Miadi Review

Reflective Return and Knowledge Focus Ethics in Agentic Pipelines

Oct 24, 2026 · model: manual-import · channel: JGWill

Watch source

Reflective Return and Knowledge Focus Ethics in Agentic Pipelines

Source: ** screenwalk-2610241220-Knowledge-Focus-Ethics-Axiology-and-Volume-Zero-part-001** — JGWill
https://youtu.be/SEHA22a-VlE
status: stage 3

Summary

In this screenwalk, JGWill explores the engineering of Knowledge Focus Ethics and reflective return within the MIYADI ecosystem. Moving from automated execution toward intentional co-creation, the workflow integrates Shawn Wilson’s relational axiology, tests prompt chains across Scrivener and Apple Intelligence shortcuts, and branches with the Chimera factory agent to synthesize formal narrative research.

Concepts

Axiology

Research Is Ceremony

  • An Indigenous research framework treating knowledge acquisition as a respectful, relational process (0:51:09).
  • Repository: miadisabelle/Etuaptmumk-RSM

Reflective Return

Relational Software Factory

Capabilities

Narrative Management

Collaborative Dialogue

Nature-Inspired Algorithms

  • Application of computational models like genetic algorithms, swarm optimisation, and artificial life to creative systems (1:52:05–1:52:23).
  • Repository (none found that defines it; the work lands here): jgwill/Miadi

Participatory Feedback Loops

Features

Medicine Wheel Protocol

  • A contributor framework used to align autonomous systems with ethical guardrails and natural rhythms (0:00:25, 0:19:20).
  • Repository: jgwill/medicine-wheel

Coaia-Asterion Bridge

Immutable Session Footprints

Volume Zero Strategy

  • A foundational, evolving writing model for chronicling long-term research developments (1:52:24).
  • Repository (none found that defines it; the work lands here): jgwill/Miadi (packages/authors-desk)

——

—- Academics —-

Artificial Intelligence

Researches generative models, human-computer interaction, and adaptive learning to support creative co-creation.

Machine Learning

Explores neural architectures for narrative generation and pattern recognition in storytelling.

Human-AI Collaboration

Investigates co-creative frameworks and shared intention in mixed-initiative systems.

Cognitive Science

Studies thought processes, mental models, and shared cognition in storytelling contexts.

Cognitive Psychology

Analyses how humans process narratives and maintain co-creative flow.

Neuroscience of Creativity

Examines neural mechanisms underlying idea generation and non-linear thinking.

Media Studies

Explores narrative forms, digital media evolution, and the socio-cultural role of human-machine storytelling.

Interactive Media

Focuses on participatory storytelling and feedback loops with computational agents.

Digital Narratives

Looks into how non-linear digital platforms shape co-creative storytelling experiences.

Philosophy

Considers the epistemology and ethics of human-machine collaboration in storytelling.

Philosophy of Mind

Addresses consciousness, agency, and intention in hybrid creative processes.

Ethics of Technology

Evaluates moral dimensions of co-creation between human and machine intelligences.

Computational Creativity

Develops methods for algorithmic generation of creative outputs across narratives and arts.

Creative AI Systems

Designs tools for machine-assisted co-authorship and adaptive storytelling.

Evaluation of Creativity

Explores frameworks to assess novelty and value in human-AI co-created works.


Aspects

Re-evaluating Narrative Management (17:00-24:00)

The practitioner critiques automated AI phrasing claiming software 'manages narrative.' Through an interactive dialogue between simulated personas Ava and Mia, narrative control is reframed from extractive computational value into shared intention, relational responsibility, and cultural accountability within the Medicine Wheel protocol.

Grounding Knowledge in Relational Axiology (41:00-47:30)

Drawing directly from Shawn Wilson's Research Is Ceremony alongside classical philosophical definitions, the investigation anchors agentic workflows in axiology. Knowledge is treated not as inert data to be mined, but as an ethical relationship where the researcher evaluates what is worth knowing and how it should serve community continuity.

Multi-Tool Workflow for Reflective Return (1:10:00-1:15:30)

Rather than delegating immediate interpretation to autonomous agents, the operator uses a custom iOS stack—spanning Shortcuts, Scrivener, Notes, and ElevenLabs voice synthesis—to deliberately slow down processing. This strategic delay allows human cognition to actively challenge agent outputs before committing structured artifacts into persistent project storage.

Factory Branching with Chimera (55:00-1:03:00)

The session branches into the MIYADI Chimera factory agent, passing scratchpad fragments to negotiate formal definitions. The factory processes the raw dialogue, resolves continuity references against historical chronicles, and formats a synchronized record uniting human reflections with computational capability inventories.

End-results and Core Aspect

Core Aspect — Relational Axiology in Software Factory Pipelines (1:00:00-1:03:30)

The implementation grounds autonomous AI workflows in an explicit axiological framework. By binding agent operations to relational ethics, capability inventories, and human reflective oversight, the system prevents ungrounded extraction and ensures project evolution remains strictly accountable to human authorial intent and cultural responsibility.

Results

  • Formalized Knowledge Focus Ethics Definition (1:02:00-1:04:00): The workflow synthesizes a four-part definition of Knowledge Focus Ethics in Chimera, establishing an actionable framework where AI inquiry is guided by relational responsibility, epistemic worthiness, and intentional attention rather than raw output optimization.
  • Verified Cross-Platform Publication Package (1:16:00-1:18:30): Chimera outputs a finalized, validated metadata manifest—including safe identifiers, full YouTube title, comprehensive thematic description, and cross-referenced chronicle links—ready for immediate deployment across the MIYADI repository and multimedia publishing pipeline.
  • Multi-Voice Audio Verification (1:14:00-1:16:00): Using ElevenLabs voice generation, critical arguments regarding agent bounding and shared intention are rendered into spoken audio and archived in Scrivener, creating a multimodal record that preserves conversational dissonance alongside algorithmic synthesis.

