The Bridge Between Humanity and What’s Coming.
A source-aware intelligence system tracking technological, scientific, economic, workforce, regulatory, infrastructure, and human-transition signals.
Current verified registry snapshot — July 28, 2026: 111 configured intelligence areas and 650 configured source entries. Registry counts evolve as sources are reviewed, repaired, added, or retired.
This repository is the public feed layer of the 4th Wave AI Intelligence Network. It converts reviewed source registries into machine-readable RSS, Atom, and JSON feeds that can support research, education, publishing, forecasting, and human-centered decision support.
The central editorial question is:
What does this mean for humanity?
The registry spans interconnected families rather than a handful of isolated topics:
- AI and agents — models, agentic systems, safety, evaluation, governance, and key builders.
- Robotics and autonomy — industrial systems, humanoids, cobots, mobility, and field deployment.
- Quantum and advanced computing — quantum systems, post-quantum security, semiconductors, edge systems, and new compute architectures.
- Biotechnology and medicine — health, bio-computation, reproductive technology, neuroscience, biosurveillance, and research integrity.
- Future of work and education — labor transition, skills, legal and healthcare transformation, and AI-native learning.
- Energy and infrastructure — grids, water, thermal systems, critical minerals, compute habitat, and infrastructure chokepoints.
- Space and planetary systems — space technology, planetary warning, Earth dynamics, ocean systems, and national research networks.
- Law, policy, security, and governance — AI oversight, courts, standards, liability, identity, provenance, and conflict monitoring.
- Human continuity and societal transition — demographics, monetary systems, resilience, access, opportunity, and the human effect.
The production pipeline currently includes:
- a complete registry rebuild every six hours;
- one concurrency lock shared by workflows that write generated files;
- a disabled legacy partial writer that cannot replace the global directory;
- registry, Python, generated-feed, and index validation;
- automatic workflow-failure issues with recovery closure;
- weekly archives and controlled workflow-history pruning.
A separate draft pull request, PR #10, adds deeper feed-health observability, source-level diagnostics, critical-lane gates, deterministic compression, and reviewed repair overlays. It remains unmerged until final review and explicit human approval.
A green workflow is not treated as proof that every lane is useful. Trust requires multiple layers:
- Registry validity — configuration is structurally valid.
- Runner reachability — sources respond from GitHub-hosted runners.
- Collection productivity — the current run discovers usable candidate links.
- Output validity — RSS, Atom, and JSON outputs parse correctly.
- Freshness — the lane has produced substantive recent material or is explicitly marked stale.
- Product quality — output should contain articles, publications, releases, or records rather than navigation and pagination pages.
- Human authority — no repair is treated as approved or merged without explicit review.
Read the full methodology in TRUST.md.
flowchart LR
A[Reviewed source registry] --> B[GitHub Actions collector]
B --> C[RSS / Atom / JSON feeds]
B --> D[Validation and health evidence]
C --> E[GitHub Pages and downstream systems]
D --> F[Warnings, critical gates, and repair work]
F --> A
H[Human approval] --> A
H --> F
See docs/ARCHITECTURE.md for the operational design and docs/OPERATIONS.md for safe maintenance procedures.
The generated directory is maintained in index.html. Individual lanes normally expose three formats:
- RSS:
feeds/<lane>.xml - Atom:
feeds/<lane>.atom.xml - JSON Feed:
feeds/<lane>.json
Examples:
Generated-file presence alone is not a health guarantee. Consult the operating evidence and known limitations before relying on a lane for consequential decisions.
| Status | Meaning |
|---|---|
| Healthy | Current run produced useful output and no configured source failed. |
| Degraded | Current output exists, but one or more sources failed, were blocked, or yielded no usable links. |
| Stale last-known-good | Existing useful output is preserved because the current run was nonproductive. |
| Empty | No usable output is available for a noncritical lane. |
| Critical stale / empty | A protected lane has crossed its failure threshold and should block false-green completion. |
| Quality warning | Output exists but may be dominated by weak, duplicate, index, or pagination entries. |
Source proposals and code changes must be evidence-based. A proposed source should identify:
- the correct intelligence lane;
- the source owner and authority type;
- the collection URL and mode;
- GitHub-runner reachability;
- parser compatibility;
- expected item yield;
- duplication and contamination risks;
- an explicit falsifier.
Read CONTRIBUTING.md before opening a pull request.
Do not report security-sensitive findings in public issues. Follow SECURITY.md.
4th Wave AI uses automation for collection and testing, not final authority. Branches, pull requests, tests, rollback paths, and explicit approval remain part of the operating model. Humans remain in charge.
This repository uses the 4th Wave AI Business Source License with a rolling change date to Apache License 2.0. Review the repository license files before reuse or contribution.