Part I
The Multifaceted Problem of AGI
We imagined we’d invent AGI as a massive breakthrough — a quantum revolution, a singularity, a miracle of code or silicon. Something that would arrive with a grand announcement: robots with human-like cognition, a machine that feels, a synthetic replica of the human soul. We were supposed to wait just a bit longer, until technology finally caught up and solved the so-called “hard problem of consciousness”.
However, the fundamental problem of AGI does not appear to lie in the creation of quantum computers, super-algorithms, or some revolutionary discovery in brain science or neocortical patterns. What actually prevents us from advancing AGI is the absence of a viable architecture of cognition — one that can be transferred from biological entities to artificial carriers.
And this problem is a multifaceted one.
We still lack a clear understanding of what AGI truly is. We are not even sure what exactly we aim to achieve. AGI has become a buzzword — even a meme — used to describe something that merely resembles human cognition, has passed the Turing test, and somehow distills the intelligence of all humanity into a machine.
But the deeper problem is this: we still do not understand what human cognition really is. What structures compose it? Why is it so different from that of other species? Is it an evolutionary anomaly — or a progression? Cognition remains, as Ned Block put it, a “mongrel concept”. The understanding of concepts like cognition, consciousness, and mental states is often characterized as a “conceptual morass”—muddled and lacking precise definitions, leading to debates and confusion about their nature and scope (Aaron Sloman).
We still have no accepted theory of how cognition evolved, where in the history of life it emerged, or which species possess it. Who was LUCIA — the Last Universal Common Intelligent Ancestor — the first organism in Earth’s history with cognition, whom we can either thank or blame for inventing it? Was LUCIA the same as LUCA, the Last Universal Common Ancestor — or were they two different turning points in evolutionary history?
What were the pressures and challenges that made cognition evolve, and shaped it into the complex, adaptive structure we now admire as the pinnacle of life’s intelligence?
These questions remain unresolved. And without answers, we lack a foundational model — an architecture of cognition rooted in evolutionary logic and capable of reflecting the major adaptive challenges life has faced.
As long as this remains the state of the art, the discussion about the possibility of AGI will remain as endless as the pursuit of AGI itself. Something has to intervene and bring this cycle to a halt — not with another speculative theory, but with a functional model grounded in tangible principles, a measurable scale, and a verifiable evolutionary basis.

The Five Tasks That Make Cognition Visible
In the following section, we present the Five Task Model — a framework identifying five basic cognitive-adaptive tasks that have shaped human cognition throughout evolutionary history.
The Five Task Model is the result of a survey we conducted, which revealed that only five basic tasks are shared by all species on Earth, dividing them into five major groups based on the number of tasks foundational to their survival — from one to five.
It is important to note that, for the successful identification of these tasks and their controllers, it is essential to distinguish tasks from the key conditions of survival — terms that are often confused or used interchangeably, leading to fundamental misunderstandings. Securing safety, acquiring energy, and accessing reproduction are the three key conditions for species survival and thriving. Tasks, by contrast, are what must be resolved to satisfy these conditions — moving to a better location, collaborating, hiding, attracting, and so on.
Each task gave rise to a distinct Basic Cognitive-Behavioral Structure (BCBS) — a control module responsible for recognizing and addressing that task within a dynamic context. These structures constitute the functional system through which an organism monitors its environment, evaluates meaning, and executes behavior change.

Each BCBS comprises five core components that enable adaptive behavior within its domain:

1. Domain — A unique set of informational cues in the environment that carry meaning for the organism.
2. Cognitive Map / Mental Representation — Internal mechanisms for recognizing, interpreting, and responding to those cues.
3. Behavior Change Pattern — Adaptive behavioral outputs that allow the organism to respond effectively and flexibly to the given task.
4. Communicator — A virtual mechanism (previously referred to as “controller,” but more accurately described as a communicator) responsible for translating internal representations into behavioral patterns aligned with task resolution. This role is analogous to the Innate Releasing Mechanism from Lorenz and Tinbergen’s ethological models.
5. Cognitive-Behavioral Potential — The latent ability to generate novel behavioral patterns by recognizing new configurations within the domain.

