ORDIVON
Menu

Now · 22 August 2026

What can be used, what has been implemented, what remains research, and what became historical.

This is a dated editorial synthesis across the project family. Project repositories remain authoritative for exact capability, operation, release, and research evidence.

Current synthesis

The public map now distinguishes capability from prototype, product, research, and history.

01 / Durable work

Host contracted; Harness and Runtime became more explicit about what they do not own.

Operational with schema-v5 Journal/CAS and exactly six public MCP Tools. Host is intentionally narrower than its historical designs: it is not a universal coordination ontology, scheduler, executor, Runtime client, source engine, cognition host, or cross-owner truth service. Operational pre-1.0. One independent Harness Journal/CAS owns bounded Run continuity; Agent-selected cognition is distinct from history, current action authority is request-bound, ambiguous Provider/Tool effects are recovered conservatively, and HarnessAgentRun is the supported normal Python composition surface. Generic Memory/RAG, semantic ranking, Provider-opaque continuation, Strategy scheduling, and domain completion remain outside Harness. Operational for owner-trusted Linux engineering work with an experimental configured Windows-native backend. Exact execution, immutable inputs, observation, cancellation, recovery, lifecycle/reclamation, and bounded evidence are current capabilities. Hostile multi-tenant isolation, arbitrary external-effect truth, and semantic Task completion remain outside Runtime.

02 / Reality-facing systems

Game and Finance now expose materially different consequences instead of serving as demonstrations of one platform.

Station Zero v2 remains the only registered current product. Station Zero v3 remains an unregistered reference/product target, but its current Product Value evidence is materially stronger than the earlier playable-proof snapshot: owner-native machine pressure can now falsify and improve product behavior, with Core completion at 5/6 deterministic profiles after the latest correction. No next product or G0 direction has been selected. Human evidence remains necessary for unresolved human-specific experience claims, not as the default next gate for every Game change. Primary Agent Surface v14 exposes seven semantic operations: observe, research, decide, execute, reconcile, performance, and owner. Agent Context v21 now distinguishes current consumer-bound obligations/current execution work from replayable historical detail. Current Finance can observe and compose real capital state, form/revalidate owner-bound decisions, preserve effect identity/reconciliation, and evaluate performance on explicit bases. External financial submission remains a separate fail-closed deployment boundary.

03 / Mediation and external relation

Media is now the top-level mediation owner; World remains a narrow cross-owner relation boundary.

Media Phase 1 (M0–M4) is closed in a stable-owner, consumer-pressure-driven posture. MF0–MF9 are frozen, MF10 is unadmitted, OMPC-v0 is the shared reference/conformance baseline, and Studio remains active. New shared implementation is admitted only if multiple real consumers demonstrate duplicated mediation machinery or a concrete falsifier. World 0.4 currently supports provider effect binding/reconciliation plus Resource, Message, and Entity trajectories and bounded WorldTaskInspector projection. Historical research remains retrievable but deleted shared layers do not reactivate by recency; new abstractions require a named current workload that direct Host + owner/provider composition cannot represent cleanly.

04 / Research and knowledge

Computing, Human, and Security retain bounded research while mature results descend into Knowledge or owner-local practice.

Current Computing separates Research, reusable Knowledge, compact Core, owner-native Projects, and Git history. SCD and Computational Possibility remain independent semantic owners inside the repository. The contracted Computer has no obligation to keep an active research line by inertia: new shared structure must survive owner placement, mature-baseline, deletion, and materially different consumer tests. HUMAN-ECON-001 remains a completed conditional reference answer; HUMAN-AI-001 is active under an external-evidence-first contraction. Human Foundations currently end at HF23 with HF24 unadmitted, and Deep Foundations end at HD9 with no preselected HD10. New work is admitted by decision-changing problems or falsifiers, not by numbering or taxonomy completion. Current Security keeps the accepted Contest/Range/Evaluation/KVM/ResearchCorpus and model-backed Actor substrate, but the default Agent surface is now the thin ordinary view rather than research chronology. Read-only preflight exposes only mechanically eligible owner operations for the current turn, while semantic choice and consequence authority remain separate. Many research-only console affordances and historical runners have been retired from the wheel or archived as reproduction apparatus. No generic Campaign/Organization, SIEM/EDR, Trust/Reputation, planner/registry/compiler/gateway, global freshness service, or cross-owner capability compiler is admitted.

Under real pressure

Experiments that can still change a boundary or delete an idea.

These questions remain active because another workload, failure, or comparison can change where responsibility belongs.

Ordivon Computing1 supporting publications

When should a finite intelligence expand its interface to Reality instead of only reasoning harder?

