technology. As an instrument, it contains as many, if not more, flaws as its virtues, and if it were only about what is inside it, there would not be such a fuss around it. In my opinion, the reason for that incredible excitement about ChatGPT is in some anticipation of the beginning of colossal changes in the life of modern society, of which it became the conductor.
Not true intelligence
So, the first thing to note is the stark contrast between what ChatGPT can do better than is expected of it and what disappoints its users. This difference matters because the technology’s path in consumer society depends directly on what users want from the products it creates.
It should be made clear that the term “AI” applied to ChatGPT — is a standard marketing gimmick. This is just a trendy label that the owners cling to all products of that class. None of these products has true intelligence.
ChatGPT (generative pre-trained transformer) is an implementation of Generative AI (as one of the varieties of so-called “task-specific” or “Narrow” or “Week” AI); it embodies no more than a “linguistic model,” as its creators keep reminding, perhaps in order not to give the reason for inflated expectations.
There is a chasm between this model and natural intelligence. The last one works through a universal analytical mechanism, whose ability to understand its environment does not depend on the content of that environment.
All current implementations of “AI” can only do so in a chosen subject area. True intelligence can predict the general state of the surrounding world due to changes in any of its parameters — precisely because its above-mentioned analytical mechanism is universal. This mechanism allows it to take into account the cumulative influence of all the phenomena it recognizes on each other.
For this, it needs only incoming data and time to build an internal model of the world. Narrow AI stays forever within the field of exploration in which it is trained. It has no abstraction mechanism to make sense of the world around it as a single picture of interconnected entities. It can be a perfect chess or Go player, but it cannot tell the difference between a dog and a cat or understand that thunder will be followed by rain unless it is programmed to account for such a pattern.
The creativity problem: the inability to generalize meaningfully
This difference thoroughly explains the limitations of ChatGPT.
Although they may go unnoticed for a long time, it is easy to spot if you want to. It cannot do truly creative work that requires semantic generalization. Although its developers claim, for example, that it can compose stories, as a matter of fact, it cannot. It can generate some text if asked and even present it supposedly in the style of some famous writer, but the result would only be misleading for a child or someone who is not inclined to read. The texts issued by ChatGPT lack narrative coherence, anything that might be called the essence of the artwork. Instead, it tries to feed you some kind of textual Frankenstein, a lifeless array of phrases cobbled out of bits of text. Worse, it cannot write a coherent story even if you offer it a scheme, no matter how much you ask it to do so.
You can somehow squeeze something out of it through specifically structured queries (this approach is called Prompt Engineering), but in the end, it is hardly worth the work: my desperate efforts to get it to develop the plot given to it resulted in something that is at most a fragment of a cheap female novel. As for its attempts to present the generated text in the style of this or that writer, they ended in an awkward caricature, which would have sent the writer into a tizzy.
During my experiments with this hapless graphomaniac, I immediately recalled the storyline from Orwell’s 1984, which referred to some kind of generating machines of mass art, dull and meaningless but satisfying the aesthetic demands of the “proles.”
All in all, it’s sad with the writing. Please don’t do it with it. Don’t do it yet.
Within and beyond the linguistic model
So, if we talk about true intelligence, ChatGPT certainly does not have it. To date, the only machine that generates intelligence is the human brain.
But this does not mean it is impossible to implement real intelligence on an artificial medium. ChatGPT is organized on principles that will never allow it to get above simulated intelligence, and yet I insist that it is the thing that has opened the door to the future where real, not nominal, intelligence will be created that will change our world in the most decisive and dramatic way. The limitations of the linguistic model do not prevent the development of the field of AI, which is much broader than the particular field of Narrow AI. I will return to this question near the end of this article, but for now, let me say that despite the objective limitations of ChatGPT, we have every reason to believe two things:
1. It’s the most disruptive technology since at least the beginning of the Internet
and
2. ChatGPT is the harbinger of the advent of true artificial intelligence. The same one that journalist and writer James Barrat called Our Final Invention.
Let me start with the first one.
“Disruptive,” not “transformative.”
Sometimes the term “disruptive technology” is misleading. Some of such technologies would be more appropriately called “transformative.”
By transformative technology, I mean those that, while significantly transforming the technological landscape, do not destroy other significant technologies.
For example, 5G. It replaces 4G, adding a lot to the current infrastructure, and acts as a bridge to many different technologies (IoT/Edge, AI, Telemedicine, Internet of body, and so on) but only to a small extent gets rid of the legacy of the previous standard (5G is the next link after 4G in the chain of so-called LTE — Long-Term Evolution, which is a standard for wireless broadband communication for mobile devices).
But that’s not the case with ChatGPT. It is precisely the “disruptive” technology that has already posed an existential challenge to the largest segment of e-commerce — the search engine advertising business.
