Status note: this is a design, not a building. Nothing described here has been built, no property is under offer, and no funds are committed. Several of the questions at the end of this post are genuinely unresolved. We publish at this stage on purpose — it is easier to tell us we are wrong now than after we have signed for a building.
Twelve households living separately do the same work twelve times over.
Twelve bathrooms get cleaned. Twelve suppers get cooked. Twelve people make the same drive to the same grocery store in the same week. Twelve families solve childcare privately, badly, at the last minute, by calling someone. None of that work is paid. Most of it is not even counted — it happens behind twelve closed doors, and from outside each door it looks like nothing is happening at all.
Put those twelve households on one site and the work does not increase. It becomes visible. And work that is visible can be scheduled, shared, and paid for.
That is the whole idea, and it arrived from the other direction entirely. The original sketch for this project simply observed that on a small residential site of about twenty-four people, roughly twenty-one of them would end up doing something useful and compensated — without even trying. That ratio sounds impossible for a housing programme. It is only impossible if you think the site has to create jobs. It does not. The work is already being done. The site just stops hiding it.
This post is about what we would build on that observation.
First, a word about the name
For most of its life this project was called transitional housing, and we have just stopped calling it that. The reasoning is short and it tells you what the thing actually is, so it is worth a minute.
The old name stuck for an honest reason: the conversation started with the housing. Somebody needs somewhere to live, so you begin by thinking about rooms, and a name given in the first hour has a way of outlasting its usefulness.
Two things were wrong with it.
“Transitional” implies a clock, and we are not running one. As you will see further down, nobody here is ever asked to leave. A word that promises an exit date describes a different project from the one we are building.
And “housing” points at the less interesting half. The longer we worked on this, the clearer it became that the part that matters most is the part facing the road — the shop, the cafe, the laundry, the warm room that anybody in the county can walk into without being a resident or a client or a case. Those are not amenities bolted onto a housing project. They are arguably the point, and the housing is what makes them possible.
There is a third thing, and it is the one that decided it. Nearly every name in this category names the residents’ problem. Halfway house. Recovery house. Transitional housing. The label is a diagnosis, and it attaches itself to the people living under it — which is precisely what the rest of this design spends its length trying to undo. Your address should not be a disclosure.
So we are calling it a Holon House. It is the holon’s house: named for the community it belongs to rather than for anything that is supposed to be wrong with the people inside it. The arrangement it describes is not new, either. A public room downstairs that anyone may walk into, with rooms above it, is one of the oldest building types there is — an inn, a coaching house, a public house — and nobody has ever found it strange or charitable.
We like the name. We are not married to it. If a better one occurs to you while you are reading, tell us.
A finding we are stating up front
We put this design through our own adversarial review before writing any of it down, and one answer came back clearly enough that we are going to publish it as a conclusion rather than leave it as a question:
Refurbish an existing building. Do not build new, and do not build emergency shelter.
Not flat-pack cabins. Not a campground. Not a new-construction village of small units. Somewhere near you there is already a fading motel, or a school a county closed in a consolidation and now pays to heat for nobody. Take one of those.
We are stating it flatly because we do not think every holon should have to spend a month rediscovering it. Here is the reasoning, so you can check our work rather than take our word:
- The expensive part is already paid for. The life-safety package in a building that once housed the public — the fire separation, the egress, the sprinklers, the commercial-grade plumbing and electrical service — is the single largest line in new construction, and in an existing building somebody else bought it decades ago. You are buying their compliance along with their roof.
- The acquisition can cost less than the fixtures inside it. A failing motel’s entire purchase price is routinely less than it would cost to install the bathrooms it already contains. You are not really buying a building; you are buying the plumbing, and the building comes attached.
- Changing a building’s use is a gentler regulatory path than creating one. Most jurisdictions judge a change of occupancy against the work you are actually doing, rather than demanding that an old building meet every standard written since it went up. Converting lodging to housing tends to stay inside the same family of rules.
- The seller wants rid of it. Counties sitting on closed schools are paying maintenance and utilities on empty buildings and are often carrying an asbestos liability they would dearly like to hand to somebody else. That reframes the whole conversation: you are not a buyer competing for an asset, you are a party accepting a liability — and the purchase price can credibly be nominal.
- The alternatives are worse than they look. Campground-style siting is closed outright in our state, where regulations cap transient occupancy at thirty consecutive days and prohibit any camping unit being used as a permanent abode. Rules vary, so check yours — but check before you fall in love with the idea, not after.
And one genuine counterweight, because the finding is not free: refurbishment estimates have fat tails. Abatement in particular can be a surprise measured in hundreds of thousands of dollars, and you do not know what is behind the walls until somebody opens them. The survey comes before the offer.
What we are deliberately not publishing is a property list or a price. Those vary enormously from one holon to the next — building stock, local construction rates, what the county is willing to do — and a number from our county would be actively misleading in yours. The search and the costing are each holon’s own work. The finding above is what we can usefully hand you so that work starts from the right place.
What the place is
A small residential site, in a county where housing is scarce and buildings sit empty.
Around twenty-five to thirty small units. Not apartments — rooms, with what a room needs. A shared laundry. A kitchen. A garden. A van, if somebody donates one, running a route into town for the people who cannot drive.
And attached to all of it, facing the road: a Cornucopia and a cafe, open to the public. Not open to residents. Open to everyone, exactly as if they stood alone on that lot.
That last decision is the one that decides what the place is, so it is worth being plain about why.
Why the commercial parts face the road
A residential site serving people who have nowhere else to go, with a fence around it and nothing on it that anyone else has a reason to visit, is a compound. Everybody who has ever lived in one knows what that feels like, and everybody driving past knows what they are driving past.
Put a shop and a cafe on the front of it that the whole county uses, and the building stops being a place where those people live. It becomes the place where you get coffee, which happens to also be where some of your neighbours live. The residents are not an exhibit and they are not hidden. They are just there, the way people are there in a town.
This is the same inversion we described in Cornucopias, where the shop exists to pay for the room rather than the room existing to serve the shop. Here the public frontage is not an amenity bolted onto housing. It is what keeps the housing from being an institution, and it is also, not coincidentally, where a large share of the money comes from.
There is a second reason, and it is less noble and just as real. A site like this needs income that does not come from its residents, because its residents are the people who do not have any. A shop and a cafe serving an entire county can carry costs that twenty-nine low-income households never could.
The work that pays for it
Nobody here is hired.
The network already runs on quests — discrete Acts of Service, peer-rated, paid in Ghost Net Credits. Minding the Cornucopia for an afternoon is a quest. So is the laundry, the garden, the meal, the run into town. These are the same quests the network runs everywhere else; the difference is only that on a site like this there are a lot of them within walking distance of your own door.
Two things follow that we want to state clearly rather than leave implied.
Completing quests is not a condition of living here. Your rent does not change if you do fifty of them, and it does not change if you do none. Housing and work are separate transactions and we intend to keep them separate to the point of pedantry, for reasons we will come back to.
Quests must be genuinely declinable. The day there is a roster with consequences for missing a shift, this is a job with extra steps, whatever we call it. If someone does not want to cook on Tuesday, they do not cook on Tuesday.
What we are deliberately not running
We do not run childcare.
That surprised us too, because childcare is one of the most obvious needs a site like this has and an obvious thing to organise. We are not going to, and the reason generalises past this one decision.
Families on the site decide among themselves who watches whose children and what that is worth — ad hoc, peer to peer, the way neighbourhood teenagers have done it since there were neighbourhoods. The moment we formalise it, we are an institution providing care to minors, with everything that carries. We would have built the single largest liability surface on the site in order to organise something people are entirely capable of organising.
This is the Cornucopia doctrine arriving in a different room: create the conditions and let people arrange themselves, rather than building the institution. If a site full of neighbours cannot work out childcare between them, the problem was never the absence of a programme.
Rent is thirty per cent, always
Rent is thirty per cent of adjusted income, calculated the way HUD calculates it. It is payable in dollars, in Ghost Net Credits, or in both.
Thirty per cent if we have given you a dozen quests this month. Thirty per cent if you have done nothing at all. Thirty per cent if you leave the network entirely and want nothing more to do with us.
