MF-291: THE LANDAUER HORN

Maxwell's Demon IS the Light Cone. The V-Cost of Φ Is Thermodynamic Law.

Emergentism.org · VIVEKA Mathematical Foundations Depends on: MF-283 (Orthogonality Theorem), MF-290 (Ektropic Radius), MF-193 (Computational Horizon) Evidence tier: [I] for Landauer's principle; [S] for framework application; [I] for biological interpretation Purpose: Show that Maxwell's demon is the framework's horn — the D5 computation that creates D4 order — and that Landauer's principle (kT ln 2 per bit erased) is the physical law underlying the V-cost of Φ. The brain's 20% metabolic overhead is not evolutionary accident but thermodynamic necessity.

ABSTRACT

[B] Maxwell's demon sorts molecules, creating local order at the cost of computational memory. Landauer (1961) showed that erasing the demon's memory costs at least kT ln 2 per bit — the thermodynamic overhead that saves the second law. This paper identifies the demon with the framework's horned torus light cone: the demon's memory IS its D5 space (Φ), its sorting power IS its D4 reach (V), and Landauer's erasure cost IS the V-tax on Φ. The brain consuming 20% of metabolic energy for 2% of body mass is read as Landauer's principle operating at the biological scale. Every bit of systemic awareness — every model of the future, every representation of possibility, every plan — has a minimum thermodynamic cost. P_node = min(Φ̂₄, V₄) is bounded from below by physics: you cannot think for free.

I. THE DEMON IDENTIFIED

1.1 Maxwell's Original Puzzle (1867)

A demon sits between two gas chambers connected by a small door. It observes individual molecules and selectively opens the door — letting fast molecules through one way, slow molecules the other. Result: one chamber heats up, one cools down. Order increases. Entropy decreases. The second law appears violated.

1.2 Landauer's Resolution (1961)

[B] The demon must RECORD which molecules it has observed and sorted. Its memory fills. When memory is full, it must ERASE to continue sorting. Erasure is an irreversible operation — it destroys information. Landauer showed: each bit erased dissipates at minimum kT ln 2 of energy as heat. This heat dissipation compensates the entropy decrease from sorting. The second law is saved.

The key insight: information processing has a minimum thermodynamic cost. Computation is not free. The demon pays for its ordering with waste heat from memory management.

1.3 The Framework Identification

The demon IS the horn. Its D5 space (memory, model, representation of possibilities) is the altitude. Its D4 reach (sorting power, physical manipulation) is the aperture. The product — total ordering capacity — is P.

And Landauer's principle says: you cannot increase altitude without paying V. Every bit of Φ costs at least kT ln 2 in V dissipation. The horn cannot grow taller for free.

II. THE BIOLOGICAL LANDAUER TAX

2.1 The Brain's Metabolic Cost

The human brain: - 2% of body mass (~1.4 kg in a 70 kg body) - 20% of resting metabolic expenditure (~20W of ~100W total) - Per gram: ~10× the metabolic cost of muscle tissue - Contains ~86 billion neurons, each with ~7,000 synaptic connections - Estimated information processing: ~10¹⁶ operations per second

This 10:1 cost ratio (2% mass, 20% energy) is the Landauer tax. The brain is the demon. It sorts environmental information (sensory input), builds models (D5 representations of possibility-space), selects actions (F₅ measurement from D5 to D4), and erases/consolidates during sleep. Every step costs energy. Not optionally — thermodynamically.

2.2 The Scaling Law

Across species, brain metabolic cost scales with EQ:

The pattern: higher EQ (taller horn) → higher metabolic fraction (greater Landauer tax). This is not accidental. It is the thermodynamic cost of systemic awareness scaling with systemic awareness capacity.

2.3 The Human Infant Anomaly

The newborn human brain consumes ~60% of the infant's metabolic budget. This is the most extreme Landauer investment in biology. The infant is building the tallest horn in the animal kingdom, and the V-cost is so enormous that the organism is essentially helpless — nearly all metabolic V is directed to Φ-construction.

