THE MEANING LAