Koher does not count who uses its tools, does not seek recognition for what it makes, and does not say what its own central decision means. Three refusals, and one reason underneath all of them: each of those instruments would replace a possibility with a fixed value, and everything afterwards would argue from the value.
There are things a practice could measure, resolve or conclude, and refuses to. This is Koher's account of which ones, and why.
The count
There are no analytics on any Koher surface. No server logs kept, no visitor count, no usage figure on any tool, no reach metric anywhere — and no decision here is ever argued from an audience number, because there are none.
The reason is not privacy. Anonymised, aggregated, self-hosted, first-party counting answers a privacy objection perfectly well, and privacy is not the objection. A count does not violate anybody. It replaces a possibility with a figure.
An unknown number can be anything. It can be zero and it can be infinite, and both stay possible only while nobody has looked. Once the figure exists it becomes the thing every later argument runs on — which tool to keep, which one was worth building, whether a quiet month means anything. None of those questions is improved by a number; they are only made easier to answer wrongly.
The figure is not more information than the field it replaced. It is less, held with more confidence.
Where this comes from
The position is older than Koher and it did not start here. It comes from Audience, written for the Museum of Vestigial Desire on 8 October 2013 — Prayas Abhinav's own long-running practice, and a separate thing from this one. Three sentences from it carry the whole argument.
"Who is the other? Who listens when we speak? Is it fair to even expect someone to do so?"
"Intensity and access are inversely related, the more the access the less the intensity and the less the access the more the intensity. … Every ear on which our message is incident dulls our message."
"We do not keep server logs, so there is no way of knowing how many people visit the website. If you do not know a number it can be anything. It can be zero and it can be infinite."
Koher inherits the position and applies it to tools. The middle sentence is why this is not a refusal of reach: access and intensity trade against each other, so the setting gets chosen for each thing made rather than maximised by default.
What may still be known
This is not a refusal to know anything, and the line is precise. Koher may know what it did: what it built, what it released, whom it wrote to, what it put up and on what day. That is memory of its own acts. What it does not get to know is what happened afterwards — who came, how many, whether it landed. Knowing what you did is not measurement.
The refusal is about reception, never about record.
The recognition
The same refusal runs one layer out. Koher does not build toward being recognised — not by a field, not by an institution, not by the apparatus that hands out standing in exchange for behaving in a legible way. It is not against being read. It declines to make the work in the shape that would get it read.
This one has a cost the count does not, and it is worth stating rather than glossing. Recognition is also how a practice is told it has a place. Refusing it means the practice is never told that, by anybody, ever.
The name
In the age of AI, reserving a role for the person and a role for the machine can be seen as an artistic gesture. It can also be seen as nothing of the kind — an ordinary engineering decision about keeping a verdict auditable, made for ordinary reasons. Both readings are available, and Koher does not say which is right.
It is the same structure a third time. An unnamed gesture can be art and can be plumbing, and both stay possible only while nobody has named it. Naming adds no fact; it removes one of the two. And Koher does not deny it either, since this is not an artistic gesture fixes the thing just as firmly, from the other side.
What has to stay true underneath: the split must keep earning its place as engineering whatever anybody reads into it. The moment it were kept for what it signifies rather than for what it does, it would be decoration.
Not three positions. One refusal, met in three places.
What it costs, and what stands in the gap
A practice with no feedback cannot detect its own drift. There is no number to correct it, no reception to disappoint it, nobody's verdict to arrive and say it has gone wrong. That is a real price and Koher pays it deliberately rather than pretending the instruments were worthless.
What stands in the gap is this surface. Positions written in advance, in public, carry more weight here than they would in a practice that could be corrected later — because they are the only thing that can be checked against. A position is a commitment made while clear-headed, that then stands against the wish to be agreed with. It is the same move as keeping judgement in readable code: fix the criterion where it can be seen, before the moment of wanting.
This page exists because the counting does not. That is the whole arrangement.
It is also a deterrent
There is a second reason to refuse the count, and it is the one that connects this position to Asked to Be a Machine.
Nobody can be required to produce the same output every time unless somebody is measuring the output. The demand that a person behave like a machine is not held up by an opinion about people. It is held up by an instrument. Handling time, tickets closed, error rate, forms processed — the figure is what turns a person into a unit, and it is also what makes them comparable to a machine in the first place. Without the number there is no comparison to lose.
So a practice that does not count cannot make that demand, even if it wanted to. Koher does not know who uses its tools, how often, or whether anyone came back. It cannot notice that somebody used it less this week, cannot design to fix that, and cannot tune anything toward regularity. There is no signal to tune on.
That is a structural fact rather than a promise. A tool built to make people more consistent needs a measure of consistency, and this practice has refused to build one. The refusal to count is what keeps the other position from quietly inverting into its opposite.
The demand needs an instrument. Refusing the instrument is not only an epistemic position; it removes what the demand runs on.
The part that is not settled
There is an argument inside this position that Koher has not resolved. Publishing it unresolved is more honest than picking a side quietly and defending it afterwards.
One reading is a refusal, and size does not enter into it. A number replaces a possibility with a figure, and that is as true of ten people as of ten million. On this reading you would decline to count at any scale, and the reason never weakens.
The other is an incapacity. Statistics needs volume. Below roughly fifteen observations a figure is noise wearing the costume of a finding, and it is better not to know when you cannot know than to try and be mistaken. On this reading the objection dissolves once the numbers are large enough to mean something.
The two agree everywhere Koher currently works, which is why the disagreement went unnoticed for a long time. They diverge on one question: would you count at ten million? The first says no, still. The second says the question stops applying.
Nothing available can decide between them, and that is a consequence of the position rather than an accident inside it. A practice that has dismissed every external instrument has also dismissed the ones that could adjudicate this. So it stays open, in public, rather than being closed by preference and then argued for.
Held open. The instrument that could settle it is one of the instruments this position refuses.
What this is not
It is not anti-reach. Narrowcasting and broadcasting are both available and both get used. The setting is chosen for each thing made, after working out what that thing holds and whom it is for. Reaching four people is not a failure if it was made for four.
It is not modesty. Nothing here is a claim about deserving attention or not deserving it. It is a claim about what a figure does to the thinking that follows it.
It is not an instruction to anybody else. This is a discipline Koher holds for its own work. Other practices count, and are right to.
An unknown number can be anything. An unnamed gesture can be either thing. Both stay possible only while nobody has fixed them, and Koher would rather keep the possibility than have the figure.