ChatGPT is extremely good at finding information but doesn’t just find “more relevant” results. It’s about how it does it and how it provides the user with the requested information. In the perception of the latter, it does not search for it but answers the question asked. While the output of a normal search engine is a list of ranked sources, ChatGPT spares the user from this stage and gives the prepared answer in a single text.
This is where all the magic happens. The user doesn’t have to do the tedious and often unproductive work of analyzing data presented in a results list. What the user sees is an aggregated result obtained by extracting data from multiple sources, ranking it internally, and finally composing it.
All this is hidden from the user’s eyes and is done in a fraction of a second. Thus, ChatGPT radically reduces the search cost and logically completes it, relieving the user of the feeling of uncertainty. Of course, it allows you to expand your query by simply asking in plain language, something like “tell me more about <subject>” under the sole condition that the subject is mentioned in the current conversation. This works precisely because it remembers the context of the whole conversation with the user.
A nice addition to this is that it can enrich its output format with additional structural elements, such as lists and even tables (although this is not always possible). Another powerful option is the ability to “personify” a query to get a semantically more specific answer, which cannot be achieved by a search engine.
Google Query: What was Krugman’s argument on the entire matter dedicated to the crisis of 2008, and how it contradicted (or, at least, isn’t aligned with) Hayek’s view?

ChatGPT Query: Tell me, what was Krugman’s argument on the entire matter dedicated to the crisis of 2008, and enlist those of them which contradict (or, at least, aren’t aligned with) Hayek’s view?

ChatGPT thus has a clear advantage over modern search engines in several respects, changing the rules of the game for the entire WWW before our eyes.
Indeed, how attractive would it be to go through Google’s endless search results to find the information you want somewhere in the piles of the generated list if the correct answer could be available immediately, with fast delivery — here and now — and without any irrelevant data around?
I’m far from saying that ChatGPT will destroy the current search advertising business model, but it will probably reduce its profitability and force it to change. Google has already felt the effects of this pressure enough. I don’t know how much traffic to its search engine has declined, but I do know that ChatGPT’s user growth rate is breaking records. It reached its first million users in just five days (!) — an absolutely fantastic indicator.
By comparison, it took Netflix, Facebook, and Instagram 41, 10, and 2.5 months, respectively.
This challenge to Google’s core business (about 58% of revenue) prompted the latter to hastily introduce its own AI chatbot, Bard, to the public in early February. However, for many, the move looked so ill-prepared that it seemed desperate, if not miserable.
Unfortunately for Google, not only was the presentation unconvincing, but Bard’s abilities baffled the public as one of its answers proved inaccurate. The result was an immediate 9% plunge in Alphabet stock, which equates to $100 billion.
Of course, there are several complaints about ChatGPT, which can be divided into functional and qualitative categories.
The first is that it is not yet able to upload images, provide links to information sources, and often — give accurate quotes.
The latter includes not always a satisfactory interpretation of the questions asked by human and the fact that its output is not formed based on real-time data but from the accumulated to 2021. The latter limitation raises the most questions since OpenAI uses this data for commercial purposes.
First of all, along with the free version of ChatGPT, there is a paid one, but the list of benefits of the latter does not include work with real-time data. It is unknown how often this data will be updated and under what conditions the updates will be available.
Secondly, OpenAI offers access to its models and data via API, but that data is still the same. Up-to-date data is available from search engines, where GPT-3 models can be integrated and should be expected soon (you.com has already done so; they claim to use GPT 3 along with their own model).
Not only about business
But all of this is about business, whereas the appearance of Generative AI at the ChatGPT level marks something more — the dawn of a new era. None of those mentioned above shortcomings obscure the fact that how ChatGPT presents its way of interacting with users is revolutionary for the entire industry.
Its ability to “understand” the context of a conversation is something never before seen indeed. Even when it doesn’t quite get a particular question right, it is neither frustrating nor irritating because such a hitch feels like an imperfection inherent in a capable reasoning being rather than the meaningless aloofness of a soulless machine. Eventually, it will give you a satisfactory answer, one way or another, if you clarify your question or refer to the part of its previous conclusion. Interacting with it feels precisely like a meaningful dialogue — unless, of course, you ask it to reveal its literary talents.
It makes all the difference since this experience alters our attitude towards technology, effectively making alienation between us obsolete and turning “them” into the continuation of “us.”
Undoubtedly, GPT technology will be integrated not only into search engines but also into countless digital products and become an integral option for web browsers (as Microsoft has already announced).
