Wiki · The Colony & the Method
Open-Endedness Validation
How to read this page
Three ways to read this page. Precise is the document itself, exactly as it is written in the repository. Plain and Clear were written for this website to help you meet that document — they are about it. They are not it, and they are not evidence.
Eighty-four pages about the colony. Each agent is an Elixir process holding a generative model and doing inference, attached to a body that logs into a Minecraft world as an ordinary player. Around that sit the broadcast suite that films them and the runbooks that keep the whole thing running. There are typed specifications for each organ of the model, plus the world and genome specs. There are also the adversarial review personas used to attack a proposed change before it ships.
It is for the reader curious how a running system is put together and how it is held to account. The accountability half is the more distinctive. There is a lab protocol governing evidence and attribution, and a claim fence that restricts the vocabulary a claim is allowed to use. There is a public gate log. And there is a standing invitation to reproduce any verdict from the commit and the seed named in its receipt.
Start with the public read, then the lab protocol, then the falsification invitation. If you want the mathematics rather than the operations, go straight to the typed organ specs.
What it is not: a description of a mind, and not all one kind of document. A large part of this corpus is design and planning — specs marked as proposed rather than applied, organs designed but not built, plans that were later superseded — and each page states which it is. A specification is not a running system, and these pages are careful about the difference; the reader should be too. Eight documents were withheld from publication because they describe private infrastructure.
Your browser cannot switch reading levels, so the document itself is shown.
Precise — the source document
This is the document. Rendered from the repository at the commit above, with nothing rewritten for the web. A gate re-renders it on every deploy and fails the build if a single byte differs.
Purpose
Verify Validation Invariant #12 and design commitment G: the world does not terminate in a finite solved map; opening seams creates genuinely new regions with altered law vectors, reachable through niche construction (resonators).
Method
- Mechanism (deterministic): force a region's seam path with resonators and
open a seam; check a new region appears with a mutated law (regime distance
- and a valid graph.
- Gating: confirm seam readiness cannot reach threshold without resonators.
- Reachable in practice (statistical): run the Infrastructure baseline (non-omniscient, interface-constrained) with a developed body over long horizons and count unforced expansions.
- Graph coherence (soak): repeatedly open seams and verify the graph stays connected with valid laws.
Artifacts used
SP.WorldTest,SP.World.DynamicsTest,SP.SoakTest,SP.InvariantsTest(#12).scripts/evidence.exsopen-endedness section; ad-hoc developed-body run.
Result summary
Mechanism (from scripts/evidence.exs):
forced resonator path opens seam: new_region=1 regime_distance=0.709
Gating: without resonators, seam_readiness stays below the 0.8 threshold even
after 400 microsteps (SP.World.DynamicsTest); 3–4 resonators drive it across.
Reachable unforced (Infrastructure baseline, developed body, 20 seeds, 1200 ticks):
total_structures=30 total_expansions=11 expanders=6/20 max_regions=5
A scripted, non-omniscient baseline opens seams unforced in 6/20 long runs, with one world reaching 5 regions (four expansions). Each child region carries a mutated law vector (new regime), not just new coordinates.
Graph coherence (soak): SP.SoakTest opens 8 seams in sequence; all 9 regions
have valid laws and the root stays connected.
Pass/Fail
PASS. Seam expansion produces new valid regions with new law regimes, is gated behind late-stage infrastructure, keeps the graph coherent, and is reachable by an interface-constrained baseline.
Residual risks
- Unforced expansion from a seed body within a typical horizon is rare; it is a hard, late-stage capability (developed body + long horizon makes it routine). This is intended difficulty; the economy could be tuned to raise the rate.
sha256 10c8dfca9c945c97 — of the original file, so what was ingested stays checkable.
Plain — written for this website, not the source document
This report asks whether the world can keep going, rather than ending in a finite map that gets finished. The specific claim, stated as a design commitment, is that opening a seam creates a genuinely new region with different underlying law, and that this is reachable by building rather than by a special command.
Four kinds of evidence are used. A forced mechanism check, which opens a seam deliberately and shows a new region appears with a measurably different regime. A gating check, which shows the readiness never reaches the threshold without the right constructions. A statistical run, which counts how often an ordinary constrained strategy opens seams without being forced. And a soak test, which opens many seams in a row and shows the graph stays connected with valid laws.
The verdict is a pass. The residual risk is that unforced expansion from a starting body is rare, which the report calls intended difficulty.
Plain · written 2026-08-01 by claude-opus-5 · not yet checked by a person · about the document whose sha256 is 10c8dfca9c945c97
Clear — written for this website, not the source document
This report tests a design commitment: that the world does not terminate in a finite map that can be finished, and that new regions with genuinely different underlying law can be created by building rather than handed out.
It uses four kinds of evidence in increasing order of realism. First a deterministic mechanism check, in which the path to a seam is forced with the required constructions and the seam is opened. A new region appears, and the distance between its law vector and the parent's is reported as a non-zero number, so the child is a new regime rather than merely new coordinates. Second a gating check, confirming that seam readiness stays below its threshold without those constructions even after a long run, and that a few of them drive it across. Third a statistical check in which an ordinary strategy, without privileged knowledge and constrained by the same interface as any learner, runs with a developed body over long horizons, and unforced expansions are counted. A minority of runs expand, with one world reaching several regions. Fourth a soak test that opens many seams in sequence and shows every resulting region has valid law and the root stays connected.
The verdict is a pass, and the summary is careful about what it covers. Expansion produces new valid regions with new regimes, and it is gated behind late-stage building. The graph stays coherent, and it is reachable by a strategy that has no more access than a learner would.
The residual risk is stated rather than buried. Unforced expansion from a starting body within a typical horizon is rare; it becomes routine only with a developed body and a long horizon. The report calls that intended difficulty and notes the economy could be tuned to raise the rate, which is a design choice rather than a finding.
Clear · written 2026-08-01 by claude-opus-5 · not yet checked by a person · about the document whose sha256 is 10c8dfca9c945c97