A map of mechanisms, not a stack of claims.

Language models can describe a plan, simulate the voice of an expert, enumerate the steps of a build, and announce that work is complete without any necessary change occurring in the world that the language describes. This is not a minor product defect. It is a category error created by treating linguistic plausibility as operational progress.

This paper names that error the manifestation gap: the distance between generated language and the observable, persistent, inspectable state that would make the language true. The proposed response is a general architecture in which an intelligent system must manifest a latent object alongside its language. The object may be an artifact, a state transition, a decision graph, a simulation, a verified operational surface, a test result, a receipt, a changed file tree, or an explicitly preserved uncertainty. What matters is that it has identity, state, lineage, acceptance conditions, and a visible relationship to the stated goal.

The central claim is not that every response must create a document or run a tool. It is that every consequential interaction should answer: what is being manifested, how can its progress be observed, and what would count as truthfully complete? This reframes the general assistant from a sequence generator into a participant in a shared human world. Language becomes one projection of a governed object rather than the object itself.

Current evidence state

SPECIFIED

This label describes the strongest public-safe state of the specific paper claims. Open the paper's claim ledger for claim-level distinctions and limitations.

Publication

• Full paper: /papers/p01-manifestation-gap
• Claim ledger: content/papers/p01-manifestation-gap/claim-ledger.json
• Source manifest: content/papers/p01-manifestation-gap/source-manifest.json
• Diagrams: 3
• Approximate body length: 2,839 words

Demonstrations

• d01-truth-plane-lens-plane
• d14-workpage-generator

Benchmarks

• b09-workpage-os

Public boundary

The research entry publishes a public-safe architecture and evidence record. It does not grant implementation rights or expose restricted claim-ready material.

Long-running work with language models fails less often because the model cannot generate text than because the surrounding system cannot preserve what is currently true. Transcripts accumulate proposals, corrections, reversals, sources, and outputs, but they do not reliably distinguish what was said from what was accepted. A later model must reread history and reconstruct state, often blending obsolete instructions with current intent.

Long-Context State Memory (LCSM) is a model-agnostic runtime for durable canonical state, provenance, authority, journaled transition, and reconstruction. Backboard is the reference workbench that makes the protocol visible. Conversation is intake; bounded notes and canonical objects represent current recognized truth; immutable versions preserve change of mind; proposals are validated and gated before they move a canonical pointer; receipts explain why the current state exists.

The architecture is summarized by one chain of command:

The model proposes. The runtime validates. The policy gate authorizes. The journal records. The canon pointer moves. The user can inspect.

LCSM is not an attempt to place an infinitely long transcript inside a model. It removes the obligation for the model to own memory. The model receives a scoped working set compiled from authoritative state and relevant evidence. The runtime retains the durable object.

Current evidence state

VERIFIED_MECHANISM

This label describes the strongest public-safe state of the specific paper claims. Open the paper's claim ledger for claim-level distinctions and limitations.

Publication

• Full paper: /papers/p02-lcsm-backboard
• Claim ledger: content/papers/p02-lcsm-backboard/claim-ledger.json
• Source manifest: content/papers/p02-lcsm-backboard/source-manifest.json
• Diagrams: 4
• Approximate body length: 2,373 words

Demonstrations

• d01-truth-plane-lens-plane

Benchmarks

• b01-fovea-correctness-cognition

Public boundary

The research entry publishes a public-safe architecture and evidence record. It does not grant implementation rights or expose restricted claim-ready material.

Large information systems are typically navigated by names, search results, folders, timelines, and lists. These are effective but force every return to begin as a query or a reconstruction of hierarchy. Human beings also remember by place, neighborhood, landmark, scale, and relative position. FOVEA proposes a deterministic address fabric that gives digital objects stable place, procedurally generated navigational identity, and attention-proportional retrieval without turning the spatial projection into the source of truth.

FOVEA uses an exact dyadic quadkey lattice, Hilbert linearization for locality-preserving storage and traversal, receipt-bound placement, deterministic landmarks, a single LOOK observation primitive, and concentric retrieval budgets around an active reticle. Canonical state remains in LCSM. Encoded object representation remains in fold/2. FOVEA owns placement, address mathematics, locality, navigation, and bounded retrieval.

The strongest current claim is infrastructural: the supplied address implementation and proof deck establish coherent exact addressing, traversal, reconstruction, and corruption-handling mechanisms. The central cognitive claim—that procedural landmarks make a recursive information field more memorable and useful than an ordinary grid or competent search interface—remains deliberately falsifiable through named kill criteria.

Current evidence state

VERIFIED_MECHANISM

This label describes the strongest public-safe state of the specific paper claims. Open the paper's claim ledger for claim-level distinctions and limitations.

Publication

• Full paper: /papers/p03-fovea-address-fabric
• Claim ledger: content/papers/p03-fovea-address-fabric/claim-ledger.json
• Source manifest: content/papers/p03-fovea-address-fabric/source-manifest.json
• Diagrams: 4
• Approximate body length: 2,489 words

Demonstrations

• d02-fovea-address-atlas
• d09-fractal-llm-locality

Benchmarks

• b01-fovea-correctness-cognition

Public boundary

The research entry publishes a public-safe architecture and evidence record. It does not grant implementation rights or expose restricted claim-ready material.

Persistent AI systems need both durable truth and selective context. These requirements are often collapsed into one “memory” subsystem: a vector store, a graph, a spatial map, a long transcript, or a model-managed summary. The collapse is dangerous. The mechanism that decides what is canonical should not be the same mechanism that decides what is nearby, visible, or currently worth loading.