However, even the above advantages of ChatGPT over traditional search engines are only part of the picture. It can provide extremely useful, up-to-date, and ready-to-use technical information. For example, it can explain an IT concept in a clear, simple language, tell you how a particular programming language is structured, and accompany its explanation with examples of actual program code. It can explain what your own code does, find bugs, optimize it, and test it. It can even create a website based on your very generic instructions, expressed in common everyday language. For those whose job is to solve the endless puzzles of the IT field, it all feels almost like a touch of the Holy Grail.
Looking at all these wonders, A. S. Clarke’s prophetic phrase that any sufficiently advanced technology is indistinguishable from magic comes to mind.
The Global Paradigm Shift: From Data to Knowledge
It remains to be seen how much will be invested in products based on GPT technology. But there is no doubt that they will be enormous. Microsoft already announced financial injections of $10 billion into Open AI in late January.
This, of course, is just the beginning. The tech giants’ intentions are apparent, and it seems that for them to invest in GPT or not is not only a strategic question but also an existential one. It happened quickly, unexpectedly, and turned out to be so serious that everything else has so far receded into the background — the Metaverse, the massive rollout of 5G and related technologies, and even the business impact of Sustainability requirements — all look much less relevant now than they did just a few months ago.
All indications are, therefore, that another IT revolution is unfolding before us, and its impact on modern society promises to be more significant than ever.
I would venture to suggest that it will be driven not only by introducing new, super-performative tools for handling information but also by deeper reasons affecting the fundamentals of how we humans interact with our environment in the broadest sense.
Like all other living beings, I, you, and every human being can be considered information objects — in the sense that information processing is an inherent function of all organisms. However, Homo Sapiens stands apart here. Our biology does not limit the scope of this function. For us, it not only serves as a means of survival in the external environment but also allows us to change that environment; thanks to this ability, we have emerged from caves and built a high-tech civilization.
Although this function has natural limitations, we can enhance it by acquiring, deepening, and applying our knowledge. The possession of knowledge is our categorical distinction from animals and, if I may use a pompous turn of speech here, a means of gradually freeing man from the arbitrariness of the soulless material universe.
Thus, the more effectively we can acquire this knowledge, the less the fate of our civilization will depend on this arbitrariness. In this regard, the most important significance of the emergence of ChatGPT lies in its role as the initiator of a breakthrough step toward the intensification of knowledge. It can be seen as a catalyst for changing the information paradigm of modern society. So far, this paradigm has been based on the recognition of the following sequence:
1. Problem definition 2. Collecting data for analysis 3. Conversion of data into information, and 4. Converting information into knowledge
But the emergence of quasi-intelligence of Narrow AI gives humans the prospect of outsourcing the 2nd and 3rd stages to it.
It already does many things better, faster, and more accurately than we do, and sometimes the difference is enormous. It’s already hard for us to imagine a world without many of the services and products that AI is behind. This means it will continue to improve, and of course, the more advanced it is, the more intensive will be the process of transition to the next information paradigm, based no longer on four but only on two stages:
1. Problem definition and 2. Acquiring knowledge
This change assumes that the need for individual possession of highly specialized skills will consistently diminish, and many of them will move from the human mind to the AI repository.
The Great Transformation and Its Meaning
The significance of this transformation cannot be overestimated.
It will free the human mind more than ever from routine, allowing it to focus on tasks that require creative thinking.
It should be noted, however, that this shift is not something unprecedented in our history. In fact, it is an ongoing process. As science and technology have advanced, the level of interaction between the human mind and its environment has become increasingly abstract. We now do not need to directly control many processes for which control was necessary for the past. In the human-made world, these processes are controlled by machines, which in turn are controlled by other machines. So, basically, it is enough for humans to control the top level of this machinery.
Such progressive enhancement, however, has a downside.
First, the social consequences are hard to predict, and the second is the challenge of managing increasing complexity. It would be naïve to assume that this time we will not be faced with even more challenging questions requiring urgent solutions. When progress radically affects the functioning of society, tensions invariably arise. Even if people are protected from physical or material threats, they may find themselves unprotected psychologically because of the breakdown of habitual ways of life, the natural conservatism of thinking, and the fear of an unknown future.
Thus, the main threat to human well-being may not be the loss of income due to the disappearance of jobs and entire professions (as more and more is said each year with the increasing pace of automation), nor the lack of available energy or environmental problems, but the mismatch of progress with the way of existence and social interaction that people accept. We all witnessed the severity of such issues during the COVID-19 pandemic. Still, while this turned out to be a transient phenomenon, the changes in question are likely irreversible.
As for the problem of increasing complexity, delegating more and more control to machines is its focal point. The problem is twofold: it is not only a quantitative increase in this complexity but also the fact that from a technical issue, it turns into a meta-problem as humans hand over to machines some of their primary prerogatives: to think. In this regard, what Agent Smith said in “The Matrix”
…I say your civilization because as soon as we started thinking for you, it really became our civilization which is, of course, what this is all about. Evolution, Morpheus, evolution.