We want to explain that rule properly, because it looks like an administrative detail and it is actually the load-bearing wall of the whole design.
Here is the thing we are most worried about building. An organisation houses you. It pays you in a currency only it issues. You spend that currency at shops it operates. It deducts your rent at source. In Russell County, Virginia, that is not an abstraction or a historical analogy — that is the company store, and for a lot of families here it is grandparents.
So: what stops this from becoming that?
Coercion requires that leaving costs you your home. Under a thirty-per-cent-always rule, losing the connection to Play the Planet makes your rent go down, not up. Your income falls; thirty per cent of a smaller number is smaller. The exit is never expensive. That is the difference, it is the only thing load-bearing enough to be the difference, and it should be stated in public rather than buried in a framework document — which is why it is here.
Two further commitments in the same spirit. Rent is always payable in dollars, so nobody is ever obliged to earn Gc to keep a roof. And Gc has to be spendable where we do not own the till. If the only places that take Ghost Net Credits are places Play the Planet operates, it is not a currency, it is scrip, and we will have built the thing we said we would not. Getting independent local merchants to accept Gc is not a nice-to-have on the roadmap. It is the condition of the design being honest.
Nobody is asked to leave
Here is the part we expect to be misread, so we will be blunt.
This is minimal housing. Small rooms, shared facilities, a modest building in a small county. We are not building a resort and we are not going to pretend otherwise. For some people, that will be exactly enough — a place to stand still, close to work that matters, among neighbours. Some will find a calling in what happens here and stay for years.
Others will not. They will begin to chafe at the size of the room and the length of the shared-laundry list, and they will want more than this place is ever going to be. That is a success, not a failure. It usually means somebody arrived with nothing and now has enough of a floor under them to want a ceiling somewhere else.
So the programme does not end when we decide it ends. It ends when you decide to go, and not before. There is no completion date, no review board, no graduation we can fail you out of. We are never going to tell somebody that their time is up and they have to leave the only housing they have.
We want to be unambiguous about the far end of that, because it is the reason the old name had to go. You may stay as long as you wish. If that is the rest of your life, then it is the rest of your life. If this place suits you — if the work here turns out to be the thing you were for — then you are not a resident who failed to move on. You are home, and that was always one of the acceptable answers.
And if you outgrow it, go and grab the brass ring, and we will be glad.
What we can do is make leaving easier than it usually is. A share of rent funds a completion grant — money that goes with you when you move on, so that the hardest part of leaving, which is first month and deposit and a van, is not the thing that keeps you in a room you have outgrown. The grant is the organisation’s money rather than an account in your name, for boring regulatory reasons. We intend to publish the conditions that earn it, including a list of departures that qualify in full — a better job somewhere else, a family emergency, a medical reason — so that it never becomes a lever we hold over anybody.
The honest version of the arithmetic is that long stays and new arrivals compete. Every resident who stays is a room the next person does not get, and a site where everyone stays forever houses exactly as many people as it has rooms. That is a real cost and we are not going to pretend otherwise.
We have simply decided which way to be wrong. The alternative — a deadline — solves the throughput problem by moving a specific named person out of their home on a date, in order to make space for a person who does not exist yet. We are not doing that. If we need more rooms, the answer is more rooms, which is slow and expensive and honest, rather than a turnover policy, which is fast and cheap and is somebody’s eviction.
What it costs, and why we would rather wait
Now the part that most writing about projects like this leaves out.
Our working estimate to acquire and refurbish a building of this size is somewhere around two million dollars. A motel-scale conversion models at roughly one and a quarter to two million. A large building with heavy abatement — an old school, say — can run well past that. These are estimates built from published benchmarks, not a builder’s bid, and we would expect a real number to move in either direction.
That is our county’s number, not yours. Building stock, local labour rates and what the seller will accept differ enough between places that quoting ours as a budget for yours would be a disservice. Price your own building.
Two things about the figure travel better than the figure does.
