This page lets you ask our website questions using generative AI.
The tool below has been grounded in our content — our glossary, our frameworks, our research, our values. It’s a way to navigate a large site, find connections between ideas, and explore what we mean by things like niche construction, regulation-first design, or the neurodiversity paradigm.
We use AI deliberately and with clear limits. Read how and why at our AI Collaboration and AI, Disability Justice, and the Ethics of Making pages.
A note on accuracy. Generative AI can misstate, confuse, or confidently produce things we didn’t say. Treat results as a starting point, not an authoritative interpretation of our work. If you need a human read on something — especially anything that touches your situation, your community, or decisions that matter — reach out to us directly. We’re here.
Not sure what to ask? Scroll down for suggestions.
This is an experiment in knowledge gardening. Want to know more? Scroll down.
An Experiment in Knowledge Gardening
Our Ask page is an unadvertised experiment some in our community are conducting. It is based on the generative AI built into wordpress.com. It is not fully conversational, so don’t expect to go back-and-forth in conversation with it. I (Ryan) think it’s strongest use will be as a discovery tool for the rest of the site, as a knowledge garden spider.
Use this page in the spirit of testing and experimenting. Feel free to share your experiences. We appreciate and encourage flow sharing. I’m openly curious how folks receive and use this page — how well it serves, how it harms.
Disclaimer: I haven’t tested this nearly enough, so I don’t know how well being backed by our website comes through in the result. It could get things terribly wrong. We’ll list problems and disclaimers on this page as we test.
Trails to Try
These questions are written in the language of the moment before the word. You don’t need our vocabulary to ask them — bringing the experience is enough. Routing it to language is the spider’s job.
Ask them as written, or better, ask them in your own words.
Notes:
These double as our test suite. Each question is a probe of whether routing actually works — does “why does switching tasks hurt” land on monotropism? Does “gifted and failing” reach spiky profiles? Does “compliance attached” find the pathology paradigm and mutual aid pages? When a routing fails, that’s a finding for the disclaimers list log: question asked → page expected → page reached.
Attention and Energy
- Why can I focus for six hours on one thing but can’t start a five-minute task?
- Why does being interrupted feel like physical pain?
- Why am I exhausted even when I did everything right?
- What is it called when rest doesn’t fix the tiredness?
- Why do I fall down research holes for things nobody asked me to learn?
Senses and Body
- Why do fluorescent lights and crowded rooms make me feel sick?
- Why can’t I tell I’m hungry until I’m shaking?
- Why do I rock, tap, pace, or hum — and is it bad?
- Why does a clothing tag feel like an emergency?
- Why does everything get loud all at once, and what happens to me after?
Communication and Relationships
- Why do conversations go wrong in both directions?
- Why am I a different person at work than at home, and why is it so exhausting?
- Why does small talk feel like a test I didn’t study for?
- Why do I lose my words when I’m overwhelmed?
- Why do my closest friendships start from shared obsessions instead of small talk?
Identity and Belonging
- Is something wrong with me, or with the places I have to spend my days?
- Why did learning I’m Autistic feel like grief and relief at the same time?
- Where do I find people like me?
- What if the gender boxes don’t fit me either?
- Why do I trust the family I chose more than the family I was given?
School, Work, and Systems
- Why am I “gifted” and failing at the same time?
- Why does the open office destroy me when it seems fine for everyone else?
- How do I ask for what I need without it being used against me?
- Why do supports always seem to come with compliance attached?
- Why does every form ask what’s wrong with me instead of what I need?
Every one of these questions has pages in the garden that name it. The vocabulary exists because people asked questions like yours, found no words waiting, and made them. Ask. The trail is there.
The Spider in the Garden
Search engines call their crawlers spiders because the web is a web. Our site is not a web in that sense. It is a knowledge garden — a richly linked landscape where ideas spread sideways, cross-pollinate, and connect across domains. A garden doesn’t need a crawler that indexes pages. It needs a spider that walks trails — one that can take a half-formed question and carry it along the associative paths of the garden until it reaches the pages that name what the asker is living.
That is the job we’re testing this tool for: not an oracle, a guide. A discovery spider for a garden too large and too interconnected to navigate by menu.
The Default Corpus Speaks About Us
Ask a general-purpose chatbot about autism and you get the pathology paradigm back. This is not a glitch. It is the training data working as designed. Large language models are trained on the public internet, and the public internet on disability is written mostly about us, not by us. Clinical literature, parent forums, compliance-based intervention marketing, tragedy framing, deficit checklists. Deficit ideology at corpus scale. The most common things the training data says about Autistic people are things we never said.
This is epistemic injustice, automated. Testimonial injustice: our accounts of our own lives are statistically outweighed by accounts written over us, so the model learns to discount our voice by default. Hermeneutical injustice: the interpretive resources our communities built — the words that make our experiences legible — are thinly represented, so the model reaches for medical-model vocabulary even when community vocabulary fits better. A model trained on that corpus reproduces both injustices at scale, at speed, with confident fluency. Fluency is the dangerous part. The deficit framing arrives sounding authoritative.
A Garden-Scoped Spider Speaks As Us
Scoping changes the epistemics. A generative tool grounded in our garden draws from a corpus that is neurodivergent- and disabled-authored, identity-first, and counter-deficit by construction. The vocabulary garden it walks is the one we planted: monotropism, the double empathy problem, spiky profiles, niche construction. When the spider answers, the conceptual ground under the answer is ours.
This matters most for the seeker who doesn’t yet have the words. Keyword search requires knowing the word. You cannot search for monotropism if no one has ever told you monotropism exists. Our glossary exists to help people find words for experiences that often go unnamed — and a garden spider extends that work to the moment before the word. It can take experience-language (“why does switching tasks hurt,” “why do conversations go wrong in both directions”) and route it to concept-language, dropping the asker at the glossary entry, the pattern, the recipe. That routing — from unnamed experience to community-built hermeneutical resource — is epistemic justice in interface form. It is the part of the work the default training data cannot do, because the default training data does not contain the resources to route to.
An Entrance, Not a Destination
A spider is a guide to the garden, never a replacement for it.
Generative AI can misstate, confuse, or confidently produce things we didn’t say — scoping reduces the source of distortion, not the possibility of it. The garden’s structure is the safeguard: every trail ends at a human-authored page, so every claim the spider makes is one click from the text it claims to summarize. Read laterally. Follow the link. The spider’s best answer is an entrance.
The default training data speaks about us. The garden speaks as us (disclaimer: grounding constrains retrieval, not the model’s priors). A spider that walks our garden carries our voice as far as the gate — and then steps aside. That is the experiment: whether generative AI, bounded by a community’s own knowledge garden, can serve discovery without supplying meaning. How we hold that boundary is documented at AI Collaboration and AI, Disability Justice, and the Ethics of Making.
Crosswalk: Bounded Competence in AI
In 2026, AI researchers adopted monotropism — a theory built to describe Autistic cognition — as an engineering principle, defining “bounded competence” as a safety feature: a model that knows what it doesn’t know. Their fix and ours rhyme. They ground the model’s training data; we ground the spider’s garden. But they bound competence at the level of the model, which fails visibly when it leaves its domain. We bound authority at the level of the interface — every trail ends at a human-authored page, so the spider carries our voice only as far as the gate. We mapped the convergence, and where we go further, here:
Spiders Rock: Our web · Our rhythm · Our way.
Our web · Our rhythm · Our way.
Helen Edgar
Spiders practice niche construction.
