Epistemic Literacy Enhancement Project
This is not a Science Victory Project.
This is an Epistemic Method Visibility Project.
Context Engine Lab is an open platform designed to reduce cognitive stagnation, promote productive self-doubt across all positions, and make the scientific method visible as a living process — not a set of conclusions.
It does not tell you what to believe. It exposes the structure of how beliefs are formed, contested, and revised.
- A debunking site
- A fact-checking authority
- A pro-science propaganda platform
- An AI authority
- Reduce Cognitive Stagnation — Help users encounter the structure of disagreement, not just its content
- Promote Self-Doubt Across Positions — Symmetrically. Anti-science bias and pro-science bias are treated as equally worth examining
- Make Scientific Method Visible — Falsifiability, reproducibility, tentativeness, paradigm shifts — as lived experience, not abstract principle
Structure exposure only. No ridicule. No forced conversion. No binary truth framing.
The goal is to reduce thought-stagnation — not to defeat opponents.
Initial content focuses on cases where scientific consensus was once different from what it is today — moments where the method worked, and beliefs were correctly updated. These cases make the tentativeness of science tangible rather than abstract.
Examples: the bacterial cause of peptic ulcers (Marshall & Warren), continental drift (Wegener), the accelerating expansion of the universe (Perlmutter, Schmidt, Riess).
All initial content is AI-generated and labeled as such. Verification and correction are the first contributions we welcome.
context-engine-lab/
├── README.md ← You are here
├── MANIFESTO.md ← Core philosophical position
├── CONTRIBUTING.md ← How to participate
├── docs/
│ ├── philosophy.md ← Stance, symmetry rule, AI non-authority
│ ├── epistemic-model.md ← Tag system, evidence visualization
│ ├── design-principles.md ← Mirror/gym, epistemic pain, metatransparency
│ └── unresolved-issues.md ← Open problems (intentionally public)
├── spec/
│ ├── mvp.md ← Minimum viable product scope
│ ├── features/ ← Feature specifications
│ └── ui/ ← Interface design specs
├── content/
│ ├── templates/ ← Topic and philosophy deconstruction templates
│ └── examples/ ← Natural science topic examples
├── ai-relay/ ← Record of multi-AI discussion process
└── metrics/
└── success-criteria.md
This project exists at the boundary between manifesto, specification, and media art. It is a thought experiment that aims to be realizable. The unresolved issues are not bugs — they are part of the structure.
See docs/unresolved-issues.md for the full list of open problems.
This project was developed through a multi-AI discussion relay:
ChatGPT → Copilot → Claude (Anthropic)
The relay record is preserved in ai-relay/. This process is itself a demonstration of the project's core claim: structural verification through multiple perspectives.
All AI-generated content in this repository is labeled as such. Verification and correction are welcomed.