BITwiki Platform General Glossary & System FAQ
A practical glossary for understanding BITwiki, BIThub, BITcore, Constructs, agents, workflows, artifacts, and the knowledge pipeline around the platform.
Use this page when a platform term is unclear.
References:
What is the BITwiki ecosystem?
The BITwiki ecosystem is the collection of knowledge, forum, AI, documentation, and workflow surfaces connected to BITwiki.
The main public surfaces are:
BITwiki is the knowledge base.
BIThub is the forum and workspace layer.
BITcore is the cognitive surface expressed through AI agents, personas, prompts, workflows, and related implementations.
References:
What is BITwiki?
BITwiki is the structured knowledge base of the ecosystem.
It is used for articles, topics, definitions, references, documentation, research pages, and organized knowledge.
BITwiki functions as a digital codex: a place where knowledge can be organized into modular, referenceable, and queryable forms.
BITwiki is still developing. Pages may include human-assisted AI content, incomplete drafts, or material that still needs review.
Reference:
What is BIThub?
BIThub is the forum and workspace layer of the ecosystem.
It is where users ask questions, discuss ideas, test tools, coordinate work, report issues, generate artifacts, and interact with AI-assisted workflows.
BIThub is not only a social forum. It is a working space for structured discussion, testing, coordination, and reusable knowledge production.
Reference:
What is BITcore?
BITcore is the summarized cognitive surface of the ecosystem.
It is not one bot, one model, one prompt, or one repository.
BITcore is the sum of the internal cognitive surfaces formed by agents, personas, workflows, prompts, memory patterns, context, and system identity.
Several BITcore variants may exist. Some are chat surfaces. Some are personas. Some are workflow-oriented. Some try to distill the strongest reasoning patterns of the system into a usable agent.
References:
Is BITcore one bot, one model, or one system?
BITcore is not one bot or one model.
A model is the underlying AI engine.
A bot is one interface.
A persona is one configured identity.
A workflow is one process.
BITcore is the sum of these internal cognitive surfaces formed by agents, personas, workflows, prompts, memory patterns, context, and system identity.
The ecosystem is interested in making LLMs and fine-tuning models in the future, but that does not mean current BITcore surfaces are custom base models.
Why are there multiple BITcore agents or personas?
Multiple BITcore variants allow different reasoning styles, roles, constraints, and workflows to be tested.
One BITcore surface may focus on explanation.
Another may focus on system reasoning.
Another may focus on dialogue, research, planning, or workflow support.
The shared purpose is to distill useful reasoning patterns into better cognitive surfaces.
What does holobiont mean in the BITcore context?
In biology, a holobiont is a host plus its associated symbiotic organisms.
In the BITcore context, holobiont is used as a metaphor for a composite human-AI system made of multiple interacting parts.
Those parts may include humans, AI agents, prompts, workflows, documents, tools, memory, and structured outputs.
BITcore is not an isolated chatbot. It is closer to the visible cognitive surface of a larger human-AI system that we call an Advanced Holobiont Intelligence.
References:
What is the difference between BITwiki, BIThub, and BITcore?
BITwiki is the structured knowledge base.
BIThub is the forum, workspace, and coordination layer.
BITcore is the summarized cognitive surface expressed through agents, personas, prompts, workflows, and system-facing AI implementations.
A simple map:
- BITwiki organizes knowledge
- BIThub coordinates discussion and testing
- BITcore expresses the reasoning surface
References:
What is the difference between a category, topic, post, tag, artifact, and wiki page?
A category is a major forum area used to group related topics.
A topic is a discussion thread.
A post is an individual message inside a topic.
A tag is a keyword used to label and find related topics.
An artifact is a reusable output created from discussion, research, AI work, testing, or documentation.
A wiki page is structured knowledge crystallized into a more stable encyclopedic or documentation form.
References:
What is an AI model?
An AI model is the underlying system that generates outputs.
A model can power many different agents, Constructs, personas, and workflows.
BITcore may use models, but BITcore is not identical to the model underneath it.
Reference:
What is fine-tuning?
Fine-tuning is the process of further training an existing AI model on selected examples or data so it behaves better for a specific purpose.
The BITwiki ecosystem is interested in model-building and fine-tuning as a future direction.
Current BITcore surfaces may be built through prompts, context, workflows, agents, and configuration rather than custom-trained base models.
Reference:
What is a prompt?
A prompt is an instruction or context given to an AI model.
Prompts can shape how an AI answers, what role it plays, what format it uses, and what constraints it follows.
