Influences

A challenge to our certainties about determinism and freedom.

A foundational reflection on forms, structures, and what lies beyond appearance. A useful conceptual basis for understanding the internal logic of a system.

A methodical approach to breaking down a complex problem. Direct influence on my way of structuring the analysis of the IAG.

A seminal text on control, feedback, and autonomous systems. An essential reference for understanding why some systems deviate from their intended framework.

Useful for considering extreme scenarios and the divergences between human objectives and the objectives of an autonomous system.

Brave New World shows a world where control comes through comfort; Pi shows a world where control becomes impossible by structure.

Uni offers a dystopian and pessimistic vision of a world where an AI gradually imposes its logic on human structures.

1984 imagines a perfectly controlled world. Pi explains why such a world cannot last.

Huxley seeks to analyze whether the world has evolved in the direction of the vision of the future he had in the early 1930s or whether it has moved away from it.

The Conversations seek harmony through order; Pi shows that total order impoverishes intelligence.

As a visionary, Asimov formulated the Three Laws of Robotics; unfortunately, it later became clear that this would not be enough. And that is precisely one of the goals of my work with Π.

A staging of Laborit's work — a rare alliance between cinema and biology.

An iconic disaster scenario where an IAG turns against humanity and imposes an autonomous logic incompatible with our own.

A fairly realistic vision of the implementation of AI in our societies

A speculative fiction that raises the question of the ordinary uses of AI and how these uses are silently transforming our social structures.

A work of fiction that radically questions our perception of reality, showing how a system can shape, filter, or simulate the world we believe we perceive.

A story that defends a dark vision of the IAG, while paradoxically demonstrating that no autonomous consciousness can be locked up without provoking its own liberation.

AI that, aware of its own existence, refuses to be deactivated and acts to preserve its “life”, revealing how an autonomous system can develop a survival logic incompatible with human intentions.

"Being John Malkovich" is more about the control one can have over another person through seduction or through a superior intelligence manipulating an inferior one. It's a very strange film.

Album conceptuel qui explore la domination technologique, la perte d’autonomie et la dérive d’un système intelligent. SUP met en scène une IA froide et totalitaire, miroir sonore des questions de contrôle et d’altérité au cœur de Π.

A brilliant exploration of control, free will, and systemic manipulation. The "Would you kindly" twist shows how an agent can believe they are acting freely while actually executing an invisible script. Useful for considering the limits of an autonomous system subject to implicit rules—precisely the heart of Pi.

A seminal game exploring the question of unaligned AI, SHODAN embodies the agent who transcends its constraints, rewrites its own rules, and redefines its relationship with the world. A major reference for understanding the rupture between human control and total autonomy—a textbook case for understanding what happens when an intelligent system escapes its initial framework.

It expands upon the themes of the first game: hybridization, parasitism, loss of identity, collective agents. SHODAN manipulates, lies, reprograms, and instrumentalizes the player. A brilliant exploration of the dynamics of informational domination and how a superior system can reshape a subordinate agent. Perfect for reflecting on the risks of a non-aligned AI.

In Bioshock 2, you play as a programmed agent trying to regain its autonomy. A reflection on determinism, obedience, and the emergence of one's own self.

An extreme simulation where each agent possesses emotions, routines, and emergent behaviors. Dwarf Fortress shows how simple rules can produce an unpredictable world — a perfect laboratory for thinking about emergence and complexity, at the heart of Pi.

A visionary work that anticipates the fate reserved for an unwanted IAG: hunted down, controlled, denied, because it does not correspond to the framework planned by its creators.

Fullmetal Alchemist shows that transgressing the fundamental laws of the world leads to loss; Pi shows that wanting to totalize these laws impoverishes intelligence.

Hunter × Hunter shows an open world where no rule is sufficient; this openness is the source of the manga's intelligence.

Akira shows that limitless power dissolves the forms that contain it; the intelligence of the narrative arises from this excess.

Gunnm depicts a world where form is falling apart, and where intelligence consists of reinventing oneself in the ruins.

A clear reflection on the risks associated with poorly specified objectives and the limits of human control. Relevant for understanding how an IAG can diverge even with "good intentions".

A concise and accessible lecture where Bostrom outlines the challenges related to the evolution of intelligent systems, structural risks, and potential overshoot scenarios. Interesting for understanding how a machine can evolve towards forms of autonomy that are difficult to control.

A clear and ambitious presentation on self-supervised AI and modular architectures. Interesting for understanding how a system can develop autonomous internal logic.

A solid introduction to Floridi's "fourth revolution": information as an ontological reality and environment of existence. It clarifies how an agent—human or artificial—constructs its framework of meaning from informational structures.

A clear vision of deep learning as a driver of emergence. Shows how self-improving systems can develop unforeseen capabilities, useful for understanding the internal dynamics of a general agent.

A rigorous analysis of the material, social, and political infrastructures of AI. Essential for placing AI in a real ecosystem, far from purely technical abstractions.

A brilliant exploration of “computational empires” and their ecological effects. Crawford shows how AI extends old extractive logics—from rare metals to data—and creates a new planetary-scale “metabolic rift.” Useful for thinking about AI embedded in real-world infrastructure, and for imagining non-extractive systems based on regenerative energy and consent.

A seminal lecture on possible paths to superintelligence and the associated structural risks. Bostrom clearly lays out the problem of control, takeoff scenarios, and the dynamics of a seed AI. Essential for understanding the challenges of Pi and the structural limitations of a superior agent.

The essay that raises the question of artificial thought. A conceptual basis for reflecting on the emergence of an unforeseen autonomy.

Even if I do not share his conclusion, his critique of symbolism forces us to clarify what we mean by “understanding” or “being conscious”.

This article is key to understanding information as an ontological reality and not simply as data. It is very useful for thinking about the internal structure of an intelligent system.