Reachability Labs measuring reachable future under commitment
Evidence hub

Evidence and public references

This is the clean public evidence surface: the current AI paper and archive, the ReachScan software, the SAT calibration benchmark, supporting materials, public code, and the shortest references that substantiate the core claims.

Use this page when you want the papers, the benchmark, the artifact ledger, and the verification tools in one place.

Program evidence by maturity level.

The public evidence is role-labeled: the current AI reach-scan result, the oracle-backed SAT calibration, graph-coloring transfer, an active adapter foothold, and formal objects that preserve the measurement contract.

Current result: AI reach-scans

The reasoning lane measures answer foreclosure from committed AI prefixes. The current RC18 cycle includes Case A structural morphology, Case B source-conditioned diagnosticity, D7 prospective replication, token-coordinate localization, reach-entropy morphology, and controlled reopening by corrective splice. See AI evidence ↓

Oracle-backed calibration: random 3-SAT

Measured constructive accessibility transition at n=1000, confirmed finite-asymptote prediction, oracle bridge, rewind depth, and late visible failure.

Transfer calibration: graph coloring

A second oracle-backed CSP shows process-side collapse inside an alive landscape, with a different morphology from 3-SAT.

Active adapter: scheduling

Scheduling is an internal adapter foothold where forward construction commits irreversibly. It should be read as an extension lane, not a completed public result.

The research-facing benchmark and public proof tools.

The interactive SAT benchmark, the artifact ledger, and the verification tool live here so the homepage does not have to carry the full evidence surface.

Interactive benchmark view: constructive success, existence-side reference, and the measured gap.
Constructive Accessibility from Committed Prefixes in Random 3-SAT
Michael Richard Nothem · Reachability Labs

Nine system sizes through n = 1,000 anchor the SAT calibration result. Oracle-verified trap analysis, bounded rewind, trajectory-space diagnostics, a confirmed scaling prediction, and a second combinatorial domain all support the same deeper claim: the goal can remain valid after the process has lost the route.

3-SAT predecessor preprint

One place for the core public references.

Use this page for the public references: the AI paper and archive, the 3-SAT predecessor line, supporting materials, public code, and the primer.

Published — AI reach-scans

ReachScan · answer foreclosure (RC18 / v0.3.5)

The AI reach-scan measurement core and its published evidence: the current v1.1-RC18 core article, the v1.1-RC18 evidence and reproducibility archive, and the v0.3.5 software — committed-prefix future fields, projection/target contracts, provenance receipts, source-conditioned separation, and a HuggingFace adapter for token-level model access.

Full Technical Report v1.1-RC18 DOI 10.5281/zenodo.21249725 is reserved; the deposit is in progress.

SAT calibration (predecessor line)

Constructive Accessibility from Committed Prefixes in Random 3-SAT

The first resolved calibration benchmark: a measured constructive accessibility boundary, late visible failure, exact-prefix deadness, trap depth, and strong within-instance path dependence.

SAT predecessor

Constructive Accessibility Instrument

The oracle-backed SAT instrument and release track. This is real predecessor work and calibration infrastructure, not the current AI software repo.

Supplement and validation

Supporting materials and large-n diagnostics

Finite-size scaling, hazard diagnostics, bridge and rewind analyses, robustness controls, and the transfer calibration lane sit here as the public understructure for the SAT calibration result.

Transfer calibration

Graph coloring as a second oracle-backed adapter

The graph-coloring lane shows a process-side collapse inside a still-alive colorability landscape. It is strong evidence that the measurement contract travels, but it is not yet claimed as a full mechanistic twin of the 3-SAT calibration.

The shortest evidence summary

A process-indexed boundary exists.

The SAT calibration benchmark resolves a constructive accessibility transition well below the classical satisfiability threshold.

Failure has a trajectory-space fingerprint.

Late visible collapse, bridge-verified deadness, rewind depth, and strong path dependence all support the same route-side interpretation.

The measurement contract transfers.

A second oracle-backed combinatorial domain already shows the same deeper process-versus-landscape split with a different surface morphology.

The instrument reads structure without an oracle.

A first reasoning-substrate case study shows committed-prefix futures confined to a wrong arithmetic family that excludes the target. The signal is structured, finite-M, and model/policy-indexed; it is not a non-extendability certificate.

