Knowledge Commons Targets: What Others Can Build, and Why It Can't Be Taken Back

Public data + basement hardware + triple-copyleft licensing = permanently secured knowledge commons.

Last Updated: March 17, 2026 License: CC-BY-SA 4.0

The Structural Argument

Three properties make 🔧🦎 ecoPrimals an irreversible knowledge commons:

1. Public data only. Every spring (validation environment) experiment uses publicly available data (NCBI, PhysioNet, NOAA, USDA, PDB, arXiv). No proprietary dataset is required. Anyone can reproduce any result without institutional access.

2. Consumer hardware. Every result runs on a $500 used RTX 3090 or equivalent. No HPC allocation, no cloud account, no institutional infrastructure required. The barrier to entry is a used gaming PC.

3. Triple-copyleft licensing ( ⚖️🔓 scyBorg). Three licenses, each enforced by an independent nonprofit:

   • AGPL-3.0-or-later (code) — enforced by the Free Software Foundation
   • ORC (game mechanics) — enforced by the Open RPG Creative Foundation
   • CC-BY-SA 4.0 (documentation) — enforced by Creative Commons

   No single entity — including the creator — can revoke any license. Any derivative must share alike.

What this means in practice: if you use ecoPrimals code, your derivative must also be open-source under AGPL-3.0. If you build game mechanics on ORC content, your mechanics are also ORC. If you derive from the docs, you attribute and share alike. The commons grows monotonically — it can never shrink.

Together: the data is free, the hardware is cheap, the code is copyleft. No one can enclose what was built. No one can build on it without contributing back.

What’s Already in the Commons

Validated Science (16,695+ Checks, All Public)

Sovereign Infrastructure (107K+ Tests, 3.2M Lines of Rust)

Targets Someone Else Could Pick Up Tomorrow

These are domains where the primals + public data + consumer hardware already provide everything needed. A domain expert with K-Nome can produce validated science without building any new infrastructure.

Tier 1: Ready Now (infrastructure exists, public data available)

Tier 2: Near-Term (1–3 months of infrastructure evolution)

Tier 3: Longer-Term (6–12 months, but the path is clear)

What Makes These Targets Permanent

The Lysogeny Protocol

Every target above is secured by the lysogeny protocol (see wateringHole/LYSOGENY_PROTOCOL.md):

1. Identify proprietary gate (e.g., AlphaFold requires Google Cloud)
2. Trace underlying math to published open research (Anderson 1958, AF2 primitives = GEMM + attention)
3. Implement from first principles under AGPL-3.0
4. Cross-validate across domains (proves generality, not domain-specific IP)
5. Document provenance chain (published paper → Python baseline → Rust → GPU → validated)
6. Publish and wait
7. Adoption lyses the proprietary gate

This is area denial, not competition. Every prospective customer who finds the open alternative is a customer the proprietary vendor never acquires. The ground contamination is permanent because AGPL-3.0 is irrevocable.

The Three-Lock Guarantee

No single entity controls all three locks. The creator cannot revoke them. A corporation cannot acquire them. A government cannot classify them (the math is published, the data is public, the code is AGPL).

Why Public Data Matters

Every 🔧🦎 ecoPrimals experiment uses data that is:

• Publicly deposited (NCBI, NOAA, USDA, PhysioNet, PDB, arXiv)
• Independently accessible (no institutional login, no API key)
• Independently verifiable (anyone can download the same data)

This means: even if every 🔧🦎 ecoPrimals repository were deleted tomorrow, anyone with the published papers and the public data could rebuild the entire validation layer from scratch. The knowledge is permanent because the evidence is permanent.

The Velocity Argument: What 10 More Months Looks Like

The 🔧🦎 ecoPrimals project produced 20,000+ checks in ~10 months. The velocity is accelerating (12 checks/day in Week 1 → 1,399 checks/day in Week 3). Extrapolating conservatively:

Each check is a validated, reproducible scientific result in the permanent commons. The commons grows faster than any single entity can enclose it.

For Someone Considering Contributing

What You Need

1. Domain expertise — K-Nome works because the human knows the science. A microbiologist reproducing antibiotic resistance papers. A soil scientist reproducing no-till studies. An immunologist reproducing cytokine data. Your expertise is the selective pressure.

2. Rust — rustup.rs, 5 minutes. No prior Rust experience required (K-Nome handles the implementation).

3. A GPU — Any Vulkan-capable card. A used RTX 2070 ($150) is sufficient for most science workloads. An RTX 3090 ($500 used) handles everything including lattice QCD.

4. Cursor IDE — The K-Nome tool. One tool, one human-AI relationship.

What You Produce

A validated, reproducible implementation of published science in your domain. Runs on any hardware. Independent of any institution. Published under AGPL-3.0. Permanently in the commons.

What Returns to You

Attribution through 🍯🌾 sweetGrass provenance braids. Every contribution is cryptographically attributed to the contributor. Every derivative that builds on your work traces back to you. CC-BY-SA 4.0 requires attribution on all derivatives of documentation. AGPL-3.0 requires source availability on all derivatives of code.

Your work stays yours. The commons uses it. Derivatives credit you. Forever.

The Knowledge Commons vs The Proprietary Model

The question is not whether sovereign scientific computing is possible. It is demonstrated. The question is how fast the commons grows. Every domain expert who picks up a primal and targets their own literature expands the commons by another validated domain.

The spore print is the record. The commons is the organism. It grows from wherever it lands.

⚖️🔓 scyBorg licensing: wateringHole/SCYBORG_PROVENANCE_TRIO_GUIDANCE.md
Lysogeny protocol: wateringHole/LYSOGENY_PROTOCOL.md
Spring repositories: github.com/syntheticChemistry/