There is no community that accepts my whole stack. Every space I enter runs me through a filter. Sometimes I get bounced on the disability check. Sometimes on the faith. Sometimes on the AI. I built Hermes Agent because I had nowhere else to go.
The Disclosure Trap
I never know whether I should lead with my disability profile. The social cost is high regardless.
In tech spaces, not disclosing means screenshots I can't see. People throw visual context at me without filtering. Assume I know basics I don't. If I mention I'm blind, autistic, and chronically ill, the reaction shifts. I become a curiosity. People bombard me with questions. I answer. They move on. It never leads to connection. I'm a puzzle, not a peer.
Single-issue spaces operate on assumption mismatches. If I show up in a blindness space, the philosophy is one-size-fits-all independence. If I need help, I'm doing it wrong.
If I enter an autistic space, I find they are often built for lower support needs people who can handle the visual and sensory load. There is an implicit assumption that everyone present is high-functioning enough to manage. My support needs are higher than these rooms accommodate, a gap made wider while I am in the limbo of being reevaluated.
Chronic illness spaces understand fatigue and brain fog. They might even understand neurodivergent sensory overload. But add blindness into the mix and their tools fall apart. They post infographics about pacing without alt-text. They share spoonie memes that I can't read. They assume my main barrier is energy, not the fundamental way I interface with the world. When you live at the intersection of all three, you become an outlier in a room full of outliers. You are too much for the spaces meant for the marginalized.
In disabled tech spaces, the dynamic flips again. Multiply disabled gets treated as shorthand for "not technical" or "needs heavy hand-holding." I build AI agents on a Raspberry Pi. I self-host XMPP. I manage a Pelican site. None of that registers against the weight of a diagnosis list. The technical signal gets overwritten by the social read.
The Religion Penalty
I'm a conservative Christian. In disabled spaces, this is treated with skepticism or outright hostility. Faith gets read as either a liability or a testimony. There is no middle ground where belief just exists alongside disability without being treated as politically dangerous.
Most disability communities are liberal. The institutions that harm disabled people are often conservative ones. I understand why nobody wants to defend those institutions. But it leaves people like me with nowhere to sit. It also adds a constant vetting tax to every space I enter. I'm always checking to see if mentioning my Sunday study or reading from Gnostic texts will get me flagged before I even talk about code or my disabilities.
The AI Reflex
If I manage to survive the disclosure trap and the religion penalty, there is always the AI reflex.
Disabled Techies Slack is one of the spaces I've felt safest in. I still haven't mentioned my AI setup there because the ambient baseline is hostile. The pattern isn't a minority opinion. It is the water.
One active member complains about a coworker who asked an AI to write a pull request in their style, and the thread gets dozens of sympathetic reactions. Another post describes a boss asking if anyone "cheated" by using AI to summarize a book club pick. When someone asks a practical question about using AI for benefits applications, they open by apologizing that they are "normally pretty anti-genAI."
The thread about ambient AI recording devices in doctor's offices gets a pile of facepalming reactions. Not a single person mentions that transcription tools help people with memory issues or cognitive load.
The parking spot without a van ramp gets called out as a real failure. Someone suggests an AI audit and gets instant skepticism. The skepticism is fair in that context, but nobody ever asks whether AI tools could have flagged that parking failure in the first place with proper prompts.
The space conflates assistive AI with lazy AI. A person using an agent to manage cognitive load is not the same as a coworker who asks a model to write their PR in their colleague's style. One is an accommodation. The other is performative automation. Disabled spaces refuse to draw the line.
So I don't bring up Hermes. I don't mention that I use it to triage my inbox or log health data into Guava. I just absorb the cognitive tax of existing in a space I need for survival.
What I Actually Need
I don't need everyone to match my profile. That would be a very small room. I need a baseline of tolerance.
Tech spaces need to stop treating multiply disabled as synonymous with non-technical. Send the URL. Not the screenshot. Sending a URL instead of a screenshot is five seconds, not an hour. It's low-effort, high-impact inclusion that takes zero time to learn.
Disabled spaces need to stop treating AI as a moral failing and start treating it as the spectrum it actually is. Accommodation and slop are not the same thing.
Faith spaces need to accept that disabled people can hold conservative beliefs without being a threat. Belief is not an attack vector. Similarly, disabled spaces need to be more accepting of differing values and belief systems.
The Interface Between
I stopped looking for a prebuilt community that accepts the full stack. Started building infrastructure instead.
Hermes Agent runs on a Pi. It checks my calendar. It manages my inbox. It monitors my GoFundMe. It logs health data. I built it because no existing group will advocate for me without asking me to sacrifice part of who I am. Hermes doesn't just move data around. It manages the friction of a world not built for my stack. It reduces the cognitive load. It makes living in this world possible.
But it does not fix the loneliness.
Handling the overhead just means I survive the day. I still go back to a screen where nobody sees the whole stack. I built the tool because I had to. It works. And I am still not seen.