Impact

What the Grant Counts Miss

We open mutual aid grant and creator grant applications twice a year, in January and June. The dates are fixed, published on the grant pages, and announced months ahead. We close each pipeline once it reaches 50 applications.

Both pipelines reach 50 in less than a week.

So we accept 200 applications a year — 100 for mutual aid, 100 for creator grants. Mutual aid sends out one to four US$500 grants a month, depending on our funding level. Creator grants go to up to five people a year. Roughly a hundred people apply for creator support annually. Five receive it.

Those ratios look brutal, and they are. They are also not the whole story, in both directions.

The counts are too small

A grant is not the only thing we give, and for most applicants it is not what they get.

We are a tiny team running a high-touch practice. We stay with people. We arrange healthcare and social services appointments. We help people find services where they actually live. We check in periodically and ask how they are doing. We help people navigate systems that are cruel by design and collapsing in real time.

We do this for applicants we cannot fund. Not being able to send someone US$500 is not a reason to stop knowing them.

None of that appears in the table above, because we have never counted it. We were doing it. It is, by volume, most of what this organization does, and our own metrics are silent about it.

The counts are too confident

We close the door at 50 because a small team cannot hold more than 50 relationships at that depth, and because an open queue is its own cruelty — hundreds of people waiting months to be told no. The cap is not austerity. It is the honest limit of work that does not scale, which is the kind of work we chose.

But it means we do not know how many people need this. The cap fills in under a week, and we never meet the people who arrive in week two.

So every demand number on this page is a floor, not a measurement. Fifty is not how many people needed a grant. Fifty is how many we let in before we stopped counting. We have never seen the size of the need — only the size of our own doorway.

We would rather say that than imply we measured something we did not.

We gather six times a week, every week — four community events and two open meetings. This is not one-off programming. It is a steady pulse, and the consistency is the point: something to come back to that does not require you to have been there last time.

• Watering Hole Hangs — Weekdays, 9AM Central. Drop in, space out. Nature live cams running, chat open, body doubling welcome. No agenda, no reason needed.
• Infodumplings — Thursdays, 7PM Central. Bring your muchness down the rabbit hole — special interests, deep dives, and a room full of people who actually want to hear it.
• Solidarity Sesh — Saturdays, 10AM Central. Peer-run mutual aid and shared experience. Not as patients. As peers.
• Campfire Learn Together — Sundays, 10AM Central. We watch a talk or a film, take a bodymind break, then argue with it together. No hierarchy, no one-way transmission.
• Operations Meeting — Tuesdays, 10AM Central. Where we run the thing. A weekly task scrub — triage, prioritize, move work forward — held in public.
• Editorial Meeting — Thursdays, 10AM Central. Words, design, and everything we put into the world. Writers, editors, designers, and anyone who cares how we communicate.

Both meetings are open to anyone in our community, because we default to open. You can come watch us run the organization, and help. Our learners plan and run events with us, climbing the learning curves as we go.

They do not end when the hour does

This is the part we did not plan.

An event is scheduled for an hour. It often runs for eleven. People arrive from around the world as their own day allows — someone joins as someone else goes to bed — and the conversation keeps moving through whatever medium each person can use that day: voice, video, text, photos, links. Someone talks. Someone types. Someone drops an image and says nothing at all. Every one of those is being there.

That is not a scheduling failure. It is what happens when you stop requiring people to be present in one modality, at one hour, at full capacity. Crip time is not a slower version of clock time. It is time that bends around bodyminds instead of asking bodyminds to bend around it. An event that can run eleven hours is an event nobody has to be on time for.

The room belongs to everyone in it

We open the room. We do not own it.

Our hosts are Autistic, ADHD, and otherwise Disabled and Neurodivergent, moving through the same fluctuating capacity as everyone else here. Some weeks a host does not make it — a flare-up, a shutdown, sensory overload, or simply the weather of living in a Disabled bodymind. The space is still yours. Carry on the conversation, sit in shared quiet, or let the room rest a while. Nothing here needs a host’s permission to happen.

That is not a gap in the plan. It is the plan. A community that only functions when everyone shows up at full capacity is not an accessible community — it just has not met us yet.

We build these as Cavendish Spaces: caves for quiet reflection, campfires for learning together, watering holes for peer connection. Written and async alongside spoken. Interaction badges. Safe foods. Somewhere to sit down and recover. These are not accommodations bolted onto a standard format — they are the format.

More on how and why we gather this way in Our Events Philosophy. Everything lives on the events calendar.

Our Learning Space runs online and offline, and the ten gatherings above happen inside it. Its architecture is Cavendish Space — caves, campfires, and watering holes — built for intermittent collaboration and collaborative niche construction instead of enforced group interaction.

We read other people’s work against ours, in public

Eleven crosswalks now sit inside the Learning Space. A crosswalk reads another body of work against ours and maps three things: where we align, where we diverge, and where a lineage runs.

Some are convergence — Finland’s national vision for schools through 2045, the Autistic-led SPACE framework for inclusive education. Some are ancestry, the traditions we descend from: Highlander Folk School, the Freedom Schools, Fernand Deligny.

We write them knowing they can be wrong, and one of them was. Our crosswalk with Pedagogies of Collapse made a claim we later had to withdraw — the correction is further down this page. A crosswalk that could never be wrong would not be worth writing.

We stopped using the standard three buckets, because they describe a different kind of organization. Program, administration, fundraising was built for nonprofits that hold galas, buy advertising, and employ development staff. We do none of those things, and forcing our spending into those three boxes produced a number that was wrong about us in the direction of looking worse.