Wait for a natural owner-level failure that satisfies an already-frozen strict burden: a real certificate/language or grounded-meta-closure case where the current closure is prospectively explicit and the repair cannot be reconstructed as ordinary composition inside the admitted substrate. Do not manufacture E1b/E2 merely because apparatus exists.

Ordivon Computing4 supporting publications

When should an Agent act, wait, observe, refuse, reconcile, or request a decision?

Repeat evidence-conditioned act/abstain trajectories in materially different domains and with state changes that make a previously impossible action become expressible; retain new non-action state only if simpler owner-native evidence cannot reconstruct the same decision.

Ordivon Finance5 supporting publications

Can Agent capital judgment remain separate from financial effect authority?

Use the current owner-capital and consumer-bound context to evaluate real opportunities without relabeling missing return/history evidence as zero risk or positive contribution. Keep external submission separate until its selected deployment profile, current credentials/signer, one-shot dispatch, and reconciliation prerequisites are explicitly admitted.

Ordivon Game0 supporting publications

Which game form deserves the next Ordivon Game product—and where do Agents add Player Value?

Continue from the cheapest claim-preserving pressure for the actual current question: use machine/Agent counterfactual and OOS evidence for structural product defects, then acquire targeted player or market evidence only when the residual claim is specifically comprehension, tension, preference, replay desire, or broader demand. Do not make one Human session the default gate or population oracle.

Ordivon Media1 supporting publications

How do we know a creative improvement is not just a searched winner or model preference?

Change observer class/provider and repeat frozen interventions across new artifact worlds. Introduce human evidence only when making a human comprehension, preference, trust, memory, or behavior claim.

Ordivon Web2 supporting publications

Can the public site explain changing research judgment without becoming a second fact system?

Finish bounded question-state adjudication, then run the full publication/build/budget/browser suite and compare the deployed source commit with current main. After that, test whether fresh readers can recover purpose, owner boundaries, current maturity, evidence limits, and the next useful destination. For high-change manual owners, prefer an owner-supplied public-authority envelope if and when that owner admits one; Web must not manufacture it on their behalf.

Dated evidence

Read the arguments and reports that changed the judgment.

A current page summarizes today. A publication preserves the evidence, limits, and interpretation at its publication or revision date.

Research note

The Hardware Interface Is Not the Experiment

What Anthropic's Model Hardware Standard, Universal Robots, Genentech, QuEra, LAP, NIST, and NASA reveal about the distinct engineering questions that remain after an Agent can operate physical equipment.

Research report

The Decision Was Valid. The Portfolio Changed.

Why Decision validity and commitment currentness are different facts, using Ordivon Finance's exact portfolio-baseline fence for state-changing proposals.

Research report

A Better-Looking Result Can Still Be Noise

What Studio R4–R6 learned about richer perception, grounded meaning, search inflation, real Web encounters, and why no single score deserves creative authority.

Research report

Correctness Is Not Isolation

How Finance C2-L separated effect correctness from credential and administrator isolation, falsified hostile-root isolation on one machine, and retained a disabled single-machine effect-correctness profile without authorizing live trading.

Current project map

10 public projects with explicit maturity and boundaries.

Open the project page for orientation; follow the repository for exact implementation, operation, release, or research authority.

implemented prototype

Ordivon Host

Operational with schema-v5 Journal/CAS and exactly six public MCP Tools. Host is intentionally narrower than its historical designs: it is not a universal coordination ontology, scheduler, executor, Runtime client, source engine, cognition host, or cross-owner truth service.

Open project ↗
implemented prototype

Ordivon Harness

Operational pre-1.0. One independent Harness Journal/CAS owns bounded Run continuity; Agent-selected cognition is distinct from history, current action authority is request-bound, ambiguous Provider/Tool effects are recovered conservatively, and HarnessAgentRun is the supported normal Python composition surface. Generic Memory/RAG, semantic ranking, Provider-opaque continuation, Strategy scheduling, and domain completion remain outside Harness.

Open project ↗
usable now

Ordivon Runtime

Operational for owner-trusted Linux engineering work with an experimental configured Windows-native backend. Exact execution, immutable inputs, observation, cancellation, recovery, lifecycle/reclamation, and bounded evidence are current capabilities. Hostile multi-tenant isolation, arbitrary external-effect truth, and semantic Task completion remain outside Runtime.