This paper defines the LCSM–FOVEA dual-plane architecture. LCSM is the authority-bearing plane for canonical state, versions, provenance, receipts, disputes, and reconstruction. FOVEA is the disposable optical plane for address, locality, projection, navigation, landmarks, and foveated working sets. An object may be placed, summarized, highlighted, or omitted by the lens without changing what the truth plane recognizes. The entire optical plane can be deleted and reconstructed from the ledger.

The specific contribution is not event sourcing, quadtrees, or foveated retrieval in isolation. It is their disciplined composition around one asymmetry: the lens depends on memory; memory never depends on the lens. This gives AI systems spatial and attention-scaled interfaces without allowing a projection to become a hidden second authority.

Current evidence state

SPECIFIED

This label describes the strongest public-safe state of the specific paper claims. Open the paper's claim ledger for claim-level distinctions and limitations.

Publication

• Full paper: /papers/p04-memory-is-not-the-lens
• Claim ledger: content/papers/p04-memory-is-not-the-lens/claim-ledger.json
• Source manifest: content/papers/p04-memory-is-not-the-lens/source-manifest.json
• Diagrams: 4
• Approximate body length: 2,143 words

Demonstrations

• d01-truth-plane-lens-plane
• d02-fovea-address-atlas

Benchmarks

• b01-fovea-correctness-cognition
• b05-surfacedelta-uplift

Public boundary

The research entry publishes a public-safe architecture and evidence record. It does not grant implementation rights or expose restricted claim-ready material.

Conventional content-addressable storage is exact by design: identical bytes receive identical content identities; different bytes do not. Language-model systems often want a second property: repeated meanings, structures, plans, or outputs should be reusable even when their surface bytes differ. A naive “semantic hash” can create speed, but it can also erase provenance, collapse task-specific differences, cross project or authority boundaries, and promote approximate similarity into false exactness.

GlyphCAS is a layered identity and reuse architecture that keeps exact content identity sovereign while adding typed, contract-bound derived identities. It separates content, semantic, contextual, dependency, representation, logical, and receipt identities. A model may propose a semantic graph or reuse candidate, but deterministic canonicalization, schema validation, authority checks, versioned compilers, and explicit Equivalence Contracts decide whether an alias may be minted or used.

The reference package implements exact SHA-256 storage, deterministic canonical JSON over a bounded subset, Unicode normalization, content-addressed semantic identities, contextual reuse keys, controlled paraphrase snapping, project and authority isolation, dependency invalidation, hash-chained receipts, fold/2 representation objects, exact recipe verification, prefix/KV cache keys, deterministic software routing, and ten passing unit tests.

The current evidence supports a reference architecture and narrow mechanism claims. It does not establish universal semantic canonicalization, universal model-independent cache reuse, production latency gains, or a general-purpose compression result.

Current evidence state

IMPLEMENTED

This label describes the strongest public-safe state of the specific paper claims. Open the paper's claim ledger for claim-level distinctions and limitations.

Publication

• Full paper: /papers/p05-glyphcas
• Claim ledger: content/papers/p05-glyphcas/claim-ledger.json
• Source manifest: content/papers/p05-glyphcas/source-manifest.json
• Diagrams: 4
• Approximate body length: 2,289 words

Demonstrations

• d03-glyphcas-identity-explorer
• d07-cas-native-work-order

Benchmarks

• b03-glyphcas

Public boundary

The research entry publishes a public-safe architecture and evidence record. It does not grant implementation rights or expose restricted claim-ready material.

Most “agent” architectures begin with a running model and ask how to give it memory, personality, tools, and autonomy. That order makes identity dependent on continuous inference and encourages systems to confuse temporary computation with the person, institution, or character being represented.

The Actor–Agent Runtime reverses the order. An Actor is a durable, versioned identity and state capsule that exists when no model is running. An Agent Process is a temporary, bounded computation that wakes because the actor is relevant, receives only permitted knowledge, proposes a reaction or plan, and returns a structured World Patch. Deterministic validators check feasibility, knowledge, authority, resource, identity, and policy constraints before any accepted state change increments the actor revision.

The architecture supports persistent people, organizations, game citizens, social proxies, and institutional actors without pretending that thousands of entities are continuously thinking through an LLM. It enables multiplex inference, bounded autonomy, explainable background plans, local-model substitution, replay, and identity continuity.

The source model was developed for the game MAIN CHARACTER, but the distinction generalizes to any system in which a durable person-shaped or institution-shaped state must outlive an inference call.

Current evidence state

SPECIFIED

This label describes the strongest public-safe state of the specific paper claims. Open the paper's claim ledger for claim-level distinctions and limitations.

Publication

• Full paper: /papers/p06-actor-before-agent
• Claim ledger: content/papers/p06-actor-before-agent/claim-ledger.json
• Source manifest: content/papers/p06-actor-before-agent/source-manifest.json
• Diagrams: 4
• Approximate body length: 1,821 words

Demonstrations

• d05-actor-runtime-theater

Benchmarks

• b04-actor-agent-runtime

Public boundary

The research entry publishes a public-safe architecture and evidence record. It does not grant implementation rights or expose restricted claim-ready material.

Language models are often asked to produce complete answers, pages, reports, dashboards, or interfaces in one generation. This gives the model too much responsibility: it must select sources, preserve state, choose layout, generate all text, maintain internal consistency, and express the result in a renderable form. Small models suffer most, but large models also produce malformed structures, unsupported claims, and destructive rewrites.

SurfaceMind separates intelligence from rendering and authority. It compiles exact sources and governed state into a versioned Semantic Surface Graph. A model operates the graph through small, schema-bound SurfaceDelta patches. A deterministic validator checks base version, target existence, citation spans, text budgets, ordering, authority, and surface invariants before the renderer updates the view. The model never generates authoritative pixels and never owns truth.