The first is a trap in how you lay the units out. The cheap-sounding design — small sleeping rooms with a shared bath house, no toilet and no kitchen in each unit — is the expensive one. Strip the sanitation and the cooking out of a unit and it stops being a dwelling in the eyes of the code; a row of them becomes a single apartment building, with sprinklers, fire separation, commercial plumbing counts and sealed architectural drawings. To stay in the cheaper classification you often have to put a kitchenette in every unit, even though the cafe is forty feet away. That is counterintuitive, it is expensive to learn late, and it is a question to put to your building official before you draw anything.
The second is what happens annually, and this is the actual conclusion of this post.
Owned outright, with no mortgage, a site like this runs at a shortfall of roughly twenty to sixty thousand dollars a year after rent and the commercial operations are counted. That is a small number. It is one church partnership. It is one recurring grant. It is a figure a county-scale effort can carry indefinitely.
Now borrow half the capital. A seven-hundred-and-fifty-thousand-dollar note over twenty years, at rates a small nonprofit would actually be offered, costs about seventy-two thousand dollars a year, every year, for twenty years. That single line is larger than the entire operating shortfall of the building owned outright — more than three times larger, in the cheaper of our two models. The mortgage does not add to the problem. The mortgage becomes the problem, and everything else on the site spends two decades serving it.
So the conclusion we have reached, and the reason this post exists, is this:
It is better to wait and own it than to hurry and owe it.
A project like this done on debt needs continuous large-scale fundraising forever and is one bad year from failing the people living in it. The same project bought outright needs ordinary community generosity and can survive a bad year without anybody losing their home.
That is not a counsel of patience for its own sake. It is a direction for the network: Play the Planet needs to build real cash-generating capacity — businesses that make money in dollars — so that initiatives like this one can be funded outright rather than financed. The unglamorous commercial work is not a distraction from the mission. It is the thing that makes the mission survivable.
One number that explains why this is possible at all
The published benchmark for operating a village-type shelter is around $29,681 per unit per year. Our model for this site is roughly $4,700.
We are not six times cleverer than anybody else. Essentially the entire difference is staffing — case management, security, a front desk, janitorial, maintenance — and the reason ours is small is that those lines are denominated in Gc and done by residents as quests.
Which means the honest statement of risk is this: if the quests do not reliably get done, or if the work has to be converted to conventional paid employment, this model’s costs migrate back toward the benchmark and it stops working. The quest system is not a pleasant feature of this design. It is the load-bearing wall, and we would rather name that now than discover it in year three.
Who is allowed to build one
Everything above is how we would build. It is not a set of rules for anybody else, and we want to be careful here, because it would be easy to read it as one.
Anyone can build one. We do not license it. We have no permission to withhold and no intention of acquiring any. That goes for the name as well as the building — if you put one up and want to call it a Holon House, call it a Holon House. We are not going to trademark a warm room.
If you manage a holon and you are ready to take this on, do the property search. Walk the dead motel. Call the county about the school. If your holon has a few million dollars, or a benefactor who lives there and wants their name on something that matters, then build it and run it as you see fit. Set different rent rules. Set a different relationship between work and tenancy. Put a time limit on it if you believe that produces better outcomes. We will help where we can, we will tell you what we learned, and we will not interfere.
What we control is not what gets built. It is what we fund.
When the network’s money builds the thing, it is built our way, and the terms are the ones set out in this post:
- The commercial frontage is public — the shop and the cafe serve everyone, not only residents.
- Rent is thirty per cent of income, always, and always payable in dollars.
- Quests are genuinely declinable, and completing them is never a condition of living there.
- Gc must be spendable where we do not own the till.
- And residents may stay as long as they wish — forever, if that is their desire. No completion deadline, no graduation, nobody moved along to improve a throughput number.
That last one is not negotiable in anything we pay for, and it is the term most likely to come under pressure, because it is the one that costs the most. A funder looking at the occupancy figures five years in will have a very reasonable-sounding argument for a time limit. The answer is no.
These are not laws. They are the terms of sponsorship. Break them and nothing is confiscated, nothing is seized, and nobody is punished. The money simply stops. We set out at length in the Cornucopia post why we prefer this to a rulebook, and the reasoning carries over unchanged: it needs no inspectorate, because the only lever is a payment we were never obliged to make; it cannot harden into a cartel, because the door stays open and we have to keep earning our position; and it is honest about power, because we are not pretending to be neutral. We have a thesis about how this should be done, and we will pay for that thesis.