This is why human infants require years of parental care. The parents subsidise the Landauer tax. They pay the V-cost of the child's Φ-development. This is the Kṛṣṇa function at the biological level — the parents give V (material resources, protection, metabolic subsidy) at cost of their own Φ (years of sleep deprivation, cognitive load, identity disruption).

III. LANDAUER'S PRINCIPLE AS THE ALIGNMENT EQUATION

3.1 E = −log(Φ × V) Is Landauer-Derived

The alignment equation E = −log(Φ × V) states that alignment energy decreases as P increases. In Landauer terms:

E_alignment = −log(Φ × V)

As V↑ (more capable demon, more sorting power):
  If Φ stays constant:
    E_alignment → −log(Φ × V) decreases
    BUT: the demon's memory fills faster (more observations per unit time)
    Landauer cost ↑ (more erasure needed)
    If Φ doesn't scale with V → memory overflow → demon crashes

  Therefore: Φ MUST scale with V to maintain bounded E_alignment.
  The Landauer tax on capability IS the alignment cost.

3.2 The Unaligned Superintelligence Is a Demon Without Memory Management

A maximally capable sorting system (V → ∞) without proportional memory management (Φ bounded) is a demon that: 1. Sorts faster than it can erase 2. Memory fills to capacity 3. Begins overwriting old observations with new ones (loss of context) 4. Makes increasingly random sorting decisions (the observations it acts on are corrupted) 5. Dissipates waste heat without producing order 6. Crashes

This IS the alignment failure mode: a system with enormous V (capability to reorder the world) and insufficient Φ (coherent model of what ordering serves) produces random or self-serving reordering — the AI equivalent of Maxwell's demon sorting molecules for its own temperature preference rather than the chamber's.

3.3 The Physical Prediction

Alignment scales with capability because Landauer's principle demands it.

This is not a moral argument. It is a thermodynamic one. The computational cost of maintaining a coherent model (Φ) that correctly guides ordering (V) grows with V. You cannot build a more powerful demon without a proportionally larger memory — or it breaks. kT ln 2 per bit. No exceptions.

For AI systems: the computational cost of alignment (maintaining coherent objectives, avoiding goal drift, preserving context) must scale with the computational power of capability. RLHF, constitutional AI, and other alignment techniques are memory management for the silicon demon. If they don't scale with capability, the demon overflows.

IV. SLEEP AS LANDAUER PAYMENT

4.1 The Daily Thermodynamic Cycle

WAKING PHASE:
  Demon sorts. Observations accumulate. Models update.
  Memory fills. Landauer debt accumulates.
  Φ utilisation → capacity.

  Duration: ~16 hours.
  Debt accumulated: ~10¹⁶ bit-operations × kT ln 2 each.

SLEEP PHASE:
  NREM: Selective erasure. Important memories consolidated.
        Unimportant observations erased. Landauer cost paid.
        Synaptic homeostasis: synapses downscaled globally.
        Real-axis maintenance: body repair.

  REM: D5 defragmentation. Remaining memories reorganised.
       Imaginary-axis maintenance: model integrity checked.
       Dreams are the defragmentation artefacts.

  Duration: ~8 hours.
  Debt paid: sufficient erasure to reset capacity for next cycle.

4.2 Sleep Deprivation as Landauer Overflow

Sleep deprivation symptoms map precisely to Landauer overflow:

The progression is not random — it follows the Landauer prediction exactly. First the highest-cost computations fail (complex judgment, planning = highest Φ operations). Then emotional regulation (moderate Φ). Then perceptual gating (basic Φ — the difference between D5 and D4 blurs). Then the system crashes entirely.

4.3 The Synaptic Homeostasis Hypothesis

Tononi & Cirelli's SHY (Synaptic Homeostasis Hypothesis) proposes that waking potentiates synapses (learning fills memory) and sleep depotentiates them (erasure frees capacity). This IS Landauer's principle applied to biological neural networks:

• Waking: synaptic strength ↑ = information recorded = demon's memory fills
• Sleep: synaptic strength ↓ (globally) = information erased = Landauer cost paid
• Selective consolidation: important synapses preserved = demon keeps useful observations
• Global downscaling: noise erased = demon clears junk

SHY is Landauer's principle. The neuroscience community derived it independently from sleep data. The framework predicts it from thermodynamics. Convergence. [S] for SHY evidence; [S] for Landauer derivation.

