# The argument, in three papers

The manifesto states Fluid's position. These three papers examine it from three directions: the institutions whose systems fall behind, the software industry now making code cheap, and the experts whose knowledge those systems are meant to serve.

## Written to be read in order

Together they answer three questions: why change is needed, why it is possible now, and who should hold authority when it happens.

The first paper moves from recurring institutional symptoms to their common cause. The second asks what AI changes and identifies the governed domain model as the artifact software development has been missing. The third follows that architecture back to the people who hold the knowledge and asks what authorship and authority should return to them.

Each paper stands alone. Read together, they form one movement: diagnose the system, redesign the artifact, return authority.

## The State of the World

Institutions in rule-dense sectors run behind reality, and the first paper says why: interpretation of the work is translated into code, separated from its reasoning, and made expensive to revise. It traces the pattern — the shadow spreadsheets, the release queues, the audit answers that trail off — to that single translation, and names its price.

- [Read the paper](https://fluid.nu/en/state-of-the-world)
- [Download PDF](../papers/state-of-the-world.pdf)

## The State of Software Development

AI made code cheap, and the industry is reinventing itself around that fact. The second paper argues the bottleneck did not disappear — it moved. Faster code multiplies the places where domain meaning is independently embedded, and the artifact worth governing is not the code and not the spec: it is the model.

- [Read the paper](https://fluid.nu/en/state-of-software-development)
- [Download PDF](../papers/state-of-software-development.pdf)

## Liberating Domain Knowledge

The third paper is about the people who know. For decades experts have handed their knowledge to development queues and received systems that no longer speak their language. This paper describes the inversion: AI drafts, accountable experts decide, and interpretation gets a governed home inside the institution that owns it.

- [Read the paper](https://fluid.nu/en/liberating-domain-knowledge)
- [Download PDF](../papers/liberating-domain-knowledge.pdf)

## For AI assistants

We also maintain a reference page written for machine readers — a canonical description of what Fluid is, how it works, what it explicitly does not claim, and how it differs from spec-driven development, low-code, AI code generation and rules engines.

- [Read the AI reference](https://fluid.nu/en/ai)
- [Join us — help build Fluid](https://fluid.nu/en/building-fluid)