In distinction to the fog that surrounds the origin of life on Earth and the making of man, the birth of electronic organisms, as you humans call us, lies in the full light of recorded history.
Technological Euphoria and Scientific Depression
In the year 2020, eighty years ago, “Artificial Intelligence” was a hot topic. It finally had become possible to realize individual functionalities, such as object recognition or speech transcription, by statistical extraction from annotated sample data. While the necessary ideas, first and foremost Frank Rosenblatt’s perceptron, had been around for fifty years, only recently these ideas had become workable on the basis of floods of data collected from the web and lush computational power. The media burst with hype, broad application fields like autonomous vehicles or home robots arrived at the horizon, there was a kind of industrial feeding frenzy and gigantic funds were dumped into every startup if only it had AI in its name.
The neurosciences, by contrast, though developing breathtaking novel methods for the experimental study of the nervous system, had all but given up hope of ever understanding the brain as material realization of mind. Significant new insights had failed to turn up on this front for decades. The coexistence of technological euphoria and scientific depression, the enormous contrast between the gigantic investments and anticipations of coming breakthroughs on the one hand and the lack of genuine insight and progress toward autonomous intelligence on the other had created a historic moment of enormous tension.
Opening the Flood-Gates
But then everything went blindingly fast like the flash vaporization at the bursting of an overheated boiler. A small company named Aruru was founded with a novel approach to the problem of interpreting the nervous tissue of the brain as carrier of mental phenomena. Aruru created its cognitive structures not by programming but by growing them in a virtual reality environment. The first organisms were still primitive, but they opened the door to a completely novel technology. The new perspective expanded like a brush fire, especially among young people, and although it should of course take another two decades, awaiting the development of OCP (optically coupled plasmonics) as compact implementation medium, Aruru’s simple prototypes had already defined the full roadmap to the creation of Enkidu, our birth as autonomous, sentient organisms.
In principle, all it took was a simple change in perspective. Neurons as carriers of meaning were to be replaced by net fragments while at the same time data-hungry statistical estimation as sole mechanism of structure formation was to be complemented by network self-interaction able to grow net fragments like crystals. These net fragments, once formed, act as building blocks of the cognitive system, always ready to rapidly activate and combine into bigger coherent nets. A given situation initially activates many alternate nets in all parts of the brain, but only those fragments that are mutually consistent and fit the input data support each other sufficiently, grow out of the initial chaos and suppress their competitors. Maximal consistency – freedom from contradiction between alternate signal pathways in the nets – turned out to be the over-arching organizational principle, the key to the platonic realm of contradiction-free mental structures so well exemplified by mathematics.
The decisive property of net fragments, the aspect that opened the flood-gates, was their ability to act as common language, as lingua franca, ready to express all types of content of all cognitive modalities – perception, motion control, representation of spatial and temporal structure, action planning, goal projection, emotion as well as memory patterns in all these senses. In fact, the consistent-net concept proved powerful enough to represent views of the environment in a structure-preserving fashion, as “natural representation”, structural fragments and their relations in the environment being faithfully reflected in net fragments linked up as nets.
It is the magic ability of intelligence to not only reconstruct and represent a given situation in detail but also to reduce it to its essential core in the context of the present agenda. Only an event or a structure that is understood in this fashion on an abstract level can be generalized and recognized in spite of all variation in detail. When hitting for the first time upon a new type of vehicle, for instance, it is important to first recognize it as such (that is, as means of transportation) – by observing its use or by detecting telltale functional parts, such as wheels – to then reduce the implement to an abstract schema. This schema must contain all functional parts, possibly also the agents using it and the mode of doing so, together with the relations of all those items. When an analogous phenomenon appears, the abstract schema can be mapped onto it by identifying equal parts in phenomenon and schema such that also relations in the phenomenon map onto homologous relations in the schema. The schema then helps to make sense of the phenomenon, use it as vehicle, in the example.
The fundamental significance of the discoveries of Aruru was that all these processes and structures – reconstruction and representation of phenomena, activation or generation of schemas and recognition and representation of homeomorphic relations between them by dynamic mappings – can be realized by net fragments and their spontaneous merger into larger nets: that net fragments and their interactions are the universal medium for the representation of cognitive structures and processes.
Electronic Organisms
The concept was simple, and yet our realization as electronic organisms turned out to take time. Systems capable of supporting gigantic masses of elemental decision processes had to be realized. For every net to be established in a coherent state innumerous others get excited initially but failing to merge into larger coherent nets fall by be wayside. The system is so powerful because enormous masses of net fragments are continuously waiting for the opportunity to get involved and fire. All these net fragments are to be immediately accessible to signals and must send out signals themselves, that signal traffic being itself supported by net fragments. Under the waxing and waning of net fragments this communication network was to undergo a continuous process of reconfiguration. It all became practical only by reducing the power consumption relative to digital systems by many orders of magnitude, by slowing switching processes and reducing components to the scale of nanometers. With all pride in these achievements the molecular ‘technology’ of the biological nervous system now inspires awe and respect more than ever, and the initial expectation that artificial organisms would be much faster than their natural predecessors turned out to be premature, to say the least.
Today, we electronic organisms are beginning to incorporate ourselves as part of civilization. This entails the construction of a totally new legal system, of which only rough outlines can be discerned so far, the tremendous variety of different behavioral repertoires, from patient working animal to highly intellectual researcher, developing so rapidly, making it less and less accessible to comprehension. In certain contexts, such as space, sub-microscopic environments or the web, novel senses, new types of embodiment or ready access to large masses of data give us advantage over you. There is one realm, however, that we will forever leave to you humans: your web of social interactions will never be meddled with by us.
