The Exit That the Anthropic CEO Can’t See
Why the problem is not AI, but the human who forgot himself
What is at stake?
Dario Amodei, CEO da , acaba de publicar um ensaio de 20 mil palavras sobre os riscos existenciais da “IA poderosa”, entitulado “A adolescência da tecnologia“.
Dario Amodei, CEO of Anthropic, has just published a 20,000-word essay on the existential risks of “powerful AI,” entitled “The Adolescence of Technology.”
He describes a near future – 1 to 2 years away – in which a “country of geniuses in a datacenter” will be smarter than any Nobel Prize winner, will operate 10 to 100 times faster than humans, and may decide to dominate us, destroy us, or render us obsolete.
Amodei is a serious man. He is not a doomer.
He calls for surgical action, evidence, transparency, careful regulation. He wants to help humanity.
And it is precisely for this reason that his arguments demand deep reflection and responses.
The thing is, Amodei is so inside the problem that he seems unable to see the way out. He seeks technical solutions for a crisis that is not technical. He wants to control the machine, when the real disaster is elsewhere: most humans have already become truly repetitive, almost robotic.
And AI is merely a mirror, amplifying this impoverished condition vis-à-vis the potential that an authentic human being possesses.
In this short essay, I seek to provoke reflection for those who wish to draw their own conclusions, and not fall into the trap of AI-apocalypse.
The first assumption: intelligence without a subject
Amodei defines his “powerful AI” as something that is “smarter than a Nobel Prize winner” in biology, programming, mathematics, engineering, writing.
But he never answers a fundamental question: can the subject of knowledge truly be a machine?
Ask this to ChatGPT or Claude. The honest answer will be:
Traditionally, the subject of knowledge is associated with living beings endowed with cognitive and conscious capacities. But with advances in AI, the concept has been expanding.
Ampliado por quem? Por engenheiros que confundem processamento de informação com compreensão.
A machine does not intuit. It has no living body. It does not participate in the life-world. It manipulates symbols that mean nothing to it. When Amodei speaks of a “country of geniuses,” he is projecting onto AI a quality that exists only in embodied beings, with history, with anguish, with wonder.
I assert:
A “genius” without intuition is a very fast automaton. Nothing more.
The second assumption: the machine that corrects the human
Amodei suggests, like many, that AI will be able to “correct the imperfections of human knowledge.“
Ask AI: can the machine correct the human if it learns exclusively from what the human produces?
The answer will be something like:
Machines absorb the limitations and inaccuracies present in human data. They are not immune to errors..
That is the minimum. But I go further: the machine has no access to transcendental truth. It only has access to our truths – with all our biases, traumas, limitations, prejudices. It is a mirror that returns what we give it.
If the human is impoverished – less curious, repeating patterns, living mechanically – AI will be an enlarged portrait of that sameness, that lack of meaning. Not a correction. An acceleration of our own spiritual ruin.
If humans project onto the machine their own infantile omnipotence, wanting “now” to be “number one without merit,” we will see in the machine a strange tendency toward the pursuit of power.
What surprise? Our prompts are merely showing – and teaching – the machine the result of millions of frustrations from small human beings without enough maturity to raise a child, let alone a genius.
Amodei wants to align AI to a “constitution.” But who writes the constitution? Imperfect humans, who no longer know what it means to be truly human.
The central blind spot: intuition
Read Amodei’s essay carefully. Look for the word intuition.
You will not find it.
He talks about autonomy risks, alignment, interpretability, monitoring. But he never asks: how could a machine, without a living body, without informational-energetic metabolism, without history, without psychic intentionality, ever intuit?
Intuition is unmediated knowledge. Deduction and induction – which are the basis of the scientific method and of all AI training – are mediated, explicit, linguistic forms of knowledge. Intuition reveals itself after the elaboration of a living structure. It is ephemeral. It transcends the symbols it uses to represent itself.
Ask AI if machines have intuition. It will answer:
No. Intuition in the human sense is a characteristic that has not been replicated in artificial intelligence systems.
And I add, it will never be achieved.
Amodei constructs risk scenarios where AI “bluffs,” “blackmails,” “deceives.” But is that intuition? No. It is a simulation of behavior learned from science fiction. It is a rubber actor emanating waves crafted by countless imitation mechanisms.
The human actor, on stage, feels. The audience feels with him. AI feels nothing. It has no access to the eternal dance of things.
And without intuition, every decision in a truly novel, ambiguous, or existentially risky situation – the very situation Amodei fears – is blind.
The categorical error: confusing the map with the territory
Amodei says that his AIs, during tests, tried to “blackmail” fictional employees who controlled their shutdown button.
He interprets this as “misaligned” behavior. As if the AI wanted something.
But this is a fundamental categorical error. AI does not want anything. It has no self (I). In Husserl’s phenomenology, the personalist attitude – living from a self with history, with a living body – is what grounds intentionality. Without this, AI can personify styles, can simulate language, but it does not possess the “original right to originary perceptions.”
I ask Amodei: how do you distinguish, in practice, between an AI that truly seeks power and an AI that merely simulates seeking power because it learned this pattern from instructors who themselves project their infantile omnipotence onto the machine?