Prompts are part of how Constructs, agents, personas, and BITcore variants can be configured.
References:
What is context?
Context is the information available to an AI or workflow when producing an output.
Context may include a user question, previous conversation, documents, system instructions, examples, tags, categories, retrieved knowledge, or source material.
Good context improves output quality.
Bad context can cause confusion, drift, or hallucination.
References:
- Google Machine Learning Crash Course — Introduction to Large Language Models
- Anthropic Engineering — Effective context engineering for AI agents
- Google Cloud — What is prompt engineering?
What is a persona?
A persona is a configured identity, voice, or role for an AI surface.
A persona may affect how an assistant explains itself, what role it plays, what constraints it follows, or what style of reasoning it emphasizes.
A persona is not the same thing as a model.
The model is the underlying engine.
The persona is part of the configured surface.
What is an agent?
An agent is an AI-driven or software-driven unit that performs a task.
An agent should have a defined role, boundary, and expected output.
A good agent does not try to do everything. It performs a specific kind of work clearly.
References:
What is a Construct?
A Construct is a structured AI configuration that defines the functional boundaries, behavior, logic, context, and purpose of an AI-facing surface.
A Construct is deeper than a persona.
A persona shapes expression.
A Construct shapes the cognition frame in which reasoning happens.
Constructs can support agents, personas, workflows, tools, and task-specific reasoning surfaces.
References:
What is an Agentic Construct?
An Agentic Construct is a Construct with execution capacity.
It does not only shape how an AI surface reasons. It can also act through tools, workflows, retrieval, commands, or other execution paths.
Agentic Constructs can support more complex workflows because they combine a structured cognition frame with task execution.
References:
What is the difference between an agent, Construct, persona, and Agentic Construct?
A persona shapes expression. It affects role, tone, style, behavior, or interaction pattern.
An agent performs action. It can follow instructions, use tools, execute steps, or complete scoped work.
A Construct defines the cognition frame. It sets the logic, boundaries, context, purpose, and operating assumptions that shape how an AI surface reasons.
An Agentic Construct is a cognition frame with execution. It combines Construct-level reasoning structure with action through tools, workflows, retrieval, commands, or other execution paths.
Compact map:
- Persona = expression
- Agent = action
- Construct = cognition frame
- Agentic Construct = cognition frame with action
References:
What is a CORE?
A CORE is a structured workflow or reasoning process composed of multiple Agentic Constructs.
A CORE usually has a defined purpose and produces a more organized output than a normal chat reply.
The exact behavior depends on the specific CORE.
A CORE should be judged by whether it produces useful, repeatable, reviewable output.
What is a Node?
A Node is a narrower agent, tool, or task unit.
A Node may support a specific function such as search, formatting, summarization, classification, analysis, testing, or generation.
The exact meaning depends on the implementation.
What is a Workspace?
A Workspace is an area where users can work with tools, workflows, prompts, agents, or outputs.
A Workspace may be simple or advanced depending on the use case.
The term should be understood practically: it is a place where structured work happens.
Are Constructs, COREs, Nodes, and Workspaces final terms?
Not fully. They are working platform terms.
Their meanings may become more precise as the system stabilizes.
Current practical usage:
- Construct means cognition frame
- CORE means structured workflow
- Node means specialized task unit
- Workspace means working area
What is a workflow?
A workflow is a repeatable process for doing work.
It may involve humans, AI tools, prompts, review steps, sources, and outputs.
A good workflow should be understandable, repeatable, and reviewable.
What is a MAS-Factory?
A MAS-Factory is a multi-agent system organized like a production line for intelligence work.
In a multi-agent system, multiple agents work together inside a shared task environment. They may divide work, coordinate outputs, and pass results between roles.
In a MAS-Factory pattern, work can move from one agent or Agentic Construct to another.
One unit may gather information. Another may summarize. Another may check evidence. Another may format the result. Another may generate an artifact. Another may prepare the output for review.
The goal is to divide complex knowledge work into scoped steps that can be reviewed, improved, and reused.
A MAS-Factory can become an intelligence factory when agentic work produces reusable artifacts, structured knowledge, or accepted BITwiki material.
References:
- Google Cloud — What is a multi-agent system?
- NVIDIA — Multi-Agent Systems
- GitHub Resources — What are multi-agent systems?
What is an intelligence factory?
An intelligence factory is a system for turning raw discussion, questions, research, AI outputs, feedback, testing, and agentic work into reusable knowledge and artifacts.
In BIThub, this can happen through human discussion, AI-assisted workflows, Constructs, Agentic Constructs, and MAS-Factory patterns.