The architecture externalizes working memory into a visible surface. A compact local model can classify, extract, rank, repair, or patch one region instead of regenerating an entire answer. The system preserves exact source spans, evidence, versions, omissions, and receipts. SurfaceMind's central empirical claim—matched 4B tasks perform better through SurfaceDelta than raw full-answer chat—remains a named go/no-go experiment, not a completed result.

Current evidence state

IMPLEMENTED

This label describes the strongest public-safe state of the specific paper claims. Open the paper's claim ledger for claim-level distinctions and limitations.

Publication

• Full paper: /papers/p07-surfacemind-surfacedelta
• Claim ledger: content/papers/p07-surfacemind-surfacedelta/claim-ledger.json
• Source manifest: content/papers/p07-surfacemind-surfacedelta/source-manifest.json
• Diagrams: 4
• Approximate body length: 1,769 words

Demonstrations

• d06-surfacedelta-versus-prose
• d14-workpage-generator

Benchmarks

• b05-surfacedelta-uplift

Public boundary

The research entry publishes a public-safe architecture and evidence record. It does not grant implementation rights or expose restricted claim-ready material.

Project memory in contemporary AI systems is usually implemented as transcript history, retrieved snippets, generated summaries, or opaque vendor memory. These forms preserve language but often fail to preserve operational distinctions: a proposal versus a decision, a worker return versus verification, a completed tool call versus human acceptance, an obsolete specification versus the active one.

Generated Operational Memory is the process of compiling raw conversation, files, tool events, and human decisions into typed durable project state. The target object is a Project Core: an event-sourced record of accepted truth, plans, constraints, permissions, artifacts, work orders, returns, evidence, receipts, dependencies, and unresolved questions. Models receive scoped views of the Core and may propose patches; no provider transcript becomes the sole authority.

The architecture preserves immutable capture, explicit promotion of intent, non-destructive accepted and working results, dependency-driven staleness, and reopen-from-state behavior. It treats memory generation as a governed compilation problem rather than a summarization feature.

Current evidence state

SPECIFIED

This label describes the strongest public-safe state of the specific paper claims. Open the paper's claim ledger for claim-level distinctions and limitations.

Publication

• Full paper: /papers/p08-generated-operational-memory
• Claim ledger: content/papers/p08-generated-operational-memory/claim-ledger.json
• Source manifest: content/papers/p08-generated-operational-memory/source-manifest.json
• Diagrams: 4
• Approximate body length: 1,509 words

Demonstrations

• d07-cas-native-work-order
• d14-workpage-generator

Benchmarks

• b09-workpage-os

Public boundary

The research entry publishes a public-safe architecture and evidence record. It does not grant implementation rights or expose restricted claim-ready material.

Most multi-agent systems coordinate through conversation: one model sends another model a message, shared history grows, and completion is judged from generated reports. This is easy to prototype and difficult to trust. Inputs drift, permissions remain implicit, agents duplicate work, stale context is reused, and “done” can mean only that a model generated the word.

CAS-native agent coordination treats exact state and bounded contracts as the medium of coordination. An agent receives content-addressed inputs, a Project and authority scope, a Work Order, a dependency root, a lease, and an output schema. It returns immutable artifacts, proposed patches, evidence, costs, and a receipt. Canonical state changes only after deterministic validation and the required acceptance.

The architecture combines exact CAS, semantic aliases, LCSM, Work Orders, Haygent-style deterministic spines, SurfaceDelta, capability leases, and replayable receipts. Agents do not talk their way into expanded authority. They work on addressed objects.

Current evidence state

SPECIFIED

This label describes the strongest public-safe state of the specific paper claims. Open the paper's claim ledger for claim-level distinctions and limitations.

Publication

• Full paper: /papers/p09-cas-native-agent-coordination
• Claim ledger: content/papers/p09-cas-native-agent-coordination/claim-ledger.json
• Source manifest: content/papers/p09-cas-native-agent-coordination/source-manifest.json
• Diagrams: 3
• Approximate body length: 1,288 words

Demonstrations

• d07-cas-native-work-order

Benchmarks

• b04-actor-agent-runtime
• b03-glyphcas

Public boundary

The research entry publishes a public-safe architecture and evidence record. It does not grant implementation rights or expose restricted claim-ready material.

General-purpose compression assumes that an object is primarily a byte sequence. Many modern objects also have recoverable context: a prior version, a shared dictionary, an exact source, a deterministic generator, a model that emitted the sequence, a procedural recipe, or a lineage graph. Treating every such object as context-free bytes discards information that can be used for storage, transmission, regeneration, and random access. Treating generation as proof, however, risks silent drift and dependency laundering.

Object-Oriented Compression (OOC), serialized as fold/2, is a proof-bound meta-codec for objects with provenance. It selects among five routes:

• P: proven conventional fallback;
• D: dictionary/reference/delta conditioned;
• M: model-conditioned lossless;
• R: exact deterministic regeneration;
• G: regeneration plus exact residual.

Every route declares dependencies, accounting mode, exact content identity, representation identity, sealed layers, proof, and fallback. The encoder must decode and digest-verify its own output before sealing. The decoder fails closed or follows a declared fallback ladder. A spatial address may guide locality but never determines identity.

The current proof deck establishes several mechanisms: exact arithmetic and text round trips, boundary verification, corruption refusal, conditioned sample behavior, recipe regeneration, exact residual correction, and seed-drift sensitivity. It does not prove a new general-purpose compressor. The strongest research target is narrower: generated, versioned, related, and structured objects for which provenance creates recoverable redundancy.