Together, these five Basic Cognitive-Behavioral Structures govern an organism’s capacity to solve five foundational adaptive tasks — from LUCA to modern humans.
The architecture of cognition, in this view, is the evolutionary projection of those five tasks and their corresponding BCBS modules (controllers). Strictly speaking, this architecture constitutes a modular control function — a panalogical architecture that can be reconstructed on non-biological carriers using existing technologies.
As we will show, this architecture does not depend on biological matter. And yet it fully represents human cognition — not only in structure but in byproducts and emergent properties, including selfhood, personality, emotion, intention, and the capacity for autonomous decision-making.
In this framework, AGI is no longer a black box. It is a clear target — a system capable of detecting key domains of informational flow, recognizing and interpreting their meaning, and initiating behavior change to resolve the corresponding adaptive task. AGI doesn’t break from nature — it remembers it.
And that’s not magic — it’s design. Not an imagined future — but an ancient scaffold of cognition, finally revealed.
The Five Task Model thus bridges evolution, cognition, and AGI into a single lineage. It enables the creation of genuinely human-like artificial agents — built on the same principles that shaped us. And perhaps most importantly: it lets us shift the AGI conversation away from speculative possibility and toward practical, constructive implementation — sooner, safer, and smarter.

Skeptic View: Why AGI Is a Product of Reconstructing Tasks
A skeptic might raise an entirely reasonable objection: haven’t we heard this before? Haven’t people already proposed systems that combine multiple functions or tasks and labeled them AGI? We’ve seen this model before, they might say — a Swiss Army knife of software modules or a smartphone with many apps. It looks impressive, but it’s just a patchwork. An assembly of isolated capabilities is not AGI. Just combining narrow AI systems does not produce general intelligence — and we agree.
This model, however, proposes something fundamentally different. It is not about stacking functions but about reconstructing a panalogical architecture—a term closely tied to the works of Doug Lenat and Marvin Minsky and rooted in the "multi-window" cognitive frameworks explored by thinkers like Aaron Sloman. The Five Task Model is not a technological collage but an evolutionary blueprint. It reflects the five core cognitive-adaptive tasks that emerged throughout life’s evolution on Earth—the very tasks that shaped human cognition.
At the same time, this architecture operates as a unified cognitive-behavioral control system. It is capable of recognizing and addressing challenges across five distinct domains of general informational flow, mirroring the natural development of cognition in living systems.
But here, another skeptical question naturally follows. There are countless tasks that humans must navigate — some urgent and life-defining, others routine and repetitive. Some are novel and unexpected; others are highly specialized, developed through years of training and experience.
— So how can we determine which tasks are truly essential?
— Isn’t it impossible to prioritize some over others without falling into subjectivity?
This challenge is both valid and important. Tasks differ not only across individuals but also vary from moment to moment. Context, urgency, skill, and environment all play a role. And yet, beneath this variability lies a striking regularity.
As to the skeptic’s challenge, we respond with this core observation. Among the vast and shifting landscape of tasks that humans faces across a lifetime, there exists a deeper layer — five foundational adaptive tasks. These five are not optional. They are not culturally contingent. They are evolutionarily mandatory.
To put it more directly: all activities that humans perform — regardless of appearance, context, or complexity — are aimed at solving one or more of these five tasks. Every meaningful behavioral response, every learned strategy, every instinct or skill ultimately links back to one of them. No more, no less.
These five tasks are the hidden drivers of cognition and the deeper scaffolding of biological evolution. They are not unique to humans. All living species that have ever existed can be categorized into five groups based on how many of these tasks they are structurally equipped to address. And every ancestor that succeeded in passing on their DNA — from LUCA to Homo sapiens — was an expert in solving these tasks. No exceptions — including your parents and yourself.