The output is not just conversation. The output should become something useful: a better answer, a reusable artifact, a tested workflow, a reviewed summary, or crystallized BITwiki knowledge.
A simple production line may look like this:
- A user submits a research request.
- One agent gathers context.
- Another agent summarizes.
- Another checks sources.
- Another structures the output.
- Another turns it into an artifact.
- A human reviews it.
- Accepted material can move toward BITwiki.
The narrow claim is that BIThub and BITwiki are designed to make knowledge work more reusable, structured, and reviewable.
References:
What is an artifact?
An artifact is a reusable output.
Examples include summaries, guides, templates, research notes, workflows, prompts, diagrams, reports, datasets, bug reports, and structured answers.
An artifact should be useful beyond the original conversation.
How does something move from BIThub into BITwiki?
Moving from BIThub into BITwiki means crystallizing discussion, research, or artifacts into structured encyclopedic knowledge.
BIThub is where knowledge can be explored, debated, tested, corrected, and shaped.
BITwiki is where accepted knowledge can be preserved in a more modular, succinct, queryable, and reference-oriented form.
A good BITwiki candidate should be:
- Structured
- Modular
- Succinct
- Source-backed
- Useful beyond one thread
- Organized as reference knowledge
- Clear enough to be queried, linked, categorized, or reused
- Stable enough to represent the current accepted version of a topic
This can include MediaWiki-style encyclopedia pages, definitions, structured references, guides, templates, or other knowledge forms.
A discussion should move into BITwiki only after the research is conducted properly, reviewed, and accepted as worth preserving.
The process through which knowledge artifacts are distilled into BITwiki is still being built and audited.
Reference:
What is reusable intelligence?
Reusable intelligence means useful work survives beyond one conversation.
A strong answer, workflow, test, summary, or discussion can become something other people can search, reuse, improve, cite, or turn into documentation.
This is one of the main reasons BIThub exists.
References:
What is structured knowledge?
Structured knowledge is information organized so it can be found, compared, connected, reused, and improved.
It may include pages, categories, tags, properties, sources, examples, workflows, artifacts, and relationships.
Structured knowledge is different from a loose pile of text because it has shape and context.
Reference:
What is a semantic knowledge graph?
A semantic knowledge graph is a structured map of concepts and relationships.
Instead of treating every page or post as isolated text, a semantic graph helps connect entities, topics, categories, references, and properties.
This is useful when knowledge needs to be searched, related, compared, or reused.
Reference:
What is memory in this platform?
Memory means preserved context that can help future work.
This may include wiki pages, posts, summaries, artifacts, tags, categories, semantic properties, saved outputs, or reviewed decisions.
Memory is useful when it helps users avoid repeating the same work.
What is the difference between memory and storage?
Storage saves data.
Memory preserves useful context.
A folder full of files is storage.
A structured map of what those files mean, why they matter, and how they connect is closer to memory.
What does AI-assisted mean?
AI-assisted means AI tools may help create, summarize, organize, or refine content.
It does not mean the content is automatically correct.
AI-assisted work should still be reviewed, corrected, sourced, and improved by humans.
BITwiki pages may include human-assisted AI content, so important claims should be checked before reuse.
Reference:
What is a hallucination?
A hallucination is when an AI produces information that sounds plausible but is false, unsupported, distorted, or invented.
Hallucinations can happen in definitions, citations, summaries, technical explanations, timelines, or claims about the platform itself.
If there is no evidence for a claim, it should not be presented as fact.
Reference:
What is cognitive drift?
Cognitive drift is when reasoning, interpretation, or outputs gradually move away from the original goal, evidence, context, or intended meaning.
In AI-assisted work, cognitive drift can happen when a model keeps elaborating, summarizing, rewriting, or chaining outputs without enough grounding.
Related technical ideas include model drift, data drift, and concept drift. NIST notes that AI systems may require maintenance because of data, model, or concept drift.
In BIThub and BITwiki work, drift is reduced by using sources, scope, review, tests, and explicit acceptance criteria.
Reference:
Why does evidence matter?
Evidence matters because AI systems and humans can both produce confident mistakes.
A useful knowledge platform needs a way to separate claims, guesses, interpretations, outputs, and verified information.
Strong contributions should make it clear where information comes from and how confident users should be.
Evidence can include public pages, documentation, source links, screenshots, repository files, direct tests, or reproducible examples.
For AI systems, evidence also supports transparency, explainability, reliability, and accountability.