Open project ↗
playable now

Ordivon Game

Station Zero v2 remains the only registered current product. Station Zero v3 remains an unregistered reference/product target, but its current Product Value evidence is materially stronger than the earlier playable-proof snapshot: owner-native machine pressure can now falsify and improve product behavior, with Core completion at 5/6 deterministic profiles after the latest correction. No next product or G0 direction has been selected. Human evidence remains necessary for unresolved human-specific experience claims, not as the default next gate for every Game change.

Open project ↗
usable now

Ordivon World

World 0.4 currently supports provider effect binding/reconciliation plus Resource, Message, and Entity trajectories and bounded WorldTaskInspector projection. Historical research remains retrievable but deleted shared layers do not reactivate by recency; new abstractions require a named current workload that direct Host + owner/provider composition cannot represent cleanly.

Open project ↗
implemented prototype

Ordivon Finance

Primary Agent Surface v14 exposes seven semantic operations: observe, research, decide, execute, reconcile, performance, and owner. Agent Context v21 now distinguishes current consumer-bound obligations/current execution work from replayable historical detail. Current Finance can observe and compose real capital state, form/revalidate owner-bound decisions, preserve effect identity/reconciliation, and evaluate performance on explicit bases. External financial submission remains a separate fail-closed deployment boundary.

Open project ↗
research

Ordivon Human

HUMAN-ECON-001 remains a completed conditional reference answer; HUMAN-AI-001 is active under an external-evidence-first contraction. Human Foundations currently end at HF23 with HF24 unadmitted, and Deep Foundations end at HD9 with no preselected HD10. New work is admitted by decision-changing problems or falsifiers, not by numbering or taxonomy completion.

Open project ↗
research

Ordivon Security

Current Security keeps the accepted Contest/Range/Evaluation/KVM/ResearchCorpus and model-backed Actor substrate, but the default Agent surface is now the thin ordinary view rather than research chronology. Read-only preflight exposes only mechanically eligible owner operations for the current turn, while semantic choice and consequence authority remain separate. Many research-only console affordances and historical runners have been retired from the wheel or archived as reproduction apparatus. No generic Campaign/Organization, SIEM/EDR, Trust/Reputation, planner/registry/compiler/gateway, global freshness service, or cross-owner capability compiler is admitted.

Open project ↗
research

Ordivon Media

Media Phase 1 (M0–M4) is closed in a stable-owner, consumer-pressure-driven posture. MF0–MF9 are frozen, MF10 is unadmitted, OMPC-v0 is the shared reference/conformance baseline, and Studio remains active. New shared implementation is admitted only if multiple real consumers demonstrate duplicated mediation machinery or a concrete falsifier.

Open project ↗
research

Ordivon Computing

Current Computing separates Research, reusable Knowledge, compact Core, owner-native Projects, and Git history. SCD and Computational Possibility remain independent semantic owners inside the repository. The contracted Computer has no obligation to keep an active research line by inertia: new shared structure must survive owner placement, mature-baseline, deletion, and materially different consumer tests.

Open project ↗

Judgments selected for review

Questions carrying the current architectural and research pressure.

This authored set highlights judgments that currently matter; date order does not decide importance.

answered

Which Agent-era world-model laws survive historical out-of-distribution pressure?

Answered for the HD0–HD9 programme. The frozen 1760–2026 campaign closed with 320 trajectories and 96 Deep Anchors across eight chronological waves. Cross-era recurrence survived only as a typed seven-family structural map plus explicit demotions, merges, replay negatives, and measurement/currentness results; the evidence did not justify ten universal success invariants or eight universal innovation laws.

answered

Does the caller-neutral Harness boundary remain valuable across different workloads and execution strategies?

Answered for the current pre-1.0 product boundary. Harness is now an independent, Host-free, caller-neutral durable cognitive execution substrate with its own Journal/CAS and supported HarnessAgentRun path. Real consumer pressure retained exact Run/turn capability separation, WorkingSet/WorkingView versus canonical history, Provider/Tool UNKNOWN recovery, structured completion as proposal rather than domain truth, and bounded capability composition; other candidate machinery such as a universal planner, generic Memory/RAG, Provider-continuation core, capability registry, and automatic knowledge/promotion services was deleted or kept local. The latest research tournament reports no next branch selected: remaining semantic frontiers are blocked on new consumers, owner transitions, or implementation diversity rather than an unresolved proof that Harness has any reusable boundary at all.

reframed

Which game structures become possible only when Agents persist and act through one World history?

Reframed after GX1 and the later Game direction work. GX1 showed materially different reachable futures and separated strategy existence from model realization; the model still failed the frozen Core-carrier continuation while a deterministic continuous-carrier policy could succeed, and a generic structured primaryObjective prototype was deleted after presentation-order sensitivity. Those results remain valid research pressure, but the programme no longer assumes that proving an Agent-native mechanic selects the next game.

answered

Which remaining friction belongs above Runtime rather than becoming another primitive?