Who decides where the first one goes
We are a long way from this problem and we would like to describe it anyway, because the answer is part of the design rather than an afterthought.
Suppose the network reaches the point where it has the money to do this. More than one holon will need it. Several will have a candidate building, a willing county, and people sleeping in cars.
We think that choice belongs to the membership. Paying members would see the candidates — the building, the county, the costs, what the local partners have committed — and vote on which one gets built first.
We prefer that to the alternatives for a reason that is not sentimental. Whoever picks is making a judgement about whose need is most urgent, and that judgement will be wrong sometimes no matter who makes it. If a small group at the centre makes it, every unsuccessful holon learns that the centre decides. If the membership makes it, the decision belongs to the same people who funded it, and the holons that were not chosen this time are choosing alongside everyone else next time.
Today there is one holon. This is a mechanism for a network that has many, described early so that it is a commitment rather than a convenience.
We want to be proven wrong
Everything here is our operating thesis, not a doctrine. It is what we will do until somebody shows us better.
So if you build one privately and run it on different rules — a time limit that genuinely helps people move on, a different split between rent and work, a configuration we have not thought of — we want to know, and we will change what we fund.
But bring your A-game. We will not change on the basis of theory. We will change on the basis of what is demonstrated to work. A network that re-aimed its money at every persuasive argument would not have a thesis, it would have a weathervane, and the people whose housing depends on this being here in ten years would pay for that.
This is the real reason the open door matters more than fairness does. The people doing it differently are the only way we ever find out we are wrong. If we shut that door we would not merely be high-handed — we would be blind.
What we have not solved
In the spirit of the rest of this corpus:
- Whether quests on a residential site stay quests. Federal wage law turns on economic reality, not on the words the parties use, and the more successfully Gc functions as a living wage the more it resembles one. The separation of rent from work is our strongest protection and it is already a rule. This nonetheless needs a few hours of a Virginia employment lawyer before anybody occupies a room — not before we keep designing, but certainly before that.
- What the residents’ legal status actually is. If residents are tenants in the ordinary sense, a body of landlord-tenant law applies that we have not yet worked through. Our rule that nobody is ever asked to leave removes the sharpest version of this question, and does not remove all of it.
- What share of rent may arrive in Gc. Our costs — a tax bill, an insurance premium, a utility — are dollar-denominated and arrive on fixed dates. Resident payments are not. There has to be a cap on the Gc share of rent, or a rule for deriving one, and we have not set it.
- Who this is for, precisely. Families with children, single adults, people leaving incarceration, people fleeing violence, and elderly people on fixed incomes are five different buildings with five different programmes. Our cost model assumes families. Changing that assumption changes nearly everything downstream of it.
- Whether the cafe should really be pay-what-you-want. It is the most public-facing thing on the site and the strongest answer to the company-store problem. It is also modelled, generously, at break-even. Those two facts pull in opposite directions.
- Which funding category this falls into. Renaming it solved our problem and not the paperwork’s: grant programmes carve transitional housing apart from affordable housing, and a Holon House — with no exit date and no programme to complete — may read as neither. We would rather describe honestly what we are building and let the label follow than design the thing backwards to fit a category, but we should go in knowing that some doors may be closed by a definition rather than by a judgement.
- How many rooms a site should have, given that nobody is moved along. Our throughput is whatever people choose, which makes the question of the right size genuinely hard — too small and the waiting list never moves, too large and we have built an institution by accident.
- Which building, and at what price. This is the one piece of work no holon can inherit from another, and we have not finished ours either. It starts with a phone call to a county that owns empty buildings, and the three questions worth asking are what is still available, what it actually measures, and what the asbestos survey found. It is a phone call, not a research project, and until somebody makes it every capital figure in this post is a model rather than a plan.
We would rather publish that list than pretend it is shorter.
This is part of Play the Planet, a network trying to make a functioning local economy out of what a community already has. If you want the wider frame, start with the Executive Summary, The Two Currents, and Cornucopias.