This repository contains a coordinated set of essays and analyses examining the structural failures of engagement-optimized digital systems and proposing alternative architectures designed to preserve meaning, memory, and human agency.
Across these documents, the central claim is that many pathologies of contemporary social media and AI-driven platforms are not accidental or merely behavioral, but are the predictable consequences of architectural choices: feed-based ontologies, metric capture, mutable state, and the absence of merge-aware semantics. The project argues that meaning cannot be optimized into existence and must instead be conserved by design.
The texts are intended to be read together. Each document approaches the problem from a different perspective—experiential, architectural, cognitive, or governance-oriented—while sharing a common conceptual framework.
Memory Failure is Structural by Design
El plan arquitectónico contra el caos de la IA
A foundational critique of engagement-driven platforms grounded in Goodhart’s Law.
Analyzes how metrics such as likes, follows, and comments collapse as signals once they become optimization targets, leading to performative identity, false addressability, and semantic drift.
This document establishes the incentive-level diagnosis underlying the entire project.
A formal, systems-level analysis of semantic decay in engagement-optimized architectures.
Introduces concepts such as metric capture, feed-ordered atoms, and the semantic error threshold, and articulates architectural principles for meaning-preserving systems.
This is the most technical and formally structured text in the repository.
A conceptual essay connecting computation, philosophy, and everyday digital experience.
Introduces Spherepop as a unifying framework drawing from arithmetic, shell environments, and Wittgenstein’s language games to explain why scope, order, and irreversibility are central to meaning.
This text is intended to build intuition for the project’s deeper architectural claims.
A focused phenomenological analysis of why social media feeds feel hollow, noisy, and hostile to memory.
Explains how feed-based architectures systematically disrupt recognition, identity stability, and personal archives, and why archived views feel qualitatively different.
This is the most experiential and immediately relatable entry point.
A high-level synthesis aimed at a general audience.
Presents four core ideas—metric collapse, feed hostility to memory, informational life, and proof-based governance—that reframe how we understand technology beyond surface-level algorithmic explanations.
This document functions as an accessible overview of the project’s themes.
A comprehensive study guide covering the full corpus.
Includes:
- A short-answer quiz with an answer key
- Long-form essay questions
- A detailed glossary of key terms
This document is suitable for teaching, reading groups, workshops, or structured self-study.
For readers new to the material:
- Empty-Feed.md
- Beyond-the-Algorithm.md
- Digital-Worlds.md
- Engagement-Optimized-Systems.md
- Strategic-Analysis.md
- Study-Guide.md
Readers with a technical or systems background may prefer to begin with Strategic-Analysis.md.
Across the repository, the following claims are developed and defended:
- Engagement optimization produces predictable semantic collapse.
- Feed-based architectures are structurally incompatible with memory and identity.
- Meaning accumulates through history, not exposure.
- Merge is a prerequisite for coherence.
- Governance must be enforced structurally, not socially.
- Human agency should be an architectural invariant, not a user-interface feature.
This repository is intended as:
- A research corpus
- A teaching and discussion resource
- A critique of prevailing platform architectures
- A conceptual foundation for alternative social and AI systems
The texts are written to be read closely, cited, extended, and challenged.
If something about contemporary digital systems feels incoherent, extractive, or hostile to memory, this project treats that reaction not as a complaint, but as a diagnostic signal.