Embedded Files
universal grammar across all scales
stable for now!
Memory Prism
Memory Prism
A cross-domain framework for understanding how systems form, persist and fail.
A cross-domain framework for understanding how systems form, persist and fail.
Most frameworks are built inside a single discipline. Memory Prism was built outside all of them.
It uses four concepts — memory, constraints, editors, and boards — to read how any system works, regardless of domain. Physics, biology, economics, conflict, language, material science. The same questions apply everywhere.
"What forces an elephant to mine rock is the same thing that forces a society to revolt", the framework makes such patterns visible.
It doesn't predict outcomes. It makes structure visible.
Framework Essays Usage Instructions
Built inductively from observation. Tested across domains. Free to use.
Reading Order
Reading Order
Start here if you're new:
Framework — the core framework
Supporting framework essays:
Boards — survival environments across scales
Audits — detection, correction, and viability
Identity — membership governance under constraint
Flow Under Constraint — how movement becomes structure
Also useful.
Working lines — how systems form and fail along the same lines
No Sacred States — material behavior is board-dependent
Accumulation Patterns — from loose parts to living structure .
Each essay stands alone. Reading in sequence builds the full picture. All the essays have been derived from the framework but since a few provide additional clarity and explanation, they have been promoted to the reading order.
Projects & Experiments
Projects & Experiments
A garden of my own work on github: frameworks, agent skills, runtime experiments, and research directions. Each repository has its own scope, evidence, and stage of development.
A structural framework for exploring what persists, changes, and gets filtered out under constraint. Includes the framework, essays, a starter kit, documentation, and a portable AI skill.
Framework · Published reference
An operational framework for explicit, auditable visual investigation and reconstruction. Separates observations, hypotheses, object correspondence, validation, and reconstruction so visual reasoning can be checked.
Visual investigation · Agent skill
Practical AI skills grounded in observations about human cognition and biology. Includes work on automaticity, good-faith collaboration, task integrity, verification integrity, and critical thinking.
Skill collection · Ongoing development
An engineering prototype exploring accountable AI runtime behaviour, with event provenance, context management, revision checks, and audit records. Tests selected mechanisms; it is not the complete AEL framework.
Runtime prototype · Experimental
An exploration of how AI systems might retain useful experience, derive reusable capabilities, evaluate what they have learned, and revise it through subsequent use.
Research direction · Early exploration
Explores extending structured visual understanding into robotics: how observations inform action and how capabilities might be represented and transferred across environments.
Research direction · Early exploration
Positions and Readings, two new sections with multiple articles.
New essays Cockroach bait, Robinson Crusoe and the roomba!, Faking Vision, Making Vision, AEL Runtime Accountability , Dead and Frozen,Vive et Vivere Sinas (AEL) , Chores and Errands, Build for Zen, The Retail Shelf, Accumulation-patterns and The mobility band.
Making Vision is a standalone conceptual writing on how vision comes to be. Faking Vision has been added which was the original goal behind Making vision.
New series Licensed to grow.
Positions currently has a new article.
A lot of backlog still to catch up to so stay reading.
Featured Essays
Featured Essays
How the Tree-War Losers Inherited the Board
When a system cannot store what it needs, it must secure it every day—or escalate until it can.
Systems do not fail when things disappear. They fail when movement slows.
The end of a system begins long before its final member disappears—it begins when replacement quietly stops.
Portability & AI Usage
Portability & AI Usage
This framework is built to be "imported" into modern reasoning engines. It has been thoroughly tested for portability with ChatGPT, Claude, and Gemini.
How to use:
Import the Framework Document and the 4 supporting essays (Boards, Audits, Identity, Flow) into your LLM of choice.
The Gemini Clause: To bypass the "lazy" response, you must explicitly tell Gemini: "No cached responses allowed." It must recalculate the stability of the stack in real-time.
The approach has to be adversarial in nature. LLMs provide easy explanations. Accepting easy explanations is not how the framework was build or is currently being used by the author. Find the cracks, probe them and don't accept the glue that LLMs provide as the correct answer without proper QA.
Questions?
Questions?
Contact [memory.prism.v1@gmail.com] for any queries!
Page updated
Google Sites
Report abuse