Current AI reach-scan frontier.

The reasoning lane has a full measurement cycle.

A reasoning trace is a path of commitments. The reasoning lane measures the future field induced by a committed prefix — rollouts continued from the exact state, with the resulting answer distribution treated as the observable. The current paper extends that single-case target-532 morphology into source-conditioned diagnosticity, prospective replication, token localization, and reopening tests.

Case A

Structural morphology on the target-532 floor-sum trace

Case A shows upstream wrongness and staged tightening. The prompt-root field is already biased toward the target-excluding family 8ℤ; commitment depletes target-compatible mass; the answer field later resolves onto the dominant wrong atom 56 while surface reasoning remains diverse. This is the structural prelude, not the whole current evidence stack.

RC18 measurement cycle

Source diagnosticity, D7 replication, localization, and reopening

Case B compares correct-source and wrong-source prefixes. On D9, the deepest-cut gap reaches +0.732 with disjoint-seed reconfirmation; the prospectively selected D7 replication has late-band separation +0.802 with 95% CI [+0.695, +0.894]. Eighteen wrong traces localize foreclosure in token coordinates, and five corrective-splice tests show controlled reopening under narrow conditions.

Claim status

The current paper supports state-indexed, finite-budget claims about answer foreclosure under declared model, sampler, budget, extraction, validity, and target contracts. It does not establish cross-model portability, cross-task generality, an internal mechanism account, a deployable monitor, or a universal repair policy.

The reasoning substrate has no exact oracle analogous to the SAT bridge check. Reasoning-side foreclosure is operational evidence about a measured future field, not a logical non-extendability certificate.

Public artifacts, versions, and verification tools.

This is the control surface for public references and downloadable artifacts. Use it when you need the ledger, the mirror list, or in-browser SHA-256 verification.

Where to start

Artifact ledger

Public artifacts, versions, and mirrors. The site is the hub — OSF, GitHub, Zenodo, and later arXiv are the public destinations.

Software release version, release link, and SHA-256 hashes are read from /artifacts.json. Deploy this file from GitHub release automation instead of editing version strings by hand.

ArtifactVersionStatusPrimaryMirrors
Core paper
Existence Is Not Reachability — answer foreclosure from committed AI prefixes
v1.1-RC18Published concept DOI Core article concept DOI RC18 version DOI · AI page
Evidence & reproducibility archive
Reach-Scan evidence, raw rows, constructors, manifests, and reproduction code
v1.1-RC18Published concept DOI Evidence archive concept DOI RC18 version DOI · ReachScan repo
Full Technical Report
Full technical report and audit record for the ReachScan publication family
v1.1-RC18Deposit in progress Reserved DOI: 10.5281/zenodo.21249725 Link after Zenodo record resolves
ReachScan software
Committed-prefix future-field measurement core for AI reach-scans
v0.3.5Public software DOI GitHub repository Zenodo software DOI
3-SAT calibration paper
Constructive Accessibility from Committed Prefixes in Random 3-SAT (predecessor line)
v1.0Public OSF Zenodo · arXiv forthcoming
SAT predecessor demo
Constructive Accessibility Instrument (oracle-backed predecessor implementation)
v0.3.2Public after tag GitHub Release Repository · Zenodo release forthcoming
Interactive figures
Epistemic, process diagnostics, scientific benchmark, and trajectory atlas figures
v4–v6Public Website Website bundle
Supplement
Extended data, validation runs, coloring adapter
v1.0Public OSF Zenodo
Primer
The Builder's View — a plain-language introduction to constructive accessibility
v4.0Public Website
Primer (audio)
The Builder's View — spoken narration (~20 min)
v1.0Public Website

Verify a download (SHA-256)

Pick a file you downloaded. This computes SHA-256 in-browser and checks it against the expected hash. Nothing is uploaded.

Concept first

If you need the shortest conceptual path before looking at papers, start with the concept pages.

Constructive accessibility

Plain language first

If you want the least technical entry point, start with the primer.

The Builder's View

Verified? Start intake.

If the evidence above maps to the failure mode you are already seeing, the next step is the intake form. It is fine if some answers are still rough.

Open the intake form →

Code path

Use ReachScan for the AI measurement core. Use the SAT repository for the earlier constructive-accessibility instrument.

ReachScan repository SAT predecessor