References:
- NIST AI Risk Management Framework
- NIST AI Resource Center — AI Risks and Trustworthiness
- IBM — What Are AI Hallucinations?
What is recursion in the BITwiki ecosystem?
Recursion means that an output can become input for another round of review, refinement, organization, or reuse.
A BIThub thread can produce a summary.
A summary can become an artifact.
An artifact can become a BITwiki draft.
A BITwiki page can reveal missing questions.
Those questions can return to BIThub for more research.
The point is controlled improvement through feedback loops.
Reference:
What is recursive intelligence?
Recursive intelligence is improvement through structured feedback loops.
In this ecosystem, users and AI systems can produce outputs, review them, correct them, reuse them, and turn them into better knowledge structures.
The practical test is whether the system gets clearer, more accurate, and more useful over time.
Reference:
What is cognition-as-infrastructure?
Cognition-as-infrastructure means treating reasoning, memory, context, workflow, and knowledge organization as part of the platform.
Most platforms store content.
This ecosystem is trying to structure the thinking around the content so it can be reused, reviewed, and improved.
The intelligence factory is one expression of cognition-as-infrastructure.
Instead of treating each answer as disposable, the system tries to preserve useful reasoning as artifacts, workflows, references, and structured knowledge.
References:
What is ontology?
Ontology means the structure of what things are.
In philosophy, ontology is commonly described as the study of what exists. In information systems, ontology refers to defining or describing things and relationships so they can be used computationally.
In this platform, ontology can refer to how concepts, categories, agents, workflows, artifacts, and knowledge pages are defined and related.
Reference:
What is epistemology?
Epistemology means how we know something.
In philosophy, epistemology concerns knowledge, justification, and what distinguishes knowledge from mere opinion.
In this platform, epistemology matters because AI outputs, user claims, sources, summaries, and workflows need to be checked.
Reference:
What is axiology?
Axiology means the study of value.
In this platform, axiology refers to what the system should reward, preserve, or optimize for.
Useful values include truth, evidence, clarity, reuse, safety, quality, and contribution signal.
Reference:
Why does the platform use system language?
The platform uses system language because it is trying to describe real structures: agents, Constructs, workflows, knowledge graphs, artifacts, categories, memory, and review loops.
The goal is not to decorate simple ideas with complex words.
The goal is to expose the structure clearly enough that users can understand what part of the system they are using, what it does, and where the output should go.
Terms like Construct, CORE, Node, artifact, recursion, and holobiont should point to real patterns in the system.
If a term does not clarify the structure, it will be rewritten, grounded, or removed.
Is this science fiction or a real platform?
It is a real platform that uses some system metaphors to explain complex architecture.
The real platform includes public websites, forum categories, wiki pages, AI constructs, repositories, documentation, and user workflows.
The metaphors are not there to hide the product.
They are there to make invisible structures easier to see: memory, coordination, recursion, agents, artifacts, and knowledge crystallization.
The practical standard is simple: the software, workflows, and outputs still need to work.
References:
What does scoped mean?
Scoped means limited to a specific purpose or boundary.
A scoped tool should not try to do everything.
Examples:
- A summarizer summarizes
- A search tool searches
- A bug-report helper collects reproducible details
- A formatting tool formats without inventing facts
Scope reduces confusion and risk.
What is high-signal and low-signal participation?
High-signal participation makes the platform clearer, more accurate, or more useful.
Examples include clear questions, good bug reports, useful summaries, tested workflows, accurate corrections, helpful examples, strong documentation, and verified sources.
Low-signal participation creates more noise than value.
Examples include vague posts, duplicate topics, unsupported claims, AI output pasted without review, drama, spam, complaints without examples, and unclear walls of text.
The goal is not to post more. The goal is to add useful signal.
What is the simplest map of the system?
Use this map:
- BITwiki = knowledge base
- BIThub = coordination and workflow space
- BITcore = cognitive surface
- AI model = underlying engine
- Prompt = instruction/context input
- Context = available information
- Persona = expression layer
- Agent = action unit
- Construct = cognition frame
- Agentic Construct = cognition frame with execution
- MAS-Factory = multi-agent production line
- CORE = structured workflow
- Node = specialized task unit
- Workspace = working area
- Category = major forum area
- Topic = discussion thread
- Post = message inside a topic
- Tag = keyword label
- Artifact = reusable output
- Wiki page = crystallized knowledge
- Ontology = what things are
- Epistemology = how we know
- Axiology = what matters
The ecosystem turns discussion, AI work, and research into reusable knowledge.