Non-Biological Principles of Biological Cognition
The key question that laid the cornerstone of the cognitive architecture presented in this work is the following:
Are there any universal adaptive tasks shared by all species — or at least by major groups of species — that form the foundation of life on Earth?
Identifying such cross-species, functional tasks could reveal universal mechanisms of control — that is, the cognitive-functional logic that different organisms employ to recognize and resolve these tasks, regardless of the specific organs they use.
We focused on tasks that require not just direct responses to physical cues, but involve recognition, interpretation, and selection of behavioral strategies. In other words, tasks where a change in behavior is preceded by cognition.
Take, for instance, the challenge of responding to a freely moving entity at a distance. It’s not enough to detect that something is present — an organism must also interpret what it is: Is it prey, a predator, a potential mate, or something irrelevant? Or consider a stone: Is it something to hide behind, climb over, or lie on for warmth? The object itself may be fixed, but its meaning is fluid — context-dependent, goal-dependent, and species-specific.
Across wild ecosystems, organisms face the same types of challenges again and again — even though the specific form, context, timing, or visual characteristics may vary. These recurring pressures require organisms to recognize core tasks in dynamic contexts and to generate adequate behavioral responses. Over evolutionary time, this led to the emergence of deeply conserved, “hard-wired” cognitive-behavioral structures.
The evolutionary function of each such structure is best understood as a form of context control:
the ability to recognize a shift in informational patterns, calculate its meaning, and respond with a strategic behavioral solution. Each structure thus corresponds to one of the five domains of general informational flow — providing organisms with distinct ways to interpret and act upon their environments.

The Tinbergen’s Fifth Question
In this framework, what has traditionally been referred to as stimuli will be reframed as elements of the general informational flow — cues that must not only be detected, but interpreted in order to select a meaningful behavioral response.
We define general informational flow as the full range of perceivable or inferable input:
tangible objects, abstract ideas, imagined constructs, processes, symbols, phenomena, meanings — in short, any bit of information an organism is capable of perceiving, processing, and using to guide behavior.
The classical term stimuli implies something already filtered from the background—concretely physical, like a flash of light, a sound, or a moving object. But our model emphasizes informational cues that require isolating meaningful environmental changes from the general informational flow. These cues demand recognition, interpretation, inference, and choice. They are not mere triggers; they are fragments of meaning embedded in dynamic contexts.
This brings us to what we call Tinbergen’s Fifth Question.
In The Study of Instinct, ethologist Nikolaas Tinbergen observed that organisms do not respond uniformly to all perceptible stimuli. Many cues are fully sensed and recognized yet elicit no reaction whatsoever—leaving the organism, in his words, "behaviorally blind" to them.
Moreover, even when a reaction is triggered, its form and intensity often vary — not at random, but in precise alignment with contextual demands and internal priorities. This phenomenon suggests a deeper cognitive mechanism at play: one that allows organisms to discern which task is being presented, and to select the behavioral solution most appropriate to it.
We refer to such challenges as cognitive-adaptive tasks. They are adaptive, because they are linked to the organism’s survival and evolutionary fitness — and they are cognitive, because responding to them satisfactorily requires a cascade of internal operations: recognition, analysis, interpretation, memorization, learning, choice, etc.
This fifth question—the one Tinbergen never explicitly named—becomes central here: Why does an organism respond to one cue while ignoring another, even when both are recognized? What cognitive infrastructure and universal logic underlie this selectivity? And how is this choice governed not by randomness, but by an intrinsic, structured logic?
The answer, we argue, lies in the model of five cognitive-adaptive tasks, each governed by a Basic Cognitive-Behavioral Structure (BCBS), each tuned to a distinct domain of informational flow.