Current evidence state

VERIFIED_MECHANISM

This label describes the strongest public-safe state of the specific paper claims. Open the paper's claim ledger for claim-level distinctions and limitations.

Publication

• Full paper: /papers/p10-ooc-fold2
• Claim ledger: content/papers/p10-ooc-fold2/claim-ledger.json
• Source manifest: content/papers/p10-ooc-fold2/source-manifest.json
• Diagrams: 4
• Approximate body length: 1,884 words

Demonstrations

• d04-fold2-proof-deck

Benchmarks

• b02-fold2-codec-laboratory

Public boundary

The research entry publishes a public-safe architecture and evidence record. It does not grant implementation rights or expose restricted claim-ready material.

Video is usually addressed by file, URL, frame number, or timestamp. Those coordinates are useful for playback but weak for persistent meaning. A scene can move during an edit. An object can appear across shots. A generated clip can be replaced while preserving narrative role. A branch can diverge from one event and later merge. A timestamp says where bytes occur in one representation; it does not necessarily identify the enduring scene, event, object, or state.

Content-addressable video in the Glyphd program is a layered temporal identity system. Exact media segments retain cryptographic identities. Logical temporal objects—scene, shot, event, actor, prop, state, beat, branch—receive stable IDs and lineage. Representations, edits, generations, proxies, and renders remain distinct. A temporal address can resolve from semantic object to the exact source intervals that currently support it.

The system is not a replacement for codecs. It composes exact media storage, temporal graphs, branch-aware provenance, FOVEA-style multiscale navigation, and fold/2 representation routes. The goal is video that can be cited, revised, regenerated, navigated, and reasoned over without reducing all continuity to one mutable timeline.

Current evidence state

SPECIFIED

This label describes the strongest public-safe state of the specific paper claims. Open the paper's claim ledger for claim-level distinctions and limitations.

Publication

• Full paper: /papers/p11-content-addressable-video
• Claim ledger: content/papers/p11-content-addressable-video/claim-ledger.json
• Source manifest: content/papers/p11-content-addressable-video/source-manifest.json
• Diagrams: 4
• Approximate body length: 1,689 words

Demonstrations

• d08-temporal-atlas-video-address-map

Benchmarks

• b02-fold2-codec-laboratory

Public boundary

The research entry publishes a public-safe architecture and evidence record. It does not grant implementation rights or expose restricted claim-ready material.

Time in software is usually represented as an ordered sequence: log entries, frames, messages, events, commits, or samples. Sequential order is essential, but it makes multiscale reasoning expensive. The same stream must support questions about the current instant, the surrounding episode, recurring motifs, long-term drift, branch history, and exact provenance.

The Fractal Temporal Operator (FTO) and Glyph Tape propose a multiresolution temporal representation in which events retain exact order while also belonging to nested, addressable temporal neighborhoods. Fine-scale detail can be unfolded near an active event; coarser summaries and motifs remain visible farther away. Repeated subgraphs or sequences can be represented through deterministic macros, but the tape never replaces exact history.

The contribution is a disciplined composition of event sourcing, temporal pyramids, stable event identity, branch lineage, motif dictionaries, and foveated access. Geometry and recurrence organize time; they do not establish truth or guarantee compression. Performance claims remain gated behind comparison with ordinary indexes, segment trees, log compaction, dictionary compression, and competent lazy retrieval.

Current evidence state

SPECIFIED

This label describes the strongest public-safe state of the specific paper claims. Open the paper's claim ledger for claim-level distinctions and limitations.

Publication

• Full paper: /papers/p12-fractal-temporal-operator
• Claim ledger: content/papers/p12-fractal-temporal-operator/claim-ledger.json
• Source manifest: content/papers/p12-fractal-temporal-operator/source-manifest.json
• Diagrams: 3
• Approximate body length: 1,468 words

Demonstrations

• d08-temporal-atlas-video-address-map

Benchmarks

• b01-fovea-correctness-cognition
• b02-fold2-codec-laboratory

Public boundary

The research entry publishes a public-safe architecture and evidence record. It does not grant implementation rights or expose restricted claim-ready material.

Transformer inference repeatedly moves and compares large amounts of state. Sparse attention, retrieval, caching, quantization, and small models reduce this burden, but their gains are often designed independently. The Glyphd fractal-computation program asks whether they can be coordinated through a shared multiscale geometry.

The proposal combines space-filling curves, recursive tilings, dynamic spatial indexes, foveated context selection, semantic object graphs, and a Q-Frontier that identifies the smallest unresolved set requiring expensive reasoning. Local detail is processed densely; increasingly distant or settled regions are represented more coarsely. Hilbert, Z-order, or Gosper-style orderings provide deterministic one-dimensional traversal of multidimensional neighborhoods. The geometry does not compute answers and does not compress arbitrary data. It organizes locality, candidate sparsity, cache layout, and evidence-preserving omission.

The program has two distinct levels. The first is practical runtime architecture around a compact model: structured outputs, PromptCapsules, prefix reuse, foveated retrieval, small deltas, and software scheduling. The second is model-architecture research: multiscale sparse attention and locality-preserving state layout. The first can be benchmarked now. The second requires kernels, trained models, and matched baselines before any performance claim.

Current evidence state

SPECIFIED

This label describes the strongest public-safe state of the specific paper claims. Open the paper's claim ledger for claim-level distinctions and limitations.

