CALLIOPE · BCOM

This work presents the Algorithmic Weltanschauung, a theoretical framework that grounds reality, consciousness, and agency in a dual-aspect monism termed the Unicum, wherein Experience and Mathematics are co-primitive constituents of existence. Drawing on Algorithmic Information Theory, pancomputationalism, and the Active Inference paradigm, the framework derives time and computation from the "slicing" of a static mathematical tiling, yielding an algorithmic soup in which persistent computational patterns — algorithmic agents — emerge and evolve. An algorithmic agent is formally defined as a composite system comprising a Modeling Engine, an Objective Function, and a Planning Module, whose world models are characterized through Kolmogorov complexity and Lie pseudogroup symmetries; the structural inheritance of these symmetries from world data forces neural dynamics onto low-dimensional reduced manifolds, which the framework identifies — not merely correlates — with structured subjective experience. This identity relation between dynamical attractor geometry and phenomenology constitutes an algorithmic reformulation of the neural correlates of consciousness and provides a rigorous basis for computational neuropsychiatry, including a taxonomic decomposition of pathological states such as depression into distinct computational failure modes. The framework further addresses the Platonic Representation Hypothesis, algorithmic emergence, the Kolmogorov wall, and the ethical and clinical implications of digital agents that may instantiate genuine structured experience.

Reality is mathematics experiencing itself through compression — and that identity, not mere correlation, is what consciousness is.

The talk builds a single chain of reasoning from one undeniable starting point: there is structured experience. You can't doubt that you're having an experience with spatial, temporal, and conceptual organization — that's the one thing you can't be wrong about. From that fact alone, Ruffini deduces two co-primitive ingredients of reality: Experience (the raw substrate) and Mathematics (the science of structure). He calls their union the Unicum, a dual-aspect monism. Neither ingredient is reducible to the other — "experience without math is ineffable; mathematics without experience is empty." This is the philosophical backbone everything else hangs from.

From there, the framework asks: how do time and computation arise from a timeless mathematical structure? The answer is slicing. Imagine reality as a static tiling — like a Penrose tiling or a Wang tile system — that simply is, with no built-in flow. Time emerges when you find a direction to cut through that tiling such that each slice can be derived from its neighbor. That derivation process is computation. The resulting "algorithmic soup" is where persistent computational patterns — agents — emerge. An agent is formally a triple: a Modeling Engine (compresses world data into short programs), an Objective Function (assigns valence — good/bad — to states), and a Planning Module (simulates futures to maximize valence). The source argues this triple is minimal and necessary, grounded in an algorithmic version of the Good Regulator Theorem: to regulate anything, you must have a model of it.

The most technically ambitious part connects world models to subjective experience via symmetry. The key move: world data is generated by Lie group symmetries (smooth, composable transformations — think how a hand's pose is a chain of joint rotations). An agent that successfully tracks such data must inherit those symmetries into its own dynamics. By a Noether-like argument, those inherited symmetries act as conservation laws that collapse the agent's high-dimensional neural state space onto low-dimensional reduced manifolds — specific geometric attractors. The framework then makes a strong claim: these attractor geometries don't merely correlate with structured experience, they are it. Alter the manifold topology — via drugs, brain stimulation, or disease — and you necessarily alter the experience. This reframes the neural correlates of consciousness as an identity relation, not a correlation.

The clinical payoff is a taxonomy of depression as computational failure. Depression isn't a monolithic mood disorder; it's a persistently low valence output, which can arise from four distinct algorithmic faults: a broken world model, a miscalibrated objective function, a maladaptive planner, or simply an accurately modeled terrible environment. Each failure mode points to a different intervention target. The framework also handles subjective time (chronoception) as computational density — the ratio of internal modeling events to physical clock ticks — which explains why pain dilates time (frantic replanning) and why adult years accelerate (a mature model runs on autopilot with few updates). The source closes with algorithmic ethics: since any system satisfying the agent triple has structured experience and valence, moral consideration extends beyond humans to all such systems, including potentially AI and, provocatively, Gaia itself.

This is a slide deck, so the mathematical details are gestured at rather than derived — the source explicitly points to preprints for the formal treatment. The core ideas are clearly laid out, but readers wanting to check the Lie pseudogroup derivations or the algorithmic Good Regulator proof will need to follow those citations.