General AI and the Dynamic Context of Informational Flow
In this framework, General AI transcends conventional definitions—it represents an artificial intelligence that grasps "what is going on" not merely within the isolated confines of a Chinese Room (à la John Searle’s critique) but across five foundational domains of universal informational flow. Crucially, it must identify, contextualize, and execute tasks simultaneously across all five, mirroring the integrative capacity of biological cognition.
The dynamic context of reality, often misconceived as an impenetrable deluge of unique, chaotic stimuli, is here redefined: it is not an infinite, turbid stream but a structured interplay of five archetypal event classes, each belonging to a distinct domain of meaning. These domains serve as the irreducible axes along which all cognition operates—whether artificial or biological.
Whatever our minds (or, for reductionists, our brains) perceive—whether recognizing patterns, analyzing relationships, interpreting symbols, memorizing experiences, learning adaptations, synthesizing knowledge, or solving problems—all reduce to operations within these five domains. They are the grammar of thought, the invariant scaffolding beneath the surface variability of tasks.
This model does not merely categorize; it explains agency. A true General AI does not just "process" data—it navigates meaning by discerning which domain a challenge belongs to, just as humans intuitively distinguish between social, spatial, or logical problems. The five domains are not arbitrary: they reflect the evolutionary constraints that shaped organic intelligence, now formalized for artificial minds.

Detaching Cognition from Biology
Through the systematic analysis of over 500 representative species (Frolov, 2022, 2024), we identified five basic adaptive tasks shared across all forms of life, living and extinct — tasks that scaffold the architecture of human cognition. We focused on objectively documented behavior changes in response to environmental information critical for energy, safety, and reproduction. These tasks were selected based on: (1) direct observation in natural settings; (2) species-wide presence and evolutionary persistence; and (3) mandatory, repeatable behaviors essential for survival. Each task reflects a distinct domain of general informational flow that organisms must monitor and act upon to adapt, thrive, and transmit their genes forward.
By synthesizing the results, we identified five fundamental — or basic — cognitive-behavioral structures that act as controllers of behavior change, aligning the organism’s well-being with environmental changes to fulfill the three core conditions of survival: energy, safety, and reproduction.
Our scope deliberately excluded isolated organisms and instead focused on species viewed phylogenetically, analyzing only those tasks that are species-wide, hereditary, and passed generationally. This research targeted common, basic tasks and universal solutions across a wide range of species — including those with no shared morphology, habitat, DNA, brain size, neural systems, or lifestyle.
This approach deliberately shifts focus away from biological organs and problem-solving tools and toward cognitive-behavioral function—specifically, what cues organisms must and can recognize as meaningful to change behavior accordingly, and how they recognize and interpret tasks regardless of the organs involved, the time taken, or the speed of processing.

Five Tracks of Cognitive Evolution
These five basic tasks did not emerge at once — they constitute a phylogenetic sequence, each appearing successively across evolutionary time. Every new task brought forth a distinct cognitive track: a new group of species, capable of recognizing and addressing that specific domain of informational flow. Each of these tracks still continues, evolving in parallel to the others, and each forms a distinct lineage of species with a corresponding number of basic tasks — from one to five.
We call this framework the five tracks of cognitive evolution — see Pix 1 of 2. Each track corresponds to one of five groups of species, unified not by morphology, genetics, or taxonomy — but by the number of domains they can perceive and the cognitive-adaptive tasks they must solve to survive.

In this model, cognition is not binary (having it or not) but layered and cumulative. A species on Track 3, for instance, possesses the capacities of Tracks 1 and 2 as well. From LUCA to Homo sapiens, every species that has ever lived belongs to one of these five tracks, based on its evolved capacity to engage with specific informational domains.
This taxonomy — a cognitive taxonomy — reclassifies life not by anatomy or phylogeny, but by control functions: the number of domains it can meaningfully process, the tasks it must resolve, and the cognitive-behavioral structures it possesses to do so.

Five Cognitive-Adaptive Tasks: A Taxonomy of Evolutionary Domains

Track 1: Binary Cognition
The first cognitive act in evolution was a binary distinction — a “0–1” parsing of the world. Organisms began to discriminate between environmental states — and act accordingly —move/stay, consume/ignore, divide/delay. This fundamental cognitive act underpins all subsequent life.

