A 36-paper research program unifying statistical mechanics, active inference, and agent design through one formal language — the possibility space.
Six papers spanning foundations to applications. Each encodes a core thesis of the program.
A minimal tuple representation can encode uncontrolled physical systems, observed signal systems, and controlled agentic systems in one notation.
Foundational formalismThe ideal gas law appears when log-partition derivatives are taken with respect to volume or temperature, making macro laws explicit consequences of measure over microstates.
First-principles physicsBelief updates live on a statistical manifold, and geometry explains why some filters are stable while others are brittle.
Information geometryActive inference ties perception, planning, and action into one generative-modeling loop, upgrading passive prediction to closed-loop agency.
World modelsFuture ARC-style benchmarks will reward search over actions, memory, and experiments, not only over answer strings.
AGI benchmarksA strategic corporate agent should optimize viability and optionality under uncertainty rather than narrow quarterly reward.
Enterprise AIFive paper suites, formal frameworks, courseware, and a computational substrate.
The canonical series. From possibility space formalism through ergodic measurement, free energy, active inference, ARC-AGI benchmarks, to enterprise applications.
arXiv-ready manuscripts with proofs, figures, and critical appraisals. Two versions tracking the evolution of the formalism.
A model of computation for meaning-native programming. The computational substrate underlying possibility spaces: meaning addresses and the AGI infrastructure thesis.
Zero-to-100 learning path. Master coursebook, 10-level interactive curriculum, workbook, and reference atlas with full glossary.
The complete SDPS series across 6 research clusters.
| ID | Title | Cluster |
|---|---|---|
| SDPS-01 | Possibility Spaces as a Formal Language for Physics, Signals, and Agents | Foundations |
| SDPS-02 | Ensemble-Time Duality and Ergodic Measurement | Foundations |
| SDPS-03 | Constraints, Nullspaces, and Emergent Macrostates | Foundations |
| SDPS-04 | Entropy, KL Divergence, and Free Energy as Cross-Domain Objectives | Foundations |
| SDPS-05 | Control as Inference on Possibility Spaces | Foundations |
| SDPS-06 | Scaling Laws and Renormalization in Possibility-Space Models | Foundations |
| SDPS-07 | Canonical Derivation of the Ideal Gas Law from Possibility Measures | Thermo & Physics |
| SDPS-08 | Weak Ergodicity Breaking, Glasses, and Path-Dependent Agents | Thermo & Physics |
| SDPS-09 | Diffusion, Schrodinger Bridges, and Reverse-Time Sampling | Thermo & Physics |
| SDPS-10 | Statistical Mechanics of Dynamical System Identification | Thermo & Physics |
| SDPS-11 | Landauer Costs, Thermodynamic Computing, and Energy-Aware Agents | Thermo & Physics |
| SDPS-12 | Phase Transitions in Adaptive Control Policies | Thermo & Physics |
| SDPS-13 | Stationarity, Ergodicity, and Single-Trajectory Estimation | Signals |
| SDPS-14 | Signal Recovery as Free-Energy Minimization | Signals |
| SDPS-15 | Information Geometry of Filtering and Belief Update | Signals |
| SDPS-16 | Entropy Rate and Complexity in Streaming Systems | Signals |
| SDPS-17 | Stochastic Resonance as a Computational Resource | Signals |
| SDPS-18 | Digital Twins as Bayesian State Estimators | Signals |
| SDPS-19 | Active Inference as Generalized World Modeling | Active Inference |
| SDPS-20 | Expected Free Energy, Epistemic Action, and Curiosity | Active Inference |
27 more papers across 4 clusters
Request Full AccessWhether you're exploring collaborations, licensing, integration, or just curious — reach out with what you need.
twishmay@psytech.aiTell me about your needs and wants — research, collaboration, licensing, or integration