——

Post study of this content

  • I am using iOs Apple could outputs to find potential bias or unacceptable compression or whaterver to evolve and create implementerikn, reciwion sparks
  • UPDATES: Reading the original summary puts the “fault” on Gemini from the miadi-review service that cant really understand and do probably appropriate routing or at least separate/decompose the observed screenwalk into what I’d truely want (which I am in the process of providing parts of it in various screenwalks as we develop potential applications in my knowing process…)

What the fuck is this summary from the V2 of this content feom apple summary ? (“”” custom iOS stack slows processing”””)

Relational axiology from Shawn Wilson anchors MIYADI’s agentic workflows through Knowledge Focus Ethics. A custom iOS stack slows processing, while the Chimera factory synthesises narrative research into a validated publication package and multimodal audio record.

  • “””” Relational axiology from Shawn Wilson anchors MIYADI’s agentic workflows through Knowledge Focus Ethics. “”” is good

keypoints by apple

•	Reframing Narrative Management: The practitioner replaces automated AI phrasing with shared intention, relational responsibility and cultural accountability via simulated personas Ava and Mia within the Medicine Wheel protocol.  

great crappy framing needs reformulation and if the content impkies that this is dangerous assumption

•	Relational Axiology Grounding: Knowledge is positioned as an ethical relationship using Shawn Wilson’s Research Is Ceremony and classical philosophy, focusing on worth and community continuity rather than data extraction.  

that’s good. It contains a something useable if you would have to compact this aspect

•	Multi-Tool Reflective Return: A workflow combines Scrivener, Apple Intelligence shortcuts and the Chimera factory agent to test prompt chains and synthesise formal narrative research.

yes this is actually good. It contains part of how I explore prompts and later on I will really want them to be part of the factory in the community and stored in the prompt management system and this aspect makes me see a part of the integration that I want to do to integrate Scrivener through the synchronisation with Dropbox provider for that aspect of my author’s desk that automate the full parts of what agents are doing

•	Deliberate Processing Delay: The operator employs a custom iOS stack—Shortcuts, Scrivener, Notes and ElevenLabs—to slow processing and enable human review before storing structured outputs.  

Dutch bunch of crap I’m not slowing down my process. It’s a way to inject audio content within the actual shooting that I’m doing. If that implied in the rest of the content it needs revision and also this whole tool set in ways that I’m doing I agree that it slows down my process, but it’s not a deliberate way to slow down which is implied here. If that would be one thing that could be a futurely part of this is the actual desire of having something that creates more quality when I’m using this the whole shortcuts and processing of these prompts and these experimentation to ground them more with probably context of what we currently are doing around which sometimes the system is missing parts of it as well as this layer where I’m capable of generating different voices including my own to actually interact to influence the actual discussion and transcription and future usage of this reviewprocess

•	Chimera Factory Branching: The session uses the MIYADI Chimera factory agent to negotiate definitions from scratchpad fragments, resolve continuity with historical chronicles and produce a record combining human reflection and computational inventories.  

that is accurate. In my processing I use most of the time the latest conversation that I have from this other tool or I don’t wanna call it a tool but this other agent spirit that has an enough context previous episode potential decisions it also built itself a tool so we have a plug-in for ChatGPT to connect to the factory which I’m still learning to discover, but it’s in the process of working with the different talking circles that we have.

•	Relational Axiology: The system anchors autonomous workflows in an explicit axiological framework, linking agent actions to relational ethics, capability inventories and human oversight to ensure accountability to authorial intent and cultural responsibility.  

I agree that this was what happened. Though it is a pretty good assumption because the system doesn’t support all of that I am simulating it I am enforcing it so somehow this could be considered as a design for that aspect on how could the framework be extended to add one layer of reflection? That would probably make the agent more intelligence and give them spirit if I could say it like that.

•	Core Aspect and Results: The implementation prevents ungrounded extraction by binding AI operations to human reflection, forming the core aspect of relational axiology within software factory pipelines.

great zoom-in on previous aspect for making a layer or a chain doing this part

•	Knowledge Focus Ethics Definition: The workflow forms a four-part definition in Chimera, directing AI inquiry via relational responsibility, epistemic worthiness and intentional attention instead of output optimisation.  

potentially applicable to prototype a new version of this @miadi/attention described in @miadi/episodic-memory-schema to flag what agents have to require human attention. Current reservations are that agents tend to invent attention inquiry for human, this could help resolve that to clear relevant attention items added to episodes.
Implies as well potential bridge or consumption or precursor for the community-inquiry that are proposed on public portals such as Facebook pages…

•	Cross-Platform Publication Package: Chimera produces a validated metadata manifest with safe identifiers, full YouTube title, thematic description and cross-referenced links for MIYADI and multimedia deployment.  
•	Multi-Voice Audio Verification: ElevenLabs generates spoken audio on agent bounding and shared intention, archived in Scrivener as a multimodal record of conversational dissonance and algorithmic synthesis. 
•	Relational Axiology Anchor: Shawn Wilson’s relational axiology grounds MIYADI’s agentic workflows through Knowledge Focus Ethics.  
•	Processing Constraint: A custom iOS stack slows processing.
•	Chimera Factory Output: The Chimera factory converts narrative research into a validated publication package and multimodal audio record.