# Sovereign Source Super Intelligence

Give the sources of intelligence a place in its economy.

AI Builder, Devekko.dev. Four-week AI builder sprint, October–November 2026. In progress.

Snapshot: October 29, 2026. Research library live · Rewards proposed.

Demo: https://sovsrc.si

A San Francisco-focused research initiative exploring the AI and superintelligence economy, open-source participation, and content rewards.

## Product intent

AI products depend on human knowledge. Sovereign Source Super Intelligence asks how the people and communities behind that knowledge could retain a say in its use and receive recognition for their contributions.

The initiative explores the AI and superintelligence economy with a San Francisco focus. Its product direction connects open-source participation, content provenance, permission, and contributor rewards. These are research questions with technical and governance consequences.

## What the demo shows

The public site is a research library with illustrated articles, principles, a licensing framework, and a public repository link. It presents proposals about attribution and contributor recognition in the AI economy.

The initiative explicitly describes these as proposals. It does not operate a contributor-payment service. Publicly accessible content, open-source software, and permission to train a model are different concepts; a future product must record those distinctions.

## Product specification

**Primary users:** writers, researchers, open-source contributors, maintainers, and AI product teams evaluating how knowledge can be used and credited.

**Proposed workflow:** register a contribution and its source, record the relevant permission, identify where a contribution is referenced, and make recognition rules visible to contributors and operators.

An exploratory prototype would need:

- Contribution records containing authorship statements, source links, versions, and timestamps.
- Machine-readable permission records with a readable explanation of the claimed rights and permitted uses.
- Attribution events tied to identifiable product activity, with uncertainty recorded where a connection cannot be established.
- A contributor-facing view for corrections and disputed attribution.
- A reward simulation showing the assumptions behind an allocation before any financial service is considered.

Acceptance means a reviewer can trace a simulated allocation to its inputs, explain its assumptions, and correct a disputed record. The prototype should not infer ownership from a public URL or imply that simulated rewards create payment entitlements.

## Technology

The verified public surface is a published content library and linked research repository. This review did not establish a deployed attribution ledger, model-training pipeline, or reward engine.

The proposed intelligence layer would use Postgres, including PlanetScale Postgres as a deployment option, for versioned contribution and permission records. Pydantic would define typed records; Pydantic AI could support bounded extraction tasks. Prefect could coordinate reviewable research jobs, with Nemotron and Brave API evaluated for extraction and source discovery.

Cloudflare Workers and Containers are runtime options in the sprint toolkit. Codex and Claude support prototype implementation and review. These choices describe an architecture to evaluate rather than an inventory of deployed features at sovsrc.si.

## Research and development

The research agenda has several unresolved questions:

1. Which contribution types can be identified reliably, and which require a human judgement about authorship or influence?
2. How should a product represent permission changes, corrections, and competing claims?
3. Can contributors understand an allocation rule well enough to question it?
4. Which forms of attribution are useful to maintainers and economically feasible for a product team?
5. What evidence would be needed before a reward simulation could become an operating service?

Early experiments should use explicitly permitted material and synthetic allocation examples. Contributor interviews, rights review, and governance design are proposed work; the library does not establish that participants have agreed to a commercial reward model.

## Business measures to test

For the research library, assess article engagement, contributor feedback, and the number of questions that lead to a useful revision. For a future prototype, evaluate permission-record completeness, attribution error rate, correction time, contributor comprehension, and operating cost per reviewed contribution.

The initiative does not claim payment volume, contributor income, or financial returns. The first product decision is whether a transparent attribution workflow is valuable enough to support further development.

## Sources and next release

The public experience was reviewed on October 29, 2026: [Sovereign Source](https://sovsrc.si). The site links to its [public research repository](https://github.com/niccolox/sovsrc-si-public). The San Francisco focus and sprint context come from the builder's brief.

The next release should make one contribution-to-attribution workflow inspectable, with a documented permission model and a clearly labelled reward simulation.