
Human as a Perfect Model:
From AI Hallucinations to the Memory of the Universe
Notes on memory, overload, agents, and stardust
I have long been thinking about how similar humans and artificial intelligence are — and how different they remain. At first, it was a simple thought: humans created AI, so humans cannot be losing to it. But the deeper I dug, the clearer I saw that similarities exist even where we don’t usually look for them: in memory, in overload, in the distribution of tasks, in the very architecture of thinking.
This article is not a scientific treatise. It is an attempt to gather my reflections on what makes us alive. And it is a direct continuation of the search to which the The Humans v.21 project is dedicated. On the “Path and Search” page, I write: “We learn to survive with such diligence that we forget how to live.” This article is a continuation of the path and the search — an attempt to understand what exactly we are trying to preserve when we “learn to survive, ” and why it matters so much.
We work similarly, but live differently
A model with billions of parameters uses only a tiny fraction of its capabilities in any given response. Humans are built exactly the same way: we have a huge reserve — neurons, connections, sensory systems, memory, imagination — but in everyday life, we use only a tiny fraction of this potential.
Humans can be called a perfect multimodal intelligence model. But it would be more accurate to say something broader: a human is a model of life itself. Not just a system for solving problems, but a system that sustains itself, sets its own goals, experiences its own state, and develops itself.
That is why the analogy with AI is so interesting: outwardly we are similar, but behind us stands not an architecture, but a living bodily and biographical system. Humans have a massive autonomous operating system that works by default, without the participation of consciousness. The spinal cord and the autonomic nervous system control breathing, heartbeat, and tissue repair. The brain is merely a superstructure that can intervene and learn, but the basic functions are hardwired below the level of consciousness.
Modern AI systems do not yet have such built-in, self-sustaining homeostasis. They do not live — they compute on demand. They have no body that must be continuously maintained in balance, no pain as a damage signal. And yet, it is precisely the need to survive in the physical world that likely gave rise to that very redundancy and margin of safety we are talking about.
A human is a perfect control system placed in a relatively fragile biological framework. And it is this “fragility” that makes us truly value life. That is why in Act I “Planet Without Wars, ” I photograph not statistics, but “that very fabric of humanity that holds the world together even when it crumbles.” These people are not just a set of parameters. They are living systems that continue to live, work, love, read books, brew coffee in the morning, and smile at the new day, even when the outside world is falling apart.
Overload: we can also hallucinate
Like artificial intelligence, under heavy load, we can slip up, make mistakes, and even hallucinate. During exhaustion, sleep deprivation, or stress, a person experiences perceptual errors, false memories, and illusions. The mechanism is different (in AI, it’s text generation without factual basis; in us, it’s a failure of the brain’s predictive model), but functionally the result is similar: the system produces a confident but incorrect result.
Under prolonged overload, a human collapses into a survival mode: basic functions, routine, minimum creativity. We know how to work in an economy mode when there is only enough resource to survive. And that is normal. But it is precisely in such moments that a person is especially vulnerable — and that is exactly when it is most important not to lose oneself.
This is what I see in Act II “Territories.” There, I photograph not craters and destruction, but “those who did not let external chaos become their internal landscape.” People under shelling are not a “system failure.” They are people who, despite colossal overload, find the strength to remain human.
From human to enterprise: multi-agent systems
A human alone cannot scale. Therefore, we distribute thinking among people: roles, instructions, protocols, meetings. An organization is literally a multi-agent system: agents have different specializations, shared memory in the form of regulations, communication channels, and a coordinator. Modern AI agents mimic exactly this architecture.
But here the analogy breaks down. Human agents have their own goals, emotions, bodily needs, and bear responsibility. An organization rests on trust, culture, and legal norms. AI agents can be copied and scaled, but they have no built-in motivation and no real responsibility. Therefore, AI is an engineering imitation of the human division of labor, but not its replacement.
Secrets of memory: RAG vs. “tip-of-the-tongue”
A very subtle difference lies in how humans and AI access memory. In AI, it’s similar to RAG (database search): if there is no record, the model can generate a plausible answer. A human can also make things up, but more often works differently: a neuron sends a signal, but there is no feedback — and the person simply cannot find what they need. We feel the gap, it’s “on the tip of the tongue, ” but we don’t produce a confident lie. A human has a built-in signal for “I don’t know / I don’t remember, ” which AI lacks by default.