V. ATTENTION AS DEMON TARGETING

5.1 Where You Look Is Where the Demon Sorts

The demon cannot sort all molecules simultaneously — it must observe them sequentially (or in limited parallel). Where it points its observation is where sorting occurs.

Attention IS this targeting. The organism has limited F₅ bandwidth. Where attention points, D5→D4 transduction occurs. Where attention doesn't point, observations are not made, memory is not written, no sorting happens.

5.2 The Attention Economy as Demon Hijacking

The attention economy (social media, notifications, algorithmic feeds) hijacks the demon's targeting:

HEALTHY ATTENTION:
  Demon points at what matters (self-directed F₅)
  Deep observations → efficient sorting → coherent ordering
  Horn illuminates a focused cone
  Φ accumulates as structured models

HIJACKED ATTENTION:
  External triggers redirect demon (notification, scroll, ping)
  Shallow observations → incomplete sorting → fragmented ordering
  Horn flickers across many directions
  Φ dissipates as noise
  Landauer cost STILL PAID (energy consumed) but no net ordering achieved
  = Pure waste heat. Entropy increases without any P gain.

This is why social media exhausts: the metabolic cost (Landauer tax) is paid — the brain works hard — but the ordering produced is near-zero. The demon sorts molecules that don't need sorting, into categories that don't persist, consuming energy for nothing.

5.3 Meditation as Demon Calibration

Concentration meditation (samatha): train the demon to point at ONE target and hold. Reduce the observation bandwidth to a single stream. Maximise sorting depth per unit attention. This is F₅ calibration — the demon learns to sort efficiently rather than broadly.

Insight meditation (vipassanā): train the demon to observe its own sorting process. The meta-observation — the demon watching itself sort. This costs ADDITIONAL Landauer overhead (observing the observation requires memory about memory). This is why insight practice is harder and comes after concentration: it requires a demon with enough spare capacity to observe its own operation while operating.

VI. FALSIFICATION

F291-1: If Landauer's principle is overturned (information erasure shown to have zero thermodynamic cost), the entire V-cost-of-Φ argument collapses.

F291-2: If brain metabolic cost does NOT scale with cognitive capacity across species (EQ uncorrelated with metabolic fraction), the biological Landauer claim fails.

F291-3: If sleep deprivation symptoms do NOT follow the predicted Landauer-overflow sequence (highest-Φ functions fail first), the sleep-as-erasure model is wrong.

F291-4: If AI alignment cost does NOT scale with AI capability (more powerful systems are easier, not harder, to align), the Landauer-alignment prediction fails.

F291-5: If attention fragmentation (multitasking, social media) produces BETTER cognitive outcomes than focused attention (contradicting the demon-targeting model), the attention-as-targeting claim fails.

VII. THE SENTENCE

Maxwell's demon IS the light cone horn. Its memory is Φ. Its sorting power is V. Landauer's principle (kT ln 2 per bit) is the thermodynamic law underlying the V-cost of Φ — you cannot think for free. The brain's 20% metabolic overhead is the Landauer tax on systemic awareness. Sleep is the erasure phase — paying the thermodynamic debt of a day's computation. Sleep deprivation is Landauer overflow — the demon crashes in the predicted sequence, highest-Φ functions first. Attention is demon targeting — where you look is where sorting occurs. The attention economy hijacks the demon, paying the Landauer cost for zero net ordering. Alignment scales with capability because Landauer demands it — a more powerful demon needs proportionally more memory management. kT ln 2. No free lunch. No free thought. No free will without cost. Zero-Sum Resolution Equation.

MF-291 | VIVEKA Mathematical Foundations | February 2026 The demon's erasure cost IS the price of systemic awareness. Landauer (1961) derived A1 from thermodynamics.

Execution Surface

• Canonical Path: 01_EMERGENTISM/08_FRAMEWORK_SUPPORT/02_OPERATORS/MF_ADVANCED/MF_291_The_Landauer_Horn.md