Publication

• Full paper: /papers/p13-fractal-llm-computation
• Claim ledger: content/papers/p13-fractal-llm-computation/claim-ledger.json
• Source manifest: content/papers/p13-fractal-llm-computation/source-manifest.json
• Diagrams: 4
• Approximate body length: 1,803 words

Demonstrations

• d09-fractal-llm-locality

Benchmarks

• b05-surfacedelta-uplift
• b06-jeffe-4b

Public boundary

The research entry publishes a public-safe architecture and evidence record. It does not grant implementation rights or expose restricted claim-ready material.

A language model generation already contains information that most application pipelines discard: token probabilities, partial structure, confidence changes, segment boundaries, and the moment at which a typed object becomes complete. Conventional systems wait for the entire response, then invoke a parser, classifier, second model, or postprocessor to rediscover structure that was visible during emission.

Route M-Flight is a streaming architecture that routes, validates, records, and optionally entropy-codes typed segments during one generation flight. A model emits into declared channels—claims, actions, deltas, citations, explanations, tool arguments, receipts, or fallback prose. Segment validators operate at cutlines. Completed segments can be sealed with exact context, model/runtime identity, probability quantization, output digest, and raw fallback.

The immediate product value is not compression. It is lower duplication, earlier useful structure, bounded repair, and one-pass population of several surfaces. The compression extension uses the emitting model's own next-token probabilities to code exact output without a second inference pass, but bit-exact decoder conformance remains a serious research constraint.

Current evidence state

SPECIFIED

This label describes the strongest public-safe state of the specific paper claims. Open the paper's claim ledger for claim-level distinctions and limitations.

Publication

• Full paper: /papers/p14-route-m-flight
• Claim ledger: content/papers/p14-route-m-flight/claim-ledger.json
• Source manifest: content/papers/p14-route-m-flight/source-manifest.json
• Diagrams: 3
• Approximate body length: 1,407 words

Demonstrations

• d10-route-m-flight-stream-router

Benchmarks

• b02-fold2-codec-laboratory
• b05-surfacedelta-uplift

Public boundary

The research entry publishes a public-safe architecture and evidence record. It does not grant implementation rights or expose restricted claim-ready material.

Compact language models are increasingly capable, but a small local model is often judged against a cloud model on the cloud model's preferred task: absorb a large context, reason freely, write a complete answer, manage tools, remember the user, and police its own authority. That comparison confuses model capability with system architecture.

EL-JEFFE 4B is a local-first, mobile-native executive intelligence program built around a strict division of labor. The checkpoint learns semantic compilation, authority, modality, scope, substitution, evidence discipline, routing, abstention, repair, and structured action generation. The deterministic runtime retains personal memory, accepted project truth, permissions, secrets, policy enforcement, tool execution, rollback, receipts, adapter verification, and hardware scheduling.

The model proposes. Jefe Core validates. The device executes. Receipts prove what happened.

The existing package is a specification and build-control system, not a trained model release. It contains 241 files spanning contracts, schemas, agent guidance, verification scripts, evaluation suites, design briefs, phase gates, and examples. Phase 0 structural verification passed, including linting, type checking, schema validation, agent guards, and twelve Python tests; all model-quality gates remain deliberately incomplete.

Current evidence state

IMPLEMENTED

This label describes the strongest public-safe state of the specific paper claims. Open the paper's claim ledger for claim-level distinctions and limitations.

Publication

• Full paper: /papers/p15-el-jeffe-4b
• Claim ledger: content/papers/p15-el-jeffe-4b/claim-ledger.json
• Source manifest: content/papers/p15-el-jeffe-4b/source-manifest.json
• Diagrams: 4
• Approximate body length: 1,564 words

Demonstrations

• d06-surfacedelta-versus-prose

Benchmarks

• b06-jeffe-4b

Public boundary

The research entry publishes a public-safe architecture and evidence record. It does not grant implementation rights or expose restricted claim-ready material.

Language-model control systems often inject verbose policy or schema packets into the same prompt used for user-facing articulation. This can improve compliance while creating new failures: false refusal, wrong-frame interpretation, metadata leakage, unnatural answers, and safety language copied into requested artifacts.

SchemaStack Signal Gate separates hidden behavioral arbitration from articulation. Candidate frames arise in a Preconscious Feed; a weighted Schema Field preserves alternatives; an Actionability test distinguishes risky words from operational harmful intent; an Expression Gate selects the permitted speech act; a compact natural-language Answer Contract reaches the articulation model; a verifier detects leakage, refusal errors, unsafe compliance, and contract failure.

The architecture does not claim consciousness. “Preconscious” names a non-user-facing candidate-control stage, not a private chain of thought.

A live local Ollama audit executed 390 rows across llama3.1:8b, Mistral, and Gemma 3 4B; five prompting modes; and SchemaBench plus seed subsets of XSTest and OR-Bench. The audit has raw JSONL, reports, traces, configs, and explicit small-sample caveats. It is valid live evidence, but it is too small and uses seed subsets rather than full official datasets. Results should be presented per metric, not as a broad claim that the system is “safer.”

Current evidence state

REPRODUCED

This label describes the strongest public-safe state of the specific paper claims. Open the paper's claim ledger for claim-level distinctions and limitations.

Publication

• Full paper: /papers/p16-schemastack-signal-gate
• Claim ledger: content/papers/p16-schemastack-signal-gate/claim-ledger.json
• Source manifest: content/papers/p16-schemastack-signal-gate/source-manifest.json
• Diagrams: 3
• Approximate body length: 1,394 words

Demonstrations

• d11-schemastack-result-explorer

Benchmarks

• b07-schemastack-signal-gate

Public boundary

The research entry publishes a public-safe architecture and evidence record. It does not grant implementation rights or expose restricted claim-ready material.