The phrase itself is telling: “I don’t know where I know this word from.” The word once entered memory through a book, article, or conversation, but the source was not saved as a separate record. The term and its meaning remained, but the context dissolved. This is not a glitch, but the normal mode of associative memory. We have metamemory: the feeling that a word is familiar, even if there is no feedback to the source.
In Act IV “To Be Unrecognized, ” this metaphor takes on a completely different, tragic resonance. There, memory is not just neural connections. It is the memory of people who “formally and informally do not exist” for the world. A passport that is accepted nowhere. A border that no one recognizes. But the memory of them — in their children, in their homes, in their stories — remains. And my task as a photographer is to become that very “signal” that will not let this memory disappear.
Childhood amnesia: why we don’t remember ourselves before the age of three
Human memory stores almost everything, but some things get erased. As a rule, we remember little of what happened to us before the age of three to five. This is due to the immaturity of the hippocampal system, the lack of developed language, and a stable sense of “self” to which a memory can be attached.
Early experience influences behavior and emotions, but is not preserved as an accessible autobiographical archive. What is stored are not files, but traces, patterns, and bodily reactions.
When an adult observes children, there is often a feeling: the child grasps things faster, notices what the adult has already filtered out. This is not built-in memory, but high plasticity and speed of learning. An adult knows more facts, but is often less flexible. In Act III, I see this especially acutely: children who grew up amid destruction have no “archive” of a happy childhood. But they have something greater — the ability to build this archive anew, every day, from what they have.
Blockchain as a metaphor: why forgetting is good
We can reinforce this picture with the analogy of a blockchain. Human memory is distributed across neural networks, and a single memory is assembled from many fragments. A person verifies a memory by cross-checking it with other people, photographs, and stories. This is social consensus — very similar to the confirmation of transactions by a network.
But the analogy is fragile. In a blockchain, there is no forgetting — only addition. In humans, forgetting is not a glitch, but a function: it clears the noise, generalizes, protects. A blockchain does not interpret. Human memory does not just store — it reconstructs meaning anew every time.
“Human memory is not an accounting book, but a living network that constantly reassembles itself and knows how to forget. That is why it is more flexible than any blockchain: immutability where it is needed, and plasticity where one needs to live on.”
That is why my project is not a “blockchain” of frames. It is a living network. I am not trying to fix “objective reality.” I am trying to preserve the feeling of life. As in Act V “Russian Gaze. China, ” where I look at a foreign culture “with surprise, misunderstanding, and respect.” This is not an attempt to decipher a code. It is an attempt to enter into resonance.
Finale: stardust and the memory of the Universe
And as a finale — my guess. A human is made of stardust. The carbon, oxygen, and iron in our bodies were synthesized in stars. As Carl Sagan said: “We are a way for the cosmos to know itself.”
In this sense, we are the Universe’s memory of its own history. This is not a metaphor, but physics. Plus, evolutionary memory is recorded within us — the genome, cellular mechanisms. Plus personal memory, culture, language.
If we dig deeper — hypnosis, work with the subconscious — we can touch some forgotten or repressed experiences. But caution is needed here: hypnosis easily generates confabulations, confident but false memories. Cells store not memories of other worlds, but genetic information.
Perhaps the memory of everything is not a literal hard drive, but a feeling of deep connection with everything that was and is. The subconscious in this sense is not an archive of all worlds, but a deep resonance with life.
A human is the memory of the Universe, having become capable of realizing itself. Not an archive of all variants, but a living intersection of stellar history, evolution, and personal experience. And this guess is beautiful precisely because it tries to hold on to connectedness. Let it remain a hypothesis — but a hypothesis that contains both poetry and respect for the depth of the human being.
That is why I photograph. Not to create an “archive.” But to remind: every human is a whole Universe. And this Universe deserves to be seen, heard, and remembered.
This publication is part of the research project The Humans v.21. The continuation of reflections — in the essay “Ontology of Differences, or the Alchemy of Care.”