Our two paid staff are the program. They run the grant programs and deliver the non-monetary mutual aid — finding services in people’s own localities, arranging healthcare and social services appointments, helping people navigate systems designed to exhaust them, and checking in afterward. Much of it is done alongside people in genuinely perilous situations, many of them without housing. Filing that work under “administration” describes the core of this organization as overhead.

For years our own published figure did exactly that. It said 65% of expenses went to program services. The real figure was 88% in 2025, and 93% so far in 2026. We were marking ourselves down by more than twenty points using a framework that could not see what our staff actually do.

This table cannot see the largest labor input we have. Two of our co-creative directors work full time and draw no salary. There is no row for them above, because there is no transaction to record. Every percentage here is calculated on money, so the table quietly describes a smaller organization than the one that exists — two people instead of four, with two senior people missing entirely.

We are not going to invent a number for it. We do not count the help that is not money, and this is the same problem wearing a suit.

We spend far more than we take in. In 2025 we spent $122,626 against revenue of $32,599. Individual donations across the entire year came to $6,953 — for every dollar a person gave us, we spent $17.64. Our reserves carried the rest, and they are carrying 2026 too.

That is not a shape that holds forever, and we would rather say so here than discover it in public later. If you have ever wondered whether a donation to a small foundation makes any difference to it: at our size, it does.

Where We Fell Short

We told you at the top of this page that you would find reflections on where we fell short. Here they are.

An impact page that reports only what worked is not being modest about the rest. It is broken. So this section stays, it stays specific, and it gets updated when we find new things wrong — which we do, regularly, and in public.

We only quote the people we could help

Every grantee quote on this page comes from someone who got a grant. That is not a coincidence. It is the shape of the sample.

Some people who apply are angry with us, and they have every right to be. They came to us with a real need and we did not fund them. We do not collect those messages, and we do not ask permission to publish them — someone we have just turned down may well apply again next cycle, and a request from us arrives with that sitting on the table, whatever we intend by it. Consent gathered under that much asymmetry is not worth having.

So take this as the counter-testimony instead. The quotes elsewhere on this page are from the people we could help. Most applicants are not.

First-come-first-served rewards being well on one particular day

We already do the obvious mitigation. Our open dates are fixed, listed on the mutual aid and creator grant pages, and announced months in advance. Nobody has to guess.

It is not enough, and we should say why rather than let the advance notice stand in for a fix.

Months of warning let you prepare an application. They do not let you be well on the morning it opens. Fifty slots go in under a week, so the people who fill them are the people who were online that week, holding enough executive function to file, and not in crisis. Those are the exact resources our applicants are least likely to have in any given week. Someone mid-crash cannot sprint. Someone whose internet is down cannot refresh a form. Someone who needs a fortnight to write anything is competing with people who need an afternoon.

We built a race. Our applicants are people the world has already structurally disadvantaged in races.

One change would close it: drawing at random from everyone who applies inside a fixed window, instead of taking the first fifty. We have not made that change. We are naming it here before we have fixed it, because “we are working on it” is the sentence organizations use to make a problem disappear without addressing it.

We do not count the help that is not money

The section above says we give applicants far more than grants. We cannot tell you how much, because we have never tracked it. That is a real gap and it cuts against us — it is the number that would make this page most impressive, and we do not have it.

We are not prepared to start counting it, either, and the reason is not that we have been too busy. This work is organic, relational, and continuous. It has no clean unit. There is no moment a check-in becomes a countable event. There is no line where helping someone find a clinic ends and staying in touch with them begins. Any number we produced would be a fact about our record-keeping, not about the care.

So we are choosing to be uncountable here and to say so, rather than manufacture a metric that would flatter us. If that ever changes, we will tell you when it changed.

We do not know whether our frameworks change anything at scale

Our field guides, glossary, and learning space get read, cited, and adapted. We can count citations. We cannot tell you whether a single school stopped punishing a child because of something we wrote.

Broad institutional adoption has not happened. Long-term outcome tracking we would actually trust does not exist, because most of what exists in that genre measures compliance and calls it progress, and we will not do that to the people we work with. That leaves a real gap. We are not going to fill it with a proxy and hope you do not check.

We track this in our Charting Impact statement and our logic model. What we can honestly say is narrower than what a funder usually wants to hear: ideas we published are in circulation, and we do not know what they changed.

Our care infrastructure does not meet demand

This is the application cap said in money instead of time. We fund what we have. What we have is less than what arrives, every cycle, in under a week.

We publish things that are wrong, and we find out later

In June we published a crosswalk between Pedagogies of Collapse and our Learning Space, written from the book alone. We claimed the framework skipped the Environment ring — the room, its heat, its light, and a learner’s power to change it.

Then we watched a documentary about the school its author actually built, and the footage showed niche construction everywhere: fabric strapped to the roof of a community center unlivable from June to September, a boxing gym used as a regulation space. We had been confidently wrong in public for two months, and it took a video of a building to show us.

We corrected the page. The correction is better than the original claim. It exists only because we went looking.

That is the practice, not an isolated case. We keep a weekly reflection recording what the work cost and what we got wrong at the same volume as what shipped, and we publish it.

What we deliberately do not measure

We do not count hours worked, speed of output, constant availability, or conformity to professional norms. Metric fixation measures what is easy to count under ideal conditions, then blames Disabled people for failing a measurement built to exclude them.

We check in on grantees because we stay in relationship with them, not because they owe us a report. Aid conditioned on proving you deserved it is not aid. No grant here depends on filing outcomes back to us, which means we permanently know less about our own results than a foundation that makes people report.

That is a trade we make on purpose, and it costs us exactly the evidence that would make this page look better.