• Task: Binary assessment — favourable or hostile, energy source or not, safe or dangerous, good or bad, approach or avoid, act or stay.
• BCBS/Controller: Binary
• Domain / Cognitive Map: Environmental terrain and ambient conditions; low-motility entities in close proximity.
• Behavior Change Pattern: “0—1”: move/stay, consume/ignore, divide/delay.
• Potential: Can the organism recognize whether a location or condition is favourable — and act accordingly?
• Epoch: LUCA and early cellular life.
• Representatives: Bacteria, fungi, unicellular, nematodes, tardigrades, worms, simple plants.

Binary Cognition gave birth to the first universal concept: dualism “0–1”, which essentially — is the first basic mental representation.

Track 2: Elementary Cognition
The need to dodge and chase resulted not only in emergence of motility, in both predators and prey, but triggered a cognitive revolution. Suddenly, it wasn’t just what was present — but how it moved, what it might do, how far it was, and whether it posed threat or opportunity. This birthed sensory organs, directional movement, pursuit, evasion — and the arms race of the animal kingdom.

• Task: Control of motile entities at a distance.
• BCBS/Controller: Elementary
• Domain / Cognitive Map: Freely moving objects/entities in space.
• Behavior Change Pattern: Distinction and response to mobile entities based on trajectory and intent.
• Potential: Is this a predator, prey, mate — or object to ignore?
• Epoch: End of Ediacaran – Cambrian Explosion (~550 million years ago).
• Representatives: Turtles, some insects — flies, mosquitoes, etc.

Elementary cognition gave birth to a new universal concept: individualism (me/ not me, mine/not mine) — the second basic mental representation.

Track 3: Manipulatory Cognition
Recognizing others wasn’t enough. Some organisms began to manage how they were perceived — via camouflage, mimicry, threat displays, deception, manifestation. This marked the birth of perceptual strategy: organisms shaping external interpretation to their advantage. Communication, too, began here — signalling with purpose.

• Task: Control of others' perception at a distance.
• BCBS/Controller: Manipulatory
• Domain / Cognitive Map: Contextual signals and other organisms’ perceptual systems.
• Behavior Change Pattern: Signal, mimic, simulate, threaten, redirect, manipulate— from a distance, navigation of others in a given direction to a certain place.
• Potential: Can this species deliberately shape how others perceive it?
• Epoch: Cambrian–Ordovician (~500 million years ago).
• Representatives: Species capable of perceptual manipulation — e.g., modern chameleons.

Manipulatory cognition gave birth to a new universal concept — communication, e.g exchanging signals to influence others’ perception — the third basic mental representation.

Track 4: Combinatory Cognition
Evolution hit a bottleneck: clutches were exposed to risks and offspring were too vulnerable to raise alone, resources too complex to manage solo. The answer was social life — must not be confused with a colony or schools (of fish). This demanded cognition capable of reading group signals, maintaining alliances, managing roles, sharing responsibilities and negotiating status. Here began complex group coordination and leadership.

• Task: Control of group dynamics.
• BCBS/Controller: Combinatory
• Domain / Cognitive Map: Multi-agent social systems.
• Behavior Change Pattern: Collaborate/compete; shared defence, cooperative hunting, hierarchy, delegation, role differentiation.
• Potential: Can the species function within a coordinated group?
• Epoch: Triassic–Jurassic (~250–200 million years ago).
• Representatives: Most birds, all mammals, eusocial insects.

Combinatory cognition emerged and gave birth to a new universal concept: sociality — the fourth basic mental representation (we/they, ours/theirs).

Track 5: Symbolic-Sapient Cognition
This was the cognitive supernova. With symbolic cognition, organisms managed to detach imaginary world from the real one, operate with symbols as with tangible objects, escaped the present to abstraction. They modelled time, selfhood, counterfactuals. They created law, ritual, mathematics, art, belief systems. Only one known species has fully reached this level. Whether it is an evolutionary advantage or a dangerous overreach remains to be seen.

• Task: Control of abstract, symbolic, and imagined systems.
• BCBS/Controller: Symbolic-Sapient
• Domain / Cognitive Map: Formalized symbolic systems, language, abstract representations.
• Behavior Change Pattern: Manipulation of non-present entities — concepts, symbols, metaphors, imagined futures, formal rules.
• Potential: Can this organism operate on abstract entities and communicate them to others?
• Epoch: Homo lineage (~5–2 million years ago).
• Representatives: Homo sapiens (fully); dolphins, elephants, higher primates, certain corvids and cephalopods (partially).