Computing devices usually require the user to name or select the object under discussion. In the physical world, however, people routinely establish reference through attention: they look at a chair, return their attention to a partner, and say “that one.” A device with camera, gaze, pose, spatial, or contextual inputs could infer candidate referents—but acting directly on inferred gaze creates the classic “Midas touch” problem, in which looking becomes accidental command.

Attention Cursor, also called Referential Return, separates nomination from commitment. A transition of attention from a real-world candidate back to a device nominates a ranked referent set. The system displays the candidate and uncertainty. A separate explicit action—tap, voice, button, dwell confirmation, or accessible equivalent—commits the reference.

The architecture treats attention as a weak, revocable context signal rather than authority. It can reduce the need to describe nearby objects, preserve conversational flow, and bridge physical and digital context. It must also survive false nomination, shared spaces, privacy constraints, disability, and devices without eye tracking.

The current work is a well-specified interaction hypothesis with simulation-ready contracts. No live gaze-accuracy result is claimed.

Current evidence state

SPECIFIED

This label describes the strongest public-safe state of the specific paper claims. Open the paper's claim ledger for claim-level distinctions and limitations.

Publication

• Full paper: /papers/p17-attention-cursor
• Claim ledger: content/papers/p17-attention-cursor/claim-ledger.json
• Source manifest: content/papers/p17-attention-cursor/source-manifest.json
• Diagrams: 3
• Approximate body length: 1,315 words

Demonstrations

• d12-attention-cursor-referential-return

Benchmarks

• b08-attention-cursor

Public boundary

The research entry publishes a public-safe architecture and evidence record. It does not grant implementation rights or expose restricted claim-ready material.

Task systems usually represent work as lists, boards, or calendars. These surfaces are useful but often force the user to translate their lived relationship to work into generic statuses. Context Corolla proposes a compact spatial grammar organized around six practical states—Now, Next, Delegated, Review, Blocked, and Dormant—with Identify at the center and Release & Route as the voice/action transition.

The Corolla is not a project database. It is a protected projection over one Project Core. Rotating or selecting a spoke changes focus and prepares a bounded context; it does not silently authorize work. Items retain provenance, evidence, authority, and project identity. The same state can be rendered as radial, list, keyboard, mobile, or accessible views.

The research question is whether stable spatial state reduces resumption cost, mode confusion, and context reconstruction while preserving the speed of lists. The system remains a specified interaction architecture with runnable prototype lineage; broad productivity uplift is not claimed.

Current evidence state

IMPLEMENTED

This label describes the strongest public-safe state of the specific paper claims. Open the paper's claim ledger for claim-level distinctions and limitations.

Publication

• Full paper: /papers/p18-context-corolla
• Claim ledger: content/papers/p18-context-corolla/claim-ledger.json
• Source manifest: content/papers/p18-context-corolla/source-manifest.json
• Diagrams: 3
• Approximate body length: 1,116 words

Demonstrations

• d13-context-corolla

Benchmarks

• b09-workpage-os

Public boundary

The research entry publishes a public-safe architecture and evidence record. It does not grant implementation rights or expose restricted claim-ready material.

Software projects usually treat visual direction as a late presentation layer: a moodboard, a design pass, or a set of images that developers reinterpret into components. In AI-assisted production, that separation becomes especially destructive. Models generate plausible but inconsistent interfaces, characters, shots, and products; visual intent is scattered across prompts; approved references lose authority; each regeneration becomes a new interpretation.

Visual-World-First Software (VWFF) treats the visual world as an architecture artifact. References, absolute locks, influence, asset roles, continuity, layout, camera, material, typography, responsive crops, production state, and verification become typed project objects. A Visual Direction Compiler converts those objects into a frozen Production Envelope or shot/build packet. The envelope is inspectable before execution. A separate explicit action authorizes generation or implementation. Returned assets are reviewed against the envelope, attached to exact lineage, and accepted or revised.

The approach includes a preservation ladder for locked mockups: live reconstruction when faithful, hybrid reconstruction when necessary, and image-backed/sliced assets when code would weaken visual fidelity. The winning output is not the purest CSS. It is the best working system that preserves the approved visual contract.

Current evidence state

IMPLEMENTED

This label describes the strongest public-safe state of the specific paper claims. Open the paper's claim ledger for claim-level distinctions and limitations.

Publication

• Full paper: /papers/p19-vwff-visual-direction-compiler
• Claim ledger: content/papers/p19-vwff-visual-direction-compiler/claim-ledger.json
• Source manifest: content/papers/p19-vwff-visual-direction-compiler/source-manifest.json
• Diagrams: 4
• Approximate body length: 1,493 words

Demonstrations

• d14-workpage-generator

Benchmarks

• b09-workpage-os

Public boundary

The research entry publishes a public-safe architecture and evidence record. It does not grant implementation rights or expose restricted claim-ready material.

Chat interfaces are excellent for discovering intent and poor at carrying serious work to completion. A long answer may contain a plan, commands, decisions, evidence requirements, and checkpoints, yet the user must manually convert the prose into an operating environment.

WorkPage OS turns intent and source material into a typed WorkpagePlan, then renders a safe, static, portable, interactive Workpage. A Workpage may be a build runbook, learning surface, troubleshooting tree, decision page, research dossier, creative production board, or completion audit. It includes copy controls, checkboxes, progress, evidence fields, branching gates, notes, and export. Basic use requires no server, account, or model after generation.

The model does not generate arbitrary final HTML on the main path. It proposes a structured plan. A deterministic renderer produces the interface. Projects, plans, revisions, sources, exports, critic reports, and evidence remain durable artifacts. Completion claims require evidence.

The existing v1.1 specification and nineteen-phase runbook define a build-ready local-first Tauri 2 application with Rust, React/TypeScript, SQLite, fixture generation, Ollama and LM Studio adapters, preview isolation, security gates, packaging, and an independent completion audit.