Answered for the current P0–P5, HP0, and C1–C10 evidence horizon. Storage pressure produced an owner/retention correction rather than a generic cleanup service, and later operational-realization pressure kept currentness, retry, evidence scope, identity, authority, retention, and reclamation separable without reopening Runtime Foundations.

answered

When does a useful research constitution earn default Agent context?

Answered for the current Quant Bedrock default-context claim. Constitution-only looked less bad than control on QB3b but remained negative; the unchanged treatment became materially worse than control on the second sealed QB3c carrier and used about 27.5% more Provider tokens. Mandatory injection has not earned default status.

answered

Which Provider-required state must survive a Tool turn without becoming Agent cognition?

Answered within Harness X2. A live Gemini falsifier returned 200 with the exact continuation and 400 INVALID_ARGUMENT when only the signature was stripped. A 940-byte continuation then survived crash/reopen, remained hidden from Agent messages, completed LIVE_CONTINUATION_OK, and caused zero Runtime redispatch. ProviderToolContinuation was the only new primitive earned across five pressured frontiers.

answered

Does canonical evidence order make an Agent understand the evidence better?

Answered for P11. Six unordered datasets collapsed 720 raw permutation digests to one stable identity, but after invalidating a truncated Tool-wire campaign and clarifying an ambiguous evaluator target, raw presentation scored 44/48 strict versus 42/48 stable list and 43/48 identity map. No semantic normalization advantage survived.

testing

When should an Agent act, wait, observe, refuse, reconcile, or request a decision?

Testing with a stronger real-domain non-action case. Finance QB6 let one Agent request exact read-only evidence rather than infer missing execution facts, then update from more_research to no_op when the current minimum contract exceeded both the frozen 10% target and existing 25% canary bound; the branch ended before simulation, Decision, Proposal, or financial write.

testing

Can Agent capital judgment remain separate from financial effect authority?

Testing at a later boundary than the original QB experiments. Finance now preserves real FinancialResources independently of current risky positions, aggregates component portfolio state with explicit FX/temporal evidence, binds owner-managed capital through replayable CapitalState/Exposure context, and blocks both stale state-changing Decisions and unsupported owner-contribution shortcuts. Context v21 routes only work pulled by current Finance consumers into authoritative obligations while external financial submission remains independently fail-closed.

answered

When does valid evidence stop being current enough to govern a decision?

Answered for the current evidence horizon. Security EC1 separated byte identity from semantic applicability; World independently separates historical occurrence from current relation/Presence; Web source-bound publication distinguishes exact owner revision from current editorial judgment and exposed stale static/test consumers; Finance further showed that an already coherent state-changing Decision loses Proposal authority when its PortfolioSnapshot baseline is no longer current. Computing Knowledge now records both HistoricalValidity != Currentness and AdmissionCurrentness != CommitmentCurrentness. These cases support owner-native revalidation and have not forced a generic Freshness, impact-graph, or automatic rebase service.

testing

How do we know a creative improvement is not just a searched winner or model preference?

Testing after R4–R6. Rich perception earned medium-local equipment, grounding proved more stable than fine labels, and R6 showed a 5,000-candidate zero-effect search could manufacture an impressive winner that failed stronger evidence. One reveal-order effect replicated directionally on pristine content for one Agent-observer class.

answered

Which World responsibilities still earn permanent ownership after deletion pressure?

Answered for the current HP0–HP8 boundary. Current retrieval improved 7/9→9/9 after historical contraction, research-only ForeignEgress/EffectPath machinery was deleted, and the same 96 mixed-failure invocations passed after deletion.

answered

Which Agent-level adversarial distinctions survive after mature security capability is imported?

Answered for the CA0–CA7 capability programme. CA6 showed a fixed script fail three of four held-out worlds while both a generic observation-driven deterministic adaptive policy and the canonical Harness/model Actor succeeded all four with the same oracle-regret vector; P1 reproduced the core adaptive-selection result under real provider/authority friction. The model did not establish a stronger residual than the thin adaptive policy. CA7 therefore closed by contraction: Campaign, Organization, persistent OpponentModel, coevolution, provider gateway, and cross-fidelity strategic law are not admitted by current evidence.

Operating discipline

Prefer observed use over speculative expansion.

New abstractions begin with a concrete failure or recurring need. They remain only when another real trajectory proves that a simpler owner or mature system cannot carry the responsibility.

Explore the research frontier