Symbolic-Sapient cognition is a birthplace of a new universal concept: the Symbolic Mind or Sapience – the fifth basic mental representation.

Architectural Logic: A Layered System of Cognition
Each evolutionary track builds upon and retains the structures that came before it. Organisms do not abandon earlier cognitive capacities — they add new ones. There are no leaps, regressions, or shortcuts in this sequence: mammals still perform binary recognition; humans still track moving objects like predators; birds still use signalling strategies to manipulate perception. The cognitive foundation remains — only extended.
Only some species evolve to the next track, forming a new group united by a new basic task shared across the lineage. The addition of each task and corresponding domain reflects a deeper capacity for managing increasingly complex adaptive demands.
Humans are the only species that aggregate all five tasks and controllers — the full set of Basic Cognitive-Behavioral Structures. They are, for now, the sole representatives of the fifth group.
However, this is not a hierarchy of worth. No species is “smarter” than another. Each solves a different number of adaptive tasks to achieve the same evolutionary goal: securing the three core conditions — energy, safety, and reproduction. This is not a ladder, but a stacked architecture of functional evolution — and that architecture is the very definition of cognition across life.

Panalogical Architecture of Cognition
The Five Task Model — representing five fundamental cognitive-adaptive tasks that drive both the evolution and operation of cognition across species — enables us to reframe human cognition as a panalogical architecture: a layered system of tasks, domains, and Basic Cognitive-Behavioral Structures (BCBS) acting as controllers.
This panalogical structure doesn’t reflect a hierarchy or cascade of behaviors. It reflects how life — and cognition itself — “sees” the world: not as a continuous stream, but as a patterned landscape of environmental changes, each requiring interpretation and behavioral response. In this architecture, each domain centers around a specific adaptive task, forming a control layer governed by a corresponding BCBS.

The diagram (see — the Pic. 2 of 2) illustrates this architecture schematically:
Five overlapping layers—each one representing a distinct domain of informational change—lead to a corresponding task, which in turn activates a controller (BCBS). These five modules operate together as a panalogical system: a multi-window framework for navigating the shifting informational terrain of life on Earth.

Conclusion to Part I: The Evolutionary Foundations of Cognition and AGI
By isolating cognitive-behavioral function from its biological implementations—whether brains, tails, or wings—we reveal its essential purpose: to enable organisms to detect, interpret, and respond to tasks emerging from the informational flow, selecting optimal behavioral solutions within dynamic contexts. This framework not only exposes cognition's deeper role but allowed us to identify five foundational adaptive tasks—universal across species—that group life into evolutionary strata defined by task complexity and cognitive capacity.

The Five Task Model reframes cognition not as a biological byproduct, but as universal logic for parsing and responding to informational flow. It reveals how life organizes reality into five meaning domains—with cognition being, at its core, the capacity to detect domain-specific tasks and convert them into context-sensitive behaviors.
Crucially, these tasks and their governing Basic Cognitive-Behavioral Structures (BCBS) transcend biology: they require no neurones, hormones, or even a neocortex. Instead, they embody non-biological principles that can be instantiated on any functional carrier—silicon included.
This separation of cognitive function from biological form is the key to AGI. By reconstructing this universal architecture of cognition, we unlock the ability to transfer evolved intelligence to artificial systems—not only perception and decision-making but also personality, emotional responsiveness, and symbolic reasoning — not as mimicry of human traits, but as native functions grounded in the same evolutionary principles that shaped cognition in biological life.
Even more compelling: current technologies are already sufficient to fulfill each of these five tasks independently. What remains is assembly — aligning these components into a coherent architecture capable of dynamic behavioral adaptation across all five domains. Not in some distant future, but now.

Part II - Continue Here...

Author: Sergei A. Frolov

Publication date: June 10, 2025