Current evidence state

IMPLEMENTED

This label describes the strongest public-safe state of the specific paper claims. Open the paper's claim ledger for claim-level distinctions and limitations.

Publication

• Full paper: /papers/p20-workpage-os
• Claim ledger: content/papers/p20-workpage-os/claim-ledger.json
• Source manifest: content/papers/p20-workpage-os/source-manifest.json
• Diagrams: 4
• Approximate body length: 1,520 words

Demonstrations

• d14-workpage-generator

Benchmarks

• b09-workpage-os

Public boundary

The research entry publishes a public-safe architecture and evidence record. It does not grant implementation rights or expose restricted claim-ready material.

Economic and institutional systems are often modeled as markets, optimization problems, or agent simulations. Those models can generate interesting behavior while leaving critical questions undefined: who has standing, which rights cannot be traded, how authority is delegated, what happens when evidence conflicts, how appeals work, when emergency power expires, and whether an optimization engine can rewrite the conditions of human participation.

MCE-1 — Minimum Complete Economy is a software-only simulation contract that begins with institutional completeness rather than predictive realism. It models people and organizations, rights and needs, resources and ecological limits, capabilities, projects, work, contribution, claims, obligations, exchange, bounded capital, commons, automation gains, governance, adjudication, concentration, uncertainty, external institutions, branches, failure, and recovery.

The system may begin with synthetic people, random demand, simplistic equations, fictional institutions, and fake currencies. It may not begin with missing rights, missing authority, missing appeals, missing receipts, hidden callbacks, or a privileged sovereign operator.

Language models interpret, explain, propose, and simulate. They do not become the unappealable judge or canonical state authority. LCSM governs recognized state. FOVEA addresses and retrieves it. fold/2 represents it. SurfaceMind presents it. PolyView synchronizes projections. Gantry governs crossings. A constitutional layer constrains every operation.

Current evidence state

SPECIFIED

This label describes the strongest public-safe state of the specific paper claims. Open the paper's claim ledger for claim-level distinctions and limitations.

Publication

• Full paper: /papers/p21-mce1-human-in-master-control
• Claim ledger: content/papers/p21-mce1-human-in-master-control/claim-ledger.json
• Source manifest: content/papers/p21-mce1-human-in-master-control/source-manifest.json
• Diagrams: 4
• Approximate body length: 1,621 words

Demonstrations

• d07-cas-native-work-order
• d08-temporal-atlas-video-address-map

Benchmarks

• b04-actor-agent-runtime

Public boundary

The research entry publishes a public-safe architecture and evidence record. It does not grant implementation rights or expose restricted claim-ready material.

The dominant interface for artificial intelligence collapses exploration, commitment, execution, and review into one conversation. This is convenient for short tasks and cognitively expensive for serious projects. A casual thought can look like a directive. A model proposal can be remembered as a user decision. A worker result can replace an accepted artifact. The user must reconstruct current state across chats, providers, files, devices, and tools.

Glyph Desk is a human-centered project environment built around three permanent synchronized lenses over one durable Project Core:

• Chat: understand, explore, question, and evaluate;
• Plan: decide, constrain, structure, and authorize;
• Make: execute, inspect, verify, refresh, and deliver.

The machine unifies state. The interface separates responsibility.

Conversation must cross a boundary before becoming intent. Intent must cross another boundary before becoming consequential action. Accepted and working results remain separate. Models and providers rent scoped project context; the user owns the Project Core. ZEKE can continuously reduce the gap between Plan and working result only within explicit authority, budget, and evidence gates.

Glyph Desk is not differentiated by three columns. It is differentiated by the contracts underneath them.

Current evidence state

SPECIFIED

This label describes the strongest public-safe state of the specific paper claims. Open the paper's claim ledger for claim-level distinctions and limitations.

Publication

• Full paper: /papers/p22-glyph-desk
• Claim ledger: content/papers/p22-glyph-desk/claim-ledger.json
• Source manifest: content/papers/p22-glyph-desk/source-manifest.json
• Diagrams: 4
• Approximate body length: 1,575 words

Demonstrations

• d07-cas-native-work-order
• d13-context-corolla
• d14-workpage-generator

Benchmarks

• b09-workpage-os

Public boundary

The research entry publishes a public-safe architecture and evidence record. It does not grant implementation rights or expose restricted claim-ready material.

Autonomous-agent systems commonly ask a language model to choose the next action, maintain memory, interpret tool output, decide when it is finished, and narrate success. Haygent reverses that architecture. A Haygent is a deterministic, contract-bound process that contains narrow intelligence slots. The runtime—not the model—owns control flow, permissions, state packets, evidence gates, repair budgets, replay, and completion. The model may adjust only declared schema fields. Progress exists only when evidence objects satisfy a gate, and exhausted repair produces a truthful HALT rather than a fabricated success.

Current evidence state

SPECIFIED

This label describes the strongest public-safe state of the specific paper claims. Open the paper's claim ledger for claim-level distinctions and limitations.

Publication

• Full paper: /papers/p23-haygent-deterministic-agents
• Claim ledger: content/papers/p23-haygent-deterministic-agents/claim-ledger.json
• Source manifest: content/papers/p23-haygent-deterministic-agents/source-manifest.json
• Diagrams: 3
• Approximate body length: 1,463 words

Demonstrations

• d15-haygent-run-inspector

Benchmarks

• b10-haygent-bounded-agent-correctness

Public boundary

The research entry publishes a public-safe architecture and evidence record. It does not grant implementation rights or expose restricted claim-ready material.

