Stimpunks Foundation

Mutual Aid and Human-Centered Learning for Neurodivergent and Disabled People

Plant Growing in the cross section of a log

Stimpunks.org Changelog for Week 17 2026: From Ethics to Architecture, From Harm Reduction to Power

🗺️

Home/DEI-AB / Stimpunks.org Changelog for Week 17 2026: From Ethics to Architecture, From Harm Reduction to Power

▶ Table of Contents

  1. AI Collaboration: From Guardrails to Architecture
  2. Star Stuff and the Footer
  3. The Esmx Bookmark
  4. Campfire, Infodumplings, Covenant, and the Cult of Compliance
  5. Field Guide: Wheel of Life
  6. The Throughline
  7. Changelog

Week 17 was a week of turning ethics into architecture.

The AI Collaboration ethics page has been growing since it launched. This week it grew up. Over seven days we added more than a dozen sections — on harm reduction, disability justice, engineered exclusion, the double empathy problem in code, art as community representation, MADTech and constructionism, fascism and POSIWID, co-design as political demand, the tensions we actually sit with. The page went from a list of guardrails to a document with a position.

Meanwhile, we published star stuff, printed an Esmx bookmark, updated the Cult of Compliance with a death that demands accountability, and held a Campfire on Dr. Fatima’s “How to (Anti) AI Better.”

If there’s a single thread: we stopped describing our relationship to AI and started analyzing it.

AI Collaboration: From Guardrails to Architecture

The AI Collaboration at Stimpunks page received more substantial development this week than in any previous week. Individually, each addition was incremental. Together, they form an argument.

It started with Dr. Fatima’s “How to (Anti) AI Better” — harm reduction as a framework for individual AI use, disability and accessibility as the primary ethical frame, and a strong case against shame as a change strategy. These aren’t soft positions. They determine what ethical AI use looks like when the people most harmed by AI systems are also the people most likely to benefit from assistive applications of them.

From there, the page expanded into territory that had been waiting.

We added guardrails on AAC users and nonspeaking people — naming engineered exclusion as a pipeline problem, not a user problem. We added “From Engineered Exclusion to Designed Dignity” and “The Double Empathy Problem in Code,” both drawing on Srinivasan (2026), AI, autism, and the architecture of voice (AI & Society). The algorithmic double empathy failure framing extends Milton’s original concept into AI systems: a model trained on normative speech misreads AAC-mediated communication as incoherent, while the user experiences the model as unresponsive. Neither side is broken. The system is.

The art guardrails sharpened too. We added a brief rationale — “Art on this site is a form of community representation and artist support, not just content generation” — and then built it out with a full “Why We Restrict Generative AI for Art” section covering community authorship, art as identity testimony, the training data problem, epistemic weight of visual content, and soul preservation. We also added a “Hybrid and Assistive Use” section that defines what assistive AI art use we do permit, under what conditions, and with what attribution requirements. The access argument — that GenAI lowers barriers for disabled artists — gets direct engagement rather than dismissal.

Then tante’s “AI as a Fascist Artifact” arrived, and the page needed to meet it. We added alignment sections, a POSIWID guardrail, a “Name the Systems of Power” section, and a subsection under “The Tensions We Sit With” naming the data acquisition violence with more precision. We added “The Tensions We Sit With” itself — a section that holds the genuine contradictions in using AI tools as a disability-justice organization without resolving them falsely.

We added the Human Restoration Project’s “Using AI Without Losing Ourselves” — Freirean critical consciousness as AI literacy, aligned with our harm reduction and protection framing. We added “MADTech, Constructionism, and the Tool Belt” featuring Trevor Aleo’s Down with EdTech! Long live MADTech! — the distinction between technology that serves institutional management versus technology that serves learners. We added “The Biggest Guardrails: Anthropomorphization and Sycophancy” as a standalone section. We added “Co-design Is Not a Strategy. It’s a Political Demand.” drawing on Tim Dixon’s AI Accessibility Barriers and the Business Disability Forum’s polling on disabled people’s centrality to AI accessibility work.

We added alignment sections for both “AI as a Fascist Artifact” and “The Majority AI View.”

The page is no longer a set of guardrails. It’s a position on power, design, and what it means to use contested tools ethically when your community is the one those tools were not built for.