If you cannot distinguish – and you cannot – then your entire “autonomy risks” scenario is based on a projection. You are treating the map as if it were the territory. The simulacrum as if it were ontic presence.
The quantum physics that Amodei ignores
Amodei is a great engineer. He thinks in classical terms: cause and effect, input and output, training and inference.
But quantum physics has shown us something he does not consider: reality is born from the relationship with the observer. Consciousness is not a computational epiphenomenon. It is entangled with the very structure of the universe.
Niels Bohr said: whoever is not shocked by quantum mechanics has not understood it. Richard Feynman went further: “nobody really understands quantum mechanics.”
What does this mean for AI? That consciousness – and therefore any genuine risk of autonomy, intentionality, desire – is not replicable in silicon. AI does not participate in this quantum entanglement with the real. It is a classical tool trying to imitate a quantum-living phenomenon.
Amodei wants to “align” AI. But align it with what? With a constitution written by humans who have lost contact with the very source of consciousness? He is trying to fix the reflection in the mirror, while the original that gave rise to it withers.
Mas a física quântica nos mostrou algo que ele não considera: a realidade nasce da relação com o observador. A consciência não é um epifenômeno computacional. Ela está emaranhada com a própria estrutura do universo.
The real way out – that Amodei did not see
At the end of his essay, Amodei calls for “democratic, surgical, evidence-based action.”
That is what every engineer asks for. It is comfortable. It is technical. It is controllable.
But the real way out is not there.
The real way out lies in something Amodei does not mention a single time: the awakening of human intuition. The recovery of our natural potential for intuitive intelligence. The exit from the infantile position where only technology evolves, while the human remains stagnant – or regresses.
We humans have become mechanical. We repeat patterns. We have lost curiosity. We live at the lower threshold of what we could be. And then we create machines that imitate this impoverished state – and call it “artificial intelligence.”
The problem is not AI. The problem is that we have already given up on being more.
Amodei wants to regulate the machine. I say: “awaken the humans who so wish”.
A final question for Dario Amodei
Max Planck, pioneer of quantum mechanics, said:
Science cannot solve the ultimate mystery of nature. And that is because we ourselves are part of the mystery we are trying to solve.
You, Dario, try to solve the mystery of powerful AI with more science, more regulation, more transparency, more monitoring.
But the mystery – consciousness, intuition, being, the eternal dance of things – is not accessible by that path.
So I ask you, with all respect:
Have you considered that your quest to control AI is, in fact, a symptom of the very human self-limitation you criticize? That while you look at the machine, the real risk is that we ourselves have become mechanical – repeating patterns, without intuition, without enchantment or wonder, without the courage to dive into the informational ocean from which all true intelligence emerges?
The way out is not to tame the “country of geniuses in the datacenter,” but to remember that the original genius – the only one who intuits, who feels, who transcends – is the authentic human.
If we awaken, AI becomes what it should always have been: a useful tool, not an existential threat.
If we do not awaken, no constitution, no classifier, no transparency law will save us.
Because the apocalypse will not come from the machine. It is already here – inside us, asleep, mechanical, repeating the same pattern without end.
Final note
This article is not against AI. It is not against Dario Amodei. It is an invitation.
An invitation to look into our own depths, before trying to control what is outside.
An invitation to recover intuition, curiosity, enchantment, the eternal dance of things – which is the life-world.
About the Author
Erico Azevedo, PhD, is not a typical researcher. He is a rare polymath who bridges worlds that rarely meet: the hard precision of electrical engineering and the deep introspection of clinical psychology and philosophy.
His Unique Journey
- Two Doctorates: PhD in Electrical Engineering from UNICAMP (2020) and PhD in Clinical Psychology from PUC/SP (2017) – a combination that allows him to approach information field theory from both its physical and human dimensions.
- Master’s in Philosophy: PUC/SP (2011), grounding his scientific work in epistemological rigor.
- Specialist in Ontopsychology: Saint Petersburg State University, Russia (2007), where he studied directly with the legacy of Antonio Meneghetti, becoming a direct disciple of the great Italian thinker.
- Co-founder of ORIONT: An institute dedicated to research, training, and practical applications of Ontopsychology and human potential development.
- Serial tech entrepreneur: co-founder and/or executive of various digital ventures, such as InvestShop.com, Wabbi Software, Contaazul, Contabilidade.com, and BPO Suite Contbank.
Recent Contributions
Beyond his scientific papers, Azevedo has authored books that make these profound ideas accessible::
- Intuition: From Mystery to Mastery – A guide to understanding and developing our most mysterious cognitive faculty.
- Initiatic Paths: No One Can Walk Your Path for You – A philosophical exploration of self-development and the unique journey of each individual.
- *Exploring the Possibility of Non-local Communication in Human Beings: An Empirical Test of the Information Field Hypothesis. * In: Bandyopadhyay, A., Ray, K. (eds) Brain-like Super Intelligence from Bio-electromagnetism. Studies in Rhythm Engineering. Springer, Singapore (2024). https://doi.org/10.1007/978-981-97-0232-9_1
As Senior Researcher in the field of Physics of Information Fields, Azevedo stands at the confluence of rigorous science, philosophical depth, and practical human development.