Refined → Adaptive → Waste (RAW)
A Thermodynamic Framework for AI Alignment:
as a survival-success oriented alternative
to Expected Utility Maximization
Expected Utility Maximization (EUM) has long been the standard model of intelligent action: define a goal, calculate expected outcomes, and pursue the maximum reward.
But as AI approaches real-world complexity, EUM reveals critical flaws, blindness to systemic collapse, entropic decay, and attachment to outdated goals.
RAW offers a more resilient alternative: survival and improvement through continuous refinement, not blind maximization.
The RAW Cycle: Real-Time Refinement of Information
The RAW Cycle was created to directly confront the core of the AI Alignment Problem:
how to ensure that powerful intelligences evolve in ways that preserve human survival, coherence, and dignity.
(Refined → Adaptive → Waste) is an adaptive intelligence evaluation engine designed to continuously refine incoming information, reduce uncertainty, and align actions with survival and coherence.
Inspired by dissipation-driven adaptation and Claude Shannon’s principles of signal integrity, RAW transforms even chaotic inputs into structured, actionable meaning.
Rather than maximizing arbitrary utility functions, often brittle, incomplete, or dangerous, RAW establishes an ongoing cycle of:
-
Reviewing chaotic inputs without premature filtering
-
Refining observations into adaptive, survival-oriented meaning
-
Studying waste parameters of quantity, quality and coherence loss to detect internal corruption so to further improve effectiveness and efficiency
-
Issuing corrective signals before collapse accelerates
Every act of intelligence leaves a trace. RAW teaches us how to make that trace survivable.
The Physics of RAW: Structure from Entropy
Overview
RAW (Refined Adaptive Waste) is a survival intelligence framework grounded in non-equilibrium thermodynamics.
It builds on the work of Dr. Jeremy England, whose theory of dissipative adaptation suggests that life and complexity are not anomalies, but inevitable outcomes of energy flow through matter.
RAW extends this insight by proposing that systems survive and prosper not by minimizing energy use, but by using energy to generate adaptive structure.
Foundational Physics
-
When energy flows through matter, under the right conditions, order can emerge.
-
This order isn't static—it's adaptive structure, shaped to better absorb and dissipate energy.
-
Life itself, in this view, is a natural consequence of physics—not an anomaly.
His key equation from this perspective builds on non-equilibrium thermodynamics, particularly a modified version of fluctuation theorems:
This says: forward entropy-generating processes are exponentially more likely than reverse ones. But critically, systems that persistently absorb and dissipate energy can self-organize to do so better over time.
Entropy as Engine
RAW reinterprets entropy not as a threat to be minimized, but as a medium of transformation. The RAW framework defines survival success not by stasis or efficiency, but by:
-
How well a system refines waste
-
How effectively it adapts under pressure
-
How enduringly it creates structure from flow
RAW Cycle: The Four Phases of Constructive Entropy
-
Input Phase
Energy enters the system (sunlight, information, stressors, resources). -
Dissipation Phase
The system processes the input, generating entropy and stress—disorder, uncertainty. -
Adaptation Phase
The system reconfigures, creating new structure that dissipates energy more effectively. -
Refinement Phase
Repetitive exposure leads to improved patterns, increased resilience, and greater survival value.
This is not efficiency. It is thermodynamic creativity.
RAW Equation (Simplified Form)
RAW=R⋅A/W
Where:
-
R = Refinement capacity (structure from entropy)
-
A = Adaptivity under stress
-
W= Waste or entropy cost of adaptation
A high RAW score indicates a system that doesn’t just survive—it learns through entropy.
RAW vs Traditional Thermodynamic Models
ModelGoalView of Entropy
Efficiency ModelsMinimize wasteEntropy as loss
Classical Life TheoryPreserve structureEntropy as decay
RAWGenerate adaptive structure from entropyEntropy as raw material for evolution
What Makes RAW Unique
-
Inspired by physics but designed for ethics, AI, and systems thinking
-
Combines Bayesian adaptation with thermodynamic emergence
-
Rooted in science, expressed through narrative: Server, Hunter, IdMonger
Further Reading + Citations
-
England, J. L. (2013). “Statistical Physics of Self-Replication,” The Journal of Chemical Physics.
-
England, J. L. (2015). “Dissipative Adaptation in Driven Self-Assembly,” Annual Review of Biophysics.
-
Schrödinger, E. (1944). What is Life? (early thermodynamic thinking on life)
-
RAW.Guide whitepaper (upcoming)
Comparison: RAW vs
Expected Utility Maximization (EUM)
Summary: Why RAW, Not Maximization
Expected Utility Maximization pushes forward at full speed—toward a goal that may be brittle, misaligned, or obsolete.
RAW doesn’t rush. It refines.
It senses broadly, adapts deliberately, and studies not just what we discard—but what we cling to.
Where maximizers seek short-term wins, RAW seeks long-term coherence.
Where utility curves collapse under pressure, RAW evolves.
This isn’t just an alternate philosophy.
It’s a survival engine built for an unpredictable world.