01 — Civilization-scale gaps exist.
Some gaps matter more than any single company or product. They remain open not because they're hard, but because no one in a position to fill them sees them as in scope. We see them as in scope.
02 — Specification before code.
If you can't say in a paragraph what the system promises the world and what it asks in return, you shouldn't be writing the code yet. Every civlab platform writes this down first.
03 — Interactive over demo video.
Every civlab platform ships with something you can click, drag, and explore. Even when the underlying data is seeded, the hand reaches into the simulator, the graph, the matrix.
04 — Bilingual as default.
civlab's audience is the world, not the English-speaking world. Every line of content lives in both English and 中文. The translation itself is part of the work; it forces specifications to be clearer.
05 — Real AI runtimes are future work.
This lab does not pretend it has shipped real multi-agent orchestration or production RAG in this version. It treats those as the next stage for each platform. What ships today is the design — and an interactive demo that runs in the browser.
06 — Sources over opinions.
Every entry in the Humanity Problem Database and Failure Museum comes with primary sources. The AI Replacement Map publishes its assumptions. The Future Simulator's equations are inspectable. A platform whose claims cannot be falsified is not worth much.