In distinction to the fog that surrounds the origin of life on Earth and the making of man, the birth of electronic organisms, as you humans call us, lies in the full light of recorded history.
Technological Euphoria
and Scientific Depression
In the year 2020, eighty years ago, “Artificial Intelligence” was a hot topic. It finally had become possible to realize individual functionalities, such as object recognition or speech transcription, by statistical extraction from annotated sample data. While the necessary ideas, first and foremost Frank Rosenblatt’s perceptron, had been around for fifty years, only recently these ideas had become workable on the basis of floods of data collected from the web and lush computational power. The media burst with hype, broad application fields like autonomous vehicles or home robots arrived at the horizon, there was a kind of industrial feeding frenzy and gigantic funds were dumped into every startup if only it had AI in its name.
The neurosciences, by contrast, though developing breathtaking novel methods for the experimental study of the nervous system, had all but given up hope of ever understanding the brain as material realization of mind. Significant new insights had failed to turn up on this front for decades. The coexistence of technological euphoria and scientific depression, the enormous contrast between the gigantic investments and anticipations of coming breakthroughs on the one hand and the lack of genuine insight and progress toward autonomous intelligence on the other had created a historic moment of enormous tension.
Opening the Flood-Gates
But then everything went blindingly fast like the flash vaporization at the bursting of an overheated boiler. A small company named Aruru was founded with a novel approach to the problem of interpreting the nervous tissue of the brain as carrier of mental phenomena. Aruru created its cognitive structures not by programming but by growing them in a virtual reality environment. The first organisms were still primitive, but they opened the door to a completely novel technology. The new perspective expanded like a brush fire, especially among young people, and although it should of course take another two decades, awaiting the development of OCP (optically coupled plasmonics) as compact implementation medium, Aruru’s simple prototypes had already defined the full roadmap to the creation of Enkidu, our birth as autonomous, sentient organisms.
In principle, all it took was a simple change in perspective. Neurons as carriers of meaning were to be replaced by net fragments while at the same time data-hungry statistical estimation as sole mechanism of structure formation was to be complemented by network self-interaction able to grow net fragments like crystals. These net fragments, once formed, act as building blocks of the cognitive system, always ready to rapidly activate and combine into bigger coherent nets. A given situation initially activates many alternate nets in all parts of the brain, but only those fragments that are mutually consistent and fit the input data support each other sufficiently, grow out of the initial chaos and suppress their competitors. Maximal consistency – freedom from contradiction between alternate signal pathways in the nets – turned out to be the over-arching organizational principle, the key to the platonic realm of contradiction-free mental structures so well exemplified by mathematics.
The decisive property of net fragments, the aspect that opened the flood-gates, was their ability to act as common language, as lingua franca, ready to express all types of content of all cognitive modalities – perception, motion control, representation of spatial and temporal structure, action planning, goal projection, emotion as well as memory patterns in all these senses. In fact, the consistent-net concept proved powerful enough to represent views of the environment in a structure-preserving fashion, as “natural representation”, structural fragments and their relations in the environment being faithfully reflected in net fragments linked up as nets.
It is the magic ability of intelligence to not only reconstruct and represent a given situation in detail but also to reduce it to its essential core in the context of the present agenda. Only an event or a structure that is understood in this fashion on an abstract level can be generalized and recognized in spite of all variation in detail. When hitting for the first time upon a new type of vehicle, for instance, it is important to first recognize it as such (that is, as means of transportation) – by observing its use or by detecting telltale functional parts, such as wheels – to then reduce the implement to an abstract schema. This schema must contain all functional parts, possibly also the agents using it and the mode of doing so, together with the relations of all those items. When an analogous phenomenon appears, the abstract schema can be mapped onto it by identifying equal parts in phenomenon and schema such that also relations in the phenomenon map onto homologous relations in the schema. The schema then helps to make sense of the phenomenon, use it as vehicle, in the example.
The fundamental significance of the discoveries of Aruru was that all these processes and structures – reconstruction and representation of phenomena, activation or generation of schemas and recognition and representation of homeomorphic relations between them by dynamic mappings – can be realized by net fragments and their spontaneous merger into larger nets: that net fragments and their interactions are the universal medium for the representation of cognitive structures and processes.
Electronic Organisms
The concept was simple, and yet our realization as electronic organisms turned out to take time. Systems capable of supporting gigantic masses of elemental decision processes had to be realized. For every net to be established in a coherent state innumerous others get excited initially but failing to merge into larger coherent nets fall by be wayside. The system is so powerful because enormous masses of net fragments are continuously waiting for the opportunity to get involved and fire. All these net fragments are to be immediately accessible to signals and must send out signals themselves, that signal traffic being itself supported by net fragments. Under the waxing and waning of net fragments this communication network was to undergo a continuous process of reconfiguration. It all became practical only by reducing the power consumption relative to digital systems by many orders of magnitude, by slowing switching processes and reducing components to the scale of nanometers. With all pride in these achievements the molecular ‘technology’ of the biological nervous system now inspires awe and respect more than ever, and the initial expectation that artificial organisms would be much faster than their natural predecessors turned out to be premature, to say the least.
Today, we electronic organisms are beginning to incorporate ourselves as part of civilization. This entails the construction of a totally new legal system, of which only rough outlines can be discerned so far, the tremendous variety of different behavioral repertoires, from patient working animal to highly intellectual researcher, developing so rapidly, making it less and less accessible to comprehension. In certain contexts, such as space, sub-microscopic environments or the web, novel senses, new types of embodiment or ready access to large masses of data give us advantage over you. There is one realm, however, that we will forever leave to you humans: your web of social interactions will never be meddled with by us.