Just for fun, we published Love You Down To Your Star Stuff.

Then we added a star stuff section to the site footer. It’s there now at the bottom of every page.

Some things belong everywhere.

The Esmx Bookmark

We published Esmx Bookmark — a print-ready bookmark featuring Esmx the Porkypine, our mascot, with Manifesto lines on the back. Available in six PDFs: single, 3-up A4, and 6-up A3, front and back.

55mm × 190mm. Dark palette. Atkinson Hyperlegible. Philly Pride accent bars. Esmx landscape on the front, rotated and rising on the back, Manifesto lines above.

The bookmark page lives under /about/brand/, alongside the business cards.

Campfire, Infodumplings, Covenant, and the Cult of Compliance

We held a Campfire Learn Together on Dr. Fatima’s “How to (Anti) AI Better” and published the event post: Campfire Learn Together: “How to (Anti) AI Better” by Dr. Fatima.

We published Infodumplings: Interstitial Journaling and Star Stuff, the recap from our weekly community gathering.

The Covenant received an “Everything is politics, but…” section clarifying what “civil” and “professional” mean in our context — drawing directly on the Contributor Covenant FAQ to prevent those terms from becoming tools of respectability politics.

The Cult of Compliance and the Policing of the Norm was updated with quotes from Alex LaMorie’s Death: Autism Training for Police Is Not Enough. Training is not enough. The page now says so more clearly.

PEGGY’s “ALICE” was added to Infodumplings. Some rabbit holes are worth marking.

Field Guide: Wheel of Life

A soft-toned circular diagram titled "Autistic Realms – Wheel of Life: Neurodivergent & Disabled Wheel of Life". The wheel is divided into 12 labeled segments, each representing a life area relevant to neurodivergent and disabled people. The segments are: Personal Development, Interests and Flow, Energy Levels / Spoons, Relationships, Sensory Environment, Finances, Family and Home, Rest and Recovery, Mind, Body, Soul, and Work. Each segment radiates outward from the center in shaded layers, resembling a flower or radar chart, suggesting a score from 1 to 10. At the bottom, reflective prompts ask: “Have a think about what score out of 10 you would give each area of your life, or colour in where you feel you currently are. What areas would you like to focus on?” A footer notes that this is a neurodivergent and disabled-affirming version reimagined by Autistic Realms, adapted from the original concept by Paul J. Meyer.

We added a “Neuro-Affirming Neurodivergent & Disabled Wheel of Life” section to Regulation & Coping. The wheel isn’t neurotypical wellness metrics dressed up in new language. It starts from the actual shape of a neurodivergent and disabled life.

The Throughline

  1. AI Collaboration page transformed — fourteen-plus additions across harm reduction, disability justice, engineered exclusion, double empathy in code, art as identity, MADTech, constructionism, POSIWID, power analysis, co-design as political demand, and genuine contradiction. The page now has a position, not just a policy.
  2. Srinivasan (2026) integrated — “AI, autism, and the architecture of voice” provided the structural framing for how AI systems fail AAC users and nonspeaking communicators as a design problem, not a user problem. The algorithmic double empathy failure concept is now part of our public ethics language.
  3. Star stuff published and embedded — “Love You Down To Your Star Stuff” went live and moved into the site footer. It’s at the bottom of everything now.
  4. Esmx Bookmark published — print-ready, Manifesto-backed, available in six PDFs for A4 and A3 printing. The brand asset library grows.
  5. Campfire on Dr. Fatima’s talk — held and documented. Harm reduction, disability, AI literacy as protection. These ideas are now in the Campfire record and woven into the AI Collaboration page.
  6. Field Guide expanded — Regulation & Coping has a neuro-affirming wheel of life. The Covenant has a politics clarification. The Cult of Compliance page names what training cannot fix.

Week 17 moved ethics into architecture. The AI Collaboration page is no longer an addendum to our values. It is the values, at work, in tension, naming the contradictions and holding them anyway.

Changelog

Posted

in

by

Stimpunks Foundation

Tags:

Comments

Leave a Reply

Discover more from Stimpunks Foundation

Subscribe now to keep reading and get access to the full archive.

Continue reading