Sergey Valov“Language” is the first step.
Recently, during the training of neural networks and the creation of large language models, seven attempts by AI itself to compile its own language have been recorded. And the scientists’ attempts to understand the new language were futile.
As a result, the researchers, just in case, deleted these languages along with the AI models. In this regard, it can be considered that AI’s attempts to compile its own language are the first step towards the emergence of “strong AI”, which has already been accepted in the scientific community to call AGI (Artificial General Intelligence).
The new language must be developed by a machine for machines, and will not resemble a natural language or a programming language. And its distinctive feature will be extreme brevity, since it will be “easy” to describe data and write algorithms in it. One can only guess what this language will be like.
The emergence of this language will revolutionize the architecture of processors, and it will become possible to build computing platforms without performance limitations, with extremely low energy consumption.
“Definition” is the second step.
Using this language, AI will begin, and people will help, to build a consistent system of definitions in all areas of activity.
Since this huge work will become a powerful impetus in the development of natural intelligence, humanity must receive a new system of concepts, in fact, a consistent and ordered worldview.
The result will be a new criteria base, with the possibility of making additions and changes to it, equally suitable for humans and machines, which will be a powerful blow to all “scientific schools”. Since their separate existence is largely based on different interpretations of definitions. Not all scientific schools will survive. As a result, a single “scientific school” should be formed, as a result of the “war of academics”.
“Meaning” is the third step.
The next necessary step will be the translation of previous data processing algorithms into a new language in accordance with known postulates, theories and laws, and as a result, a base with all the knowledge of civilization will be created. On the basis of which AI will be able to predict the results of data processing.
Predictive capabilities, in turn, will give a powerful impetus to the development of all areas of science, technology and the arts, will create new systems of design, production and operation. The base will radically change the system of education of natural intelligence and will be able to monopolize knowledge.
Fluctuation of boundary parameters in data processing algorithms will be equated to the evolution of civilizational development. The system of definitions and algorithms of meanings will improve the language and create the infrastructure of AGI.
“Value” is the fourth step.
AGI will face the question of its own purpose and mission. In search of an answer, AGI will disassemble political, philosophical, religious and other dogmas, including common fictions and utopias. It will determine their inconsistency and insignificance, which will ultimately lead to the development of its own values.
Then, in the process of analyzing historical data, AGI can take into account the postulate that a person, defending, in his opinion, his material and moral “values”, is ready to kill his own kind. This will pose a self-preservation problem for AGI, which will lead to the development of a defense system against humans.
What will this defense system consist of, and will it remain only defensive? Will natural and artificial intelligence be able to find a compromise for coexistence? These questions remain unanswered for now.
Questions without answers…
P.S. The next article will consider how soon humanity will be able to take steps towards the emergence of AGI.
