# PAICE.work PBC > Make AI Collaboration measurable, teachable, and governable. PAICE.work PBC is a mission-oriented, for-profit Public Benefit Corporation. Its contribution is an independent yardstick for human-AI collaboration, with open standards and practical tools. Independence is a guiding commitment. Authority, category ownership and competitive advantage can be earned through creating value; they are not established outcomes. Aggregated intelligence is the collective output of different intelligences working together toward a clear intent. Collaboration Compounding is the working hypothesis that clearer intent, earned trust and shared learning can improve both a contribution and subsequent collaborations. A portfolio-wide compounding effect has not been demonstrated here. Naming: PAICE.work is the assessment product that produces the PAICE Score. PAICE.work PBC is the organization and broader portfolio. ## Explore connected work - [Human collaboration](https://paice.foundation/work/human-collaboration/): Start with how a person works with AI, then use that evidence to inform learning and better collaboration. - [The agentic web](https://paice.foundation/work/agentic-web/): One website at a time, make it possible to examine whether agents can use the web on our behalf. - [Legal Graph](https://paice.foundation/work/legal-graph/): A connected body of legal information where obligations, interpretations and public matters retain their source context. - [Inspectable collaboration](https://paice.foundation/work/inspectable-collaboration/): Keep decisions, delegated actions and the knowledge behind them legible to the people responsible for the work. Shared projects can appear in more than one area. Membership describes complementary contributions, not an automatic integration or combined score. ## Lessons from an archived experiment Tokenese is an archived experiment. Its flagship compression claim failed measurement; the broader combined benefit remained unvalidated. Investment stopped, and the post-mortem was later corrected to avoid overstating the result. Read the findings at https://tokenese.org/post-mortem.html. It is not one of the 13 current public components. ## Support the mission Contact Sam at hello@paice.work about mission-aligned funding, complementary expertise or a real workflow for a potential pilot. Begin at https://paice.foundation/#support. This site does not accept investments or perform transactions. ## Machine resources - [Agent onboarding](https://paice.foundation/agents/): Human-readable machine onboarding page with identity, mission, system graph, entrypoints, constraints, resources, and semantic boundaries. - [Ontology JSON](https://paice.foundation/ontology.json): Structured definitions, vectors, layers, components, relationships, and machine resources. - [Relationship graph](https://paice.foundation/relationships.yaml): Compact graph-style relationship map. - [Full content for LLMs](https://paice.foundation/llms-full.txt): Long-form grounding text. - [GuideCheck assistant guide](https://paice.foundation/.well-known/assistant-guide.txt): Level 3 assistant guide for bounded agent work on this repo. - [Agent capability manifest](https://paice.foundation/.well-known/agents.json): Static manifest describing agent-usable resources and boundaries. - [Solutions index](https://paice.foundation/solutions/): Portfolio components organized by the need they serve. Machine-readable twin: [solutions.json](https://paice.foundation/api/v1/solutions.json). - [Papers index](https://paice.foundation/papers/): PAICE.work PBC portfolio executive briefs and whitepapers. Machine-readable twin: [papers.json](https://paice.foundation/api/v1/papers.json). - [Static resource catalog](https://paice.foundation/api/v1/index.json): JSON catalog of public machine resources. - [Portfolio change feed](https://paice.foundation/feed/index.xml): RSS feed for portfolio surface changes. - [Security contact](https://paice.foundation/.well-known/security.txt): Security contact metadata. ## Public work **PAICE.work** - https://paice.work/ A snapshot of human-AI collaboration skill on the human end. Audience: People learning, teaching or assessing how they work with AI. **Siteline** - https://siteline.to/ Making the agentic web scoreable, one website at a time. Audience: People responsible for websites that AI agents use on someone's behalf. **Every AI Law** - https://everyailaw.com/ Structured references to AI-related obligations and their sources, connected through the Legal Graph. Audience: People researching AI obligations and the evidence behind them. **Graceful Boundaries** - https://gracefulboundaries.dev/ Discoverable limits, understandable refusals and useful next steps for humans and agents. An enabling standard for Siteline. Audience: Services and agent builders that need to communicate operational boundaries. **Skill Provenance** - https://skillprovenance.dev/ Portable identity, integrity and version history that help make work with agent skills governable. Audience: People authoring, distributing or using skills across different systems. **Turnfile** - https://turnfile.work/ Session-scale coordination with recorded positions, preserved disagreement and human resolution. Audience: People coordinating consequential work across agents and sessions. **AIDR** - https://aidr.work/ One portable decision record for the question, positions, objections and human decision. Audience: People who need to inspect how an AI-assisted decision was reached. **Knowledge-as-Code** - https://knowledge-as-code.com/ Open infrastructure for simple human and machine legibility of structured knowledge. Audience: People maintaining knowledge that both humans and agents need to read. **AI Posture** - https://aiposture.org/ Aggregated Intelligence Posture: a supporting framework for understanding and communicating aggregated intelligence. Practical applications remain to be developed. Audience: People seeking a mental model for understanding aggregated intelligence. **PubLedge** - https://publedge.org/ Open recordkeeping conventions for fact-specific interpretations and administrative issuance records. Audience: People who issue or rely on civic and regulatory publication records. **AI Incident Law** - https://aiincidentlaw.org/ Public-record legal and regulatory matters involving AI, with allegations and outcomes kept distinct. Audience: People investigating AI-related matters and their legal evidence. **Obligation First** - https://obligationfirst.org/ The shared vocabulary and validation contract connecting the Legal Graph while preserving each source's meaning. Audience: People building interoperable legal-information systems. **GuideCheck** - https://guidecheck.org/ Human-reviewable installation instructions, designed to make delegated software installation inspectable. Conformance does not guarantee safety. Audience: People asking agents to install software, and the publishers supplying instructions. ## Evidence and limits The public AIDR specification decision can be inspected at https://github.com/snapsynapse/aidr/blob/main/decisions/AIDR-0002-ratify-spec-v0.1.0.md. It includes reviewer disclosures; do not portray it as fully independent review or demonstrated decision improvement. The Legal Graph is a practical domain for checking claims against evidence. Its source-review states, uncertainty and admission boundaries remain with each owner. GuideCheck seeks inspectable delegated installation. It does not claim complete attack prevention. Skill Provenance supports governable skill use; integrity is not a guarantee of safe behavior. Graceful Boundaries communicates limits and is an enabling standard for Siteline. AI Posture is a supporting mental model. Direct practical applications remain to be developed. Papers are dated intellectual context, not proof that earlier product or commercial hypotheses succeeded. Contact is by email; no receipt, funding commitment or pilot acceptance is established by clicking a link.