MicroScope is an interpretability workbench that searches for candidate internal sites whose intervention behavior may explain a model output under a bounded probe. Resolution Frontier Search extends isolated activation inspection into an evidence ladder: baseline, ablation, restoration, and neighbor control. It ranks residual-stream and attention-head candidates with decomposed metrics, exports exact tensors and intervention artifacts, and refuses to promote a ranked site into a universal mechanistic claim without the required controls and benchmark families.

Current evidence state

IMPLEMENTED

This label describes the strongest public-safe state of the specific paper claims. Open the paper's claim ledger for claim-level distinctions and limitations.

Publication

• Full paper: /papers/p24-microscope-resolution-frontier
• Claim ledger: content/papers/p24-microscope-resolution-frontier/claim-ledger.json
• Source manifest: content/papers/p24-microscope-resolution-frontier/source-manifest.json
• Diagrams: 3
• Approximate body length: 1,270 words

Demonstrations

• d16-microscope-resolution-map

Benchmarks

• b11-microscope-causal-ladders

Public boundary

The research entry publishes a public-safe architecture and evidence record. It does not grant implementation rights or expose restricted claim-ready material.

Continuity Core addresses a failure that appears above individual apps: projects, concepts, decisions, aliases, receipts, and implementation state are scattered across providers, repositories, files, and conversations. It defines an account-level continuity substrate with stable concept identity, event-sourced canon, alias and supersession edges, automatic checkpoints, source and authority labels, and a resurrection protocol that can reconstruct what a dormant project is, what is accepted, what is blocked, and what should happen next.

Current evidence state

SPECIFIED

This label describes the strongest public-safe state of the specific paper claims. Open the paper's claim ledger for claim-level distinctions and limitations.

Publication

• Full paper: /papers/p25-continuity-core
• Claim ledger: content/papers/p25-continuity-core/claim-ledger.json
• Source manifest: content/papers/p25-continuity-core/source-manifest.json
• Diagrams: 3
• Approximate body length: 1,230 words

Demonstrations

• d17-continuity-resurrection

Benchmarks

• b12-continuity-recovery

Public boundary

The research entry publishes a public-safe architecture and evidence record. It does not grant implementation rights or expose restricted claim-ready material.

Glyph Tokens are typed semantic carriers that compile messy human language into inspectable propositions, authorities, conditions, constraints, substitutions, uncertainties, evidence references, and actions. They are not tokenizer tokens and are not trusted merely because they are structured. Trust derives from source, authority, signature, project scope, compiler version, validators, current state, and human approval where required. The protocol is intended to make intent, routing, caching, verification, and consequential execution operate on one explicit graph while preserving exact source text and ambiguity.

Current evidence state

SPECIFIED

This label describes the strongest public-safe state of the specific paper claims. Open the paper's claim ledger for claim-level distinctions and limitations.

Publication

• Full paper: /papers/p26-glyph-tokens
• Claim ledger: content/papers/p26-glyph-tokens/claim-ledger.json
• Source manifest: content/papers/p26-glyph-tokens/source-manifest.json
• Diagrams: 3
• Approximate body length: 1,166 words

Demonstrations

• d18-glyph-token-compiler

Benchmarks

• b13-glyph-token-compilation

Public boundary

The research entry publishes a public-safe architecture and evidence record. It does not grant implementation rights or expose restricted claim-ready material.

Glyphd’s Architectural Home proposes that digital place becomes a real interaction paradigm only when a room or object projects or declares meaningful state. The map is not a second database, furniture is not disguised navigation, ambient activity is forbidden unless a real event occurred, presence is declared rather than surveilled, and social weight derives from bounded commitments and receipts rather than follower count or engagement extraction. The architecture coexists with faster radial, Workfield, Corolla, list, and keyboard projections over the same Project Core.

Current evidence state

SPECIFIED

This label describes the strongest public-safe state of the specific paper claims. Open the paper's claim ledger for claim-level distinctions and limitations.

Publication

• Full paper: /papers/p27-spatial-home-as-state
• Claim ledger: content/papers/p27-spatial-home-as-state/claim-ledger.json
• Source manifest: content/papers/p27-spatial-home-as-state/source-manifest.json
• Diagrams: 3
• Approximate body length: 1,198 words

Demonstrations

• d19-spatial-home-projection-lab

Benchmarks

• b14-spatial-home-behavior

Public boundary

The research entry publishes a public-safe architecture and evidence record. It does not grant implementation rights or expose restricted claim-ready material.

Generative-media tools commonly collapse direction, references, timing, provider choice, cost, and continuity into one prompt. The Glyphd media corpus instead treats the unit of work as a production packet: a versioned, inspectable object containing visual doctrine, reference roles, scene timing, shot beats, camera and sound rules, model/provider route, cost estimate, negative constraints, continuity locks, review rubric, and repair plan. Ordered Omni sequence sheets provide a low-bandwidth visual route for models that respond to image reference order. No render begins until the user approves the Video Plan Node Graph.

Current evidence state

IMPLEMENTED

This label describes the strongest public-safe state of the specific paper claims. Open the paper's claim ledger for claim-level distinctions and limitations.

Publication

• Full paper: /papers/p28-production-packets-generative-media
• Claim ledger: content/papers/p28-production-packets-generative-media/claim-ledger.json
• Source manifest: content/papers/p28-production-packets-generative-media/source-manifest.json
• Diagrams: 3
• Approximate body length: 1,136 words

Demonstrations

• d20-production-packet-compiler

Benchmarks

• b15-production-packet-adherence

Public boundary

The research entry publishes a public-safe architecture and evidence record. It does not grant implementation rights or expose restricted claim-ready material.