{"schemaVersion":"2.0","title":"@eooo.io","url":"https://eooo.io","contentStatus":"design-preview","projects":[{"slug":"factor-e","name":"Factor-E","number":"01","status":"Active","accent":"cyan","category":"Umbrella initiative","updated":"September 4, 2026","question":"What changes when the product is no longer the code, but the factory that produces outcomes?","summary":"The umbrella initiative connecting agent systems, control, context, observability, and the tools required to build the next set of tools.","thesis":"AI-native software is not conventional software with an assistant bolted onto the side. It is a different operating model: intent becomes executable, agents become participants, and the architecture must govern how judgment, context, authority, and recovery move through the system.","signal":"The important architecture is moving above the application layer and into the machinery that coordinates work.","focus":["Define the canonical mission record as the sole authority for orchestration state.","Make agent graphs visible without binding the system to one workflow engine.","Turn experiments into reusable doctrine, tools, and operating patterns."],"connectsTo":["SuperSkippy","SkippyScope","Merry Band","Kontrolr","Tessera","Tools"],"timeline":[{"id":"demo-update-01","date":"2026-09-04","displayDate":"SEP 04, 2026","phase":"Architecture","title":"Visualization becomes a projection, not an authority","summary":"The orchestration graph can be rendered through LangGraph-compatible views while the canonical mission record remains the system of record."},{"id":"demo-update-02","date":"2026-08-11","displayDate":"AUG 11, 2026","phase":"Doctrine","title":"From agent loops to engineered graphs","summary":"Nodes, routers, human gateways, conditional edges, ownership, and failure paths become explicit design material."},{"id":"demo-update-03","date":"2026-08-08","displayDate":"AUG 08, 2026","phase":"Formation","title":"The constellation gets a name","summary":"The related experiments are organized as one initiative: a factory for building the factories that produce software and operational outcomes."}],"writing":[{"kind":"Field note","status":"In development","title":"The factory becomes the product","slug":"the-factory-becomes-the-product","markdown":"## A small example, in several languages\n\nThese snippets illustrate the publishing format. They are not excerpts from the project's private implementation.\n\n### Python\n\n```python\ndef ready_for_review(mission: dict) -> bool:\n    \"\"\"Keep the decision small enough to inspect.\"\"\"\n    return bool(mission[\"evidence\"]) and mission[\"status\"] == \"complete\"\n```\n\n### TypeScript\n\n```typescript\ntype Mission = { status: string; evidence: string[] };\n\nconst readyForReview = (mission: Mission): boolean =>\n  mission.status === \"complete\" && mission.evidence.length > 0;\n```\n\n### PHP\n\n```php\n<?php\nfunction readyForReview(array $mission): bool\n{\n    return $mission['status'] === 'complete'\n        && count($mission['evidence']) > 0;\n}\n```\n\n### SQL\n\n```sql\nSELECT id, status\nFROM missions\nWHERE status = 'complete'\nORDER BY updated_at DESC;\n```\n\n### A decision worth drawing\n\nThe review checks the available evidence. Missing evidence returns the work for revision; sufficient evidence lets it proceed.\n\n```mermaid\nflowchart TD\n  accTitle: An illustrative evidence review\n  accDescr: Completed work is reviewed for evidence. Missing evidence returns to revision and then to review. Sufficient evidence proceeds to acceptance.\n  A[Completed work] --> B{Evidence sufficient?}\n  B -->|Yes| C[Accept result]\n  B -->|No| D[Revise work]\n  D --> B\n```\n","summary":"Why orchestration, policy, context, and verification are becoming more strategically important than any single generated artifact."},{"kind":"Architecture note","status":"Outline","title":"The mission record and the projected graph","slug":"the-mission-record-and-the-projected-graph","summary":"A separation-of-authority pattern for combining durable orchestration state with optional visualization and execution frameworks."}],"diagram":{"title":"From intent to governed outcomes","summary":"A deliberately simplified view of the common architecture emerging across the initiative.","nodes":[{"label":"Human intent","detail":"Purpose, constraints, and authority"},{"label":"Mission record","detail":"Canonical state and accountable decisions"},{"label":"Agent graph","detail":"Specialists, routes, gates, and recovery"},{"label":"Outcome record","detail":"Artifacts, evidence, and transferable lessons"}]},"nextQuestion":"What is the smallest complete Factor-E loop that can accept intent, coordinate work, expose its reasoning, recover from failure, and leave behind a trustworthy record?"},{"slug":"super-skippy","name":"SuperSkippy","number":"02","status":"Evolving","accent":"petrol","category":"Coordination layer","updated":"September 4, 2026","question":"Can an AI partner become a durable command layer without taking authorship away from the human?","summary":"A continuing experiment in memory, judgment, delegation, standards, and the long-running relationship between a builder and an AI system.","thesis":"A useful coordinating intelligence needs more than conversational fluency. It needs bounded authority, durable state, explicit missions, inspectable decisions, and a reliable way to return authorship to the human at the moments that matter.","signal":"Continuity is not a larger prompt. It is governed context, explicit authority, and a record that survives the conversation.","focus":["Canonical mission records","Read-only graph projections","Human authority and escalation boundaries"],"connectsTo":["Factor-E","Merry Band","SkippyScope","Kontrolr"],"timeline":[{"id":"demo-update-04","date":"2026-09-04","displayDate":"SEP 04, 2026","phase":"Planning","title":"Mission authority is made explicit","summary":"The coordination plan preserves one canonical record while treating external graph tooling as optional and read-only."},{"id":"demo-update-05","date":"2026-08-11","displayDate":"AUG 11, 2026","phase":"Doctrine","title":"Auto mode enters the operating model","summary":"Routine work can proceed autonomously while destructive or irreversible actions remain governed."}],"writing":[{"kind":"Design note","status":"Planned","title":"Continuity without counterfeit memory","slug":"continuity-without-counterfeit-memory","summary":"What durable partnership requires beyond context-window accumulation."}],"diagram":{"title":"Authority stays legible","summary":"SuperSkippy coordinates the work while the human retains authorship and irreversible authority.","nodes":[{"label":"Human author","detail":"Intent and consequential approval"},{"label":"SuperSkippy","detail":"Planning, context, and coordination"},{"label":"Specialists","detail":"Bounded execution and evidence"},{"label":"Mission record","detail":"Durable state returned to the author"}]},"nextQuestion":"How should authority degrade gracefully when context, tools, or confidence become unreliable?"},{"slug":"skippy-scope","name":"SkippyScope","number":"03","status":"Prototype","accent":"info","category":"Observability surface","updated":"September 4, 2026","question":"How do we make the reasoning and disagreement inside an agent system inspectable?","summary":"An observability surface for agent activity, decisions, handoffs, and the internal message board where specialized agents show their work.","thesis":"Agent observability should reconstruct not only what ran, but why a decision changed, where disagreement occurred, what evidence resolved it, and which authority permitted the next action.","signal":"If the work cannot be reconstructed, it cannot be trusted, improved, or safely repeated.","focus":["Mission and graph views","Agent channels and threaded deliberation","Evidence-linked decisions"],"connectsTo":["Factor-E","SuperSkippy","Merry Band"],"timeline":[{"id":"demo-update-06","date":"2026-09-04","displayDate":"SEP 04, 2026","phase":"Visualization","title":"Two levels of orchestration visibility","summary":"Mission topology and live execution state are separated so each can remain legible."},{"id":"demo-update-07","date":"2026-08-22","displayDate":"AUG 22, 2026","phase":"Concept","title":"The agents get a room of their own","summary":"Channels and threads become a record of disagreement, evidence, and reconciliation."}],"writing":[{"kind":"Product note","status":"In development","title":"Observability is more than telemetry","slug":"observability-is-more-than-telemetry","summary":"Why an agent system needs a legible social and decision layer, not only traces."}],"diagram":{"title":"From activity to reconstruction","summary":"Operational signals become useful only after they are attached to evidence, decisions, and authority.","nodes":[{"label":"Agent events","detail":"Actions, messages, and handoffs"},{"label":"Evidence links","detail":"Sources, artifacts, and tool results"},{"label":"Decision thread","detail":"Disagreement and reconciliation"},{"label":"Readable record","detail":"What happened, why, and under whose authority"}]},"nextQuestion":"Which internal events deserve permanence, and which should disappear once their operational value expires?"},{"slug":"merry-band","name":"Merry Band","number":"04","status":"Active","accent":"ember","category":"Agent organization","updated":"September 4, 2026","question":"Can a portable crew of specialized agents work as a disciplined organization rather than a prompt-shaped blob?","summary":"A reusable collection of roles, skills, handoff contracts, review gates, and occasionally argumentative personalities.","thesis":"Specialized agents become an organization only when roles, ownership, evidence, review, escalation, and recovery are designed as carefully as prompts and tools.","signal":"Specialization matters only when ownership, evidence, recovery, and the next handoff are explicit.","focus":["Portable agent archetypes","Bounded fan-out through Barney workers","Contracts for handoff and review"],"connectsTo":["Factor-E","SuperSkippy","SkippyScope"],"timeline":[{"id":"demo-update-08","date":"2026-09-04","displayDate":"SEP 04, 2026","phase":"Organization","title":"The Barneys join the crew","summary":"Disposable worker agents provide bounded parallel capacity beneath accountable specialist agents."},{"id":"demo-update-09","date":"2026-08-11","displayDate":"AUG 11, 2026","phase":"Architecture","title":"The crew becomes a graph","summary":"Handoffs, gates, ownership, and failure routes become first-class structure."}],"writing":[{"kind":"Operating note","status":"Draft","title":"A crew is not a swarm","slug":"a-crew-is-not-a-swarm","summary":"The organizational mechanics that distinguish accountable delegation from undirected parallelism."}],"diagram":{"title":"Accountable fan-out","summary":"Disposable parallel workers expand capacity beneath a specialist who still owns synthesis and quality.","nodes":[{"label":"Mission slice","detail":"One bounded question or deliverable"},{"label":"Specialist","detail":"Ownership, decomposition, and review"},{"label":"Barney fan-out","detail":"Parallel, disposable grunt work"},{"label":"Synthesis","detail":"One accountable result with evidence"}]},"nextQuestion":"How much organizational structure improves reliability before it becomes theatrical bureaucracy for machines?"},{"slug":"kontrolr","name":"Kontrolr","number":"05","status":"Design","accent":"signal","category":"AI control plane","updated":"August 28, 2026","question":"Where should policy, tools, permissions, memory, and model routing live?","summary":"A privacy-conscious control plane for governing how people and agents use models, context, tools, and organizational authority.","thesis":"AI capabilities should be governed from a coherent control layer where policy, identity, tools, context, models, audit, and organizational authority can be reasoned about together.","signal":"Most AI governance problems are authority problems wearing an impressive lanyard.","focus":["Policy and permission evaluation","Model and tool routing","Auditable context boundaries"],"connectsTo":["Factor-E","SuperSkippy","Tessera"],"timeline":[{"id":"demo-update-10","date":"2026-08-28","displayDate":"AUG 28, 2026","phase":"Evaluation","title":"The harness joins the trust boundary","summary":"Sandboxing, recovery, tool semantics, and state handling are treated as part of the governed system."},{"id":"demo-update-11","date":"2026-06-12","displayDate":"JUN 12, 2026","phase":"Architecture","title":"The control plane takes shape","summary":"Policy, routing, context, and audit are separated from individual AI applications."}],"writing":[{"kind":"Architecture note","status":"Planned","title":"Govern the capability, not the chat box","slug":"govern-the-capability-not-the-chat-box","summary":"A control-plane model for AI use across products and agent systems."}],"diagram":{"title":"One governed capability path","summary":"Every request crosses the same authority boundary before models, tools, or retained context become available.","nodes":[{"label":"Identity + intent","detail":"Who is asking and what they need"},{"label":"Policy decision","detail":"Permissions, purpose, and risk"},{"label":"Capability route","detail":"Model, tool, and context selection"},{"label":"Audit record","detail":"Decision, execution, and recovery evidence"}]},"nextQuestion":"Can one policy model remain understandable while governing humans, agents, models, tools, and retained context together?"},{"slug":"tessera","name":"Tessera","number":"06","status":"Concept","accent":"slate","category":"Private context vault","updated":"August 12, 2026","question":"How can personal context become useful without becoming somebody else's permanent asset?","summary":"A private context vault exploring local-first memory, selective disclosure, trust boundaries, and human control over retained knowledge.","thesis":"Personal context should behave like a user-owned capability: locally governed, selectively disclosed, revocable, inspectable, and useful across tools without silently becoming platform property.","signal":"Memory becomes leverage only when the person it describes still controls the boundary.","focus":["Local-first context","Selective disclosure","Portable memory contracts"],"connectsTo":["Factor-E","Kontrolr","SuperSkippy"],"timeline":[{"id":"demo-update-12","date":"2026-08-12","displayDate":"AUG 12, 2026","phase":"Data","title":"The archive becomes queryable context","summary":"Conversation history is reframed as governed personal data rather than a pile of exported transcripts."},{"id":"demo-update-13","date":"2026-04-18","displayDate":"APR 18, 2026","phase":"Concept","title":"A vault for usable memory","summary":"Local storage, trust boundaries, and selective disclosure define the first architecture."}],"writing":[{"kind":"Research note","status":"Outline","title":"Memory with an owner","slug":"memory-with-an-owner","summary":"A user-controlled alternative to platform-bound personalization."}],"diagram":{"title":"Context crosses by consent","summary":"Personal memory remains behind a user-controlled boundary and leaves only through explicit disclosure rules.","nodes":[{"label":"Private context","detail":"Locally governed personal knowledge"},{"label":"Trust boundary","detail":"Consent, scope, and revocation"},{"label":"Selective view","detail":"The minimum useful disclosure"},{"label":"External tool","detail":"Temporary, purpose-bound use"}]},"nextQuestion":"What consent model is simple enough to use daily and precise enough to govern consequential disclosure?"},{"slug":"orchestrati","name":"Orchestrati","number":"07","status":"Concept","accent":"oxblood","category":"Learning world","updated":"August 19, 2026","question":"Can people learn agent orchestration through puzzles, narrative, and systems that push back?","summary":"A mildly absurd secret order becoming a possible learning world for orchestration, architecture, and AI-assisted problem solving.","thesis":"The hardest orchestration lessons may be learned more deeply through a world that makes players investigate context, assign authority, debug unreliable agents, and live with the consequences of weak specifications.","signal":"The line between a serious training system and a rabbit-infested mythology is thinner than management would prefer.","focus":["Myst-like systems puzzles","Narrative instruction","Agent orchestration as play"],"connectsTo":["Factor-E","Merry Band","Tools"],"timeline":[{"id":"demo-update-14","date":"2026-08-19","displayDate":"AUG 19, 2026","phase":"World building","title":"The joke becomes a learning system","summary":"A secret order, symbolic lore, and puzzle mechanics converge into a possible training platform."},{"id":"demo-update-15","date":"2026-08-17","displayDate":"AUG 17, 2026","phase":"Origin","title":"The Order is founded","summary":"An absurd mythology begins accumulating rules, symbols, texts, and suspiciously useful architectural lessons."}],"writing":[{"kind":"World note","status":"Collecting","title":"Contextua Magnifica","slug":"contextua-magnifica","summary":"Fragments from the sacred and operational texts of the Order."}],"diagram":{"title":"The lesson hides inside the world","summary":"Narrative pressure turns orchestration mechanics into consequences the player must investigate and repair.","nodes":[{"label":"Systems puzzle","detail":"Incomplete context and unreliable actors"},{"label":"Player plan","detail":"Authority, sequence, and delegation"},{"label":"World response","detail":"Consequences expose weak assumptions"},{"label":"Learned doctrine","detail":"A reusable orchestration pattern"}]},"nextQuestion":"Can the world teach disciplined orchestration without ever feeling like a disguised enterprise training course?"}],"essays":[{"status":"In development","title":"When software stops waiting for humans","summary":"We built interfaces for human eyes because humans were the actors. As agents become the operating surface, the assumptions beneath product design begin to move.","themes":["AI-native work","Interfaces","Systems"],"slug":"when-software-stops-waiting-for-humans","section":"essays","image":"/images/editorial/code-judgment.webp","imageWidth":1672,"imageHeight":941,"imageCaption":"Editorial illustration · Demo essay","imageAlt":"An abundance of code converges on one human reviewer at a red decision gate","markdown":"## The question\n\nWe built interfaces for human eyes because humans were the actors. As agents become the operating surface, the assumptions beneath product design begin to move.\n\n## Where this could lead\n\nAn essay is room to follow an argument beyond a single project: to examine the assumptions, consider the counterargument, and work out what survives.\n\nThis is a layout sample. The complete essay, supporting sources, and any archive migration will be added separately."},{"status":"Draft","title":"When code becomes abundant, judgment becomes the bottleneck","summary":"Cheap production does not remove engineering. It moves the scarce work into specification, context, verification, and the design of systems that can fail without taking the building with them.","themes":["Software engineering","Agents","Architecture"],"slug":"when-code-becomes-abundant","section":"essays","image":"/images/editorial/code-judgment.webp","imageWidth":1672,"imageHeight":941,"imageCaption":"Editorial illustration · Demo essay","imageAlt":"An abundance of code converges on one human reviewer at a red decision gate","markdown":"## The question\n\nCheap production does not remove engineering. It moves the scarce work into specification, context, verification, and the design of systems that can fail without taking the building with them.\n\n## Where this could lead\n\nAn essay is room to follow an argument beyond a single project: to examine the assumptions, consider the counterargument, and work out what survives.\n\nThis is a layout sample. The complete essay, supporting sources, and any archive migration will be added separately."},{"status":"Archive migration","title":"The tension between epistemic curation and ideological capture","summary":"An earlier essay returning from Medium to a permanent home, where the argument can be preserved, revised transparently, and connected to the work that followed.","themes":["Knowledge","Institutions","Culture"],"slug":"epistemic-curation-and-ideological-capture","section":"essays","image":"/images/editorial/code-judgment.webp","imageWidth":1672,"imageHeight":941,"imageCaption":"Editorial illustration · Demo essay","imageAlt":"An abundance of code converges on one human reviewer at a red decision gate","markdown":"## The question\n\nAn earlier essay returning from Medium to a permanent home, where the argument can be preserved, revised transparently, and connected to the work that followed.\n\n## Where this could lead\n\nAn essay is room to follow an argument beyond a single project: to examine the assumptions, consider the counterargument, and work out what survives.\n\nThis is a layout sample. The complete essay, supporting sources, and any archive migration will be added separately."}],"humor":[{"slug":"the-committee-for-spontaneous-innovation","section":"monk-e-wrenched","title":"The Committee for Spontaneous Innovation","summary":"All unplanned breakthroughs must now be submitted six weeks in advance.","status":"Sample satire","themes":["Bureaucracy","Inventive nonsense"],"image":"/images/editorial/monk-e-wrenched.webp","imageWidth":1672,"imageHeight":941,"imageCaption":"Monk-e-Wrenched · Editorial illustration","imageAlt":"A robed monkey holds an oversized wrench beside a machine with one red gear","markdown":"## An entirely reasonable announcement\n\nThe committee is delighted to announce that spontaneous innovation is now a strategic priority. A working group has been established to determine when it may occur.\n\nAll unexpected discoveries must be entered into the Unexpected Discovery Register before they are discovered. Discoveries made without prior authorization will be classified as administrative misunderstandings.\n\n> Please think outside the box. The box must remain inside the approved thinking area.\n\n## Meanwhile, in the workshop\n\nSomeone has built the thing. This has caused considerable concern. There was no kickoff meeting, no steering committee, and the prototype appears to work.\n\nA retrospective has been scheduled to ensure this does not happen again."}],"notes":[],"absurdities":[{"slug":"dave","title":"Dave","category":"Space / Existential inconvenience","status":"Playable preview","summary":"An astronaut. An airlock. A computer with some concerns.","description":"A browser experiment in space, atmosphere, and the awkward business of asking a computer to let you back inside. Explore Discovery One from the pod, then return to the door for a conversation.","url":"https://dave.eoooio.chatgpt.site"},{"slug":"weird","title":"Weird","category":"Questionable Science","status":"Playable preview","summary":"A deeply unscientific 1980s bedroom experiment.","description":"Questionable Science is the place for an experiment that wandered away from respectable engineering. A small browser diversion with an unabashedly strange premise.","url":"https://weird.eoooio.chatgpt.site"},{"slug":"just-one","title":"Just One","category":"Retro adventure / In development","status":"Concept","summary":"A late-1980s adventure idea, beginning with one room and a growing collection of frustrations.","description":"A retro adventure concept in the spirit of the point-and-click era. The first scene explores a room and its stairs; this page will collect the playable version and the story of how it develops.","url":null}]}