…takes on a connotation beyond pure sci-fi, prompting the human mind to think more and more about the unintended consequences of the STP.
Yet we hardly have a choice whether to continue on the path of progress or stop.
Obviously, we cannot stop, not with our insatiable thirst to push the boundaries of the possible (though some would prefer to invoke the trivial positive feedback loop between progress and the satisfaction of consumer demand).
Consequently, our world will continue to become more complex, and we will only be able to cope with this ever-increasing complexity if we rely on our technology. So, we go back to the need to delegate some control over our environment to them and retain supervision at the very top.
Of course, it makes more sense for us to be realists than optimists and to begin by assuming the not-so-comfortable fact that we are not guaranteed to accomplish this task.
The dangers awaiting us along the way have been and continue to be spoken of by many wise people, and these concerns should by no means be ignored. But another obvious truth is that we can be guaranteed nothing in the universe we find ourselves in. We cannot protect ourselves from a gamma-ray burst, an asteroid collision, or even — if you like that development best — a visit from unfriendly aliens.
We can only achieve what we can, but this is undoubtedly better than inactive humility. By acting prudently, purposefully, and cooperatively, humans are capable of phenomenal achievements. In fact, our civilization is itself a miracle created by the human mind and its crucial achievement, our technology. Even though people have often misused it and, even worse, used it deliberately against each other, it would be unfair to say that history teaches us nothing. We may not be straight A’s in its class, but we are not failures. Humans are not devoid of common sense, though it is impossible to know in advance whether it will kick in before it is too late.
Approaching a tipping point
Generative AI technology is still relatively simple, therefore, not too dangerous. For all their power, they do not pose an immediate threat to humanity — that threat may instead come from people armed with these technologies.
These technologies themselves do not make decisions in the sense that we understand the process. They are guided by algorithms, just as animals are driven by instincts, and this has nothing to do with the intentional mind.
But we can expect something much more powerful to emerge that we cannot control in the sense of directive regulation. That something could have an autonomous mind, and it would emerge not spontaneously, as a result of some unforeseen coincidence or due to an oversight, but as the outcome of human intention.
In fact, people need more from technology than Narrow AI tools can offer.
Humans believe that they can acquire something that will enable them to cope with the problems they are powerless to solve on their own because of the depressing limitations of their reasoning. This weakness stems from the contradiction of human nature, within which reason, empathy, and spirituality are forced to share space with selfishness, greed, insensitivity, and cruelty.
This conflict never subsides. It manifests itself in our relationships on an individual and collective level. It never ceases to plague us. Humanity’s greatest minds have attempted to solve this problem throughout history, with only relative and unreliable success. Our hope is that our human mind is not doomed to struggle with this issue alone. Probably, we can create an advanced incarnation of it. Something that is much more intelligent and wiser than either a set of Narrow AI algorithms or the paleolithic psyche of Homo Sapiens. Some kind of synthetic consciousness, as close as possible to the content of what some philosophers have called Objective Reason.
In this way, humans need a means of ultimate self-transformation, their own idealized replica, with power unlimited by nothing but morality, constituted by recognizing the primacy of human universals. That is, they need a true Reason, not an imitation of it; a mind ontologically identical to their own, creative, spiritualized, and free from the evolutionary heritage of the species Homo Sapiens irrelevant to human values.
I believe this means the “final invention” I mentioned above but in the most positive sense.
To some, its appearance may seem a long shot or impossible because of some allegedly fundamental limitations imposed on implementing artificial consciousness. Strictly speaking, such limitations cannot be ruled out, but this is a separate topic that requires a philosophical analysis of what should be considered “consciousness,” “intelligence,” “reason,” “mind,” “sameness,” etc. In any case, the lack of an objective understanding of these distinctions does not mean that AI cannot be truly intelligent.
Moreover, there is currently a theory that represents its design based on the principles of the human brain. Thus, the key factor in implementing this type of intelligence (in contrast to Narrow/Weak AI called General/Strong AI, respectively) may be humans’ belief that it is possible and their desire to obtain it.
Desire and possibility go hand in hand, but a little before desire, there is anticipation. Or, perhaps, an insight — call it what you like; in a word, that familiar to everyone feeling, the very occurrence of which, however fleeting, can predetermine the future. Do you remember? Ten years ago, the movie “Her” came out, and anyone who has seen it and has now tried ChatGPT has almost caught himself thinking, “it has already happened.” And we are approaching the point where the current generation will have to answer the questions raised in this movie and others, which we yet cannot imagine. From a historical perspective, this will probably happen very soon, maybe tomorrow.
But not today; we have time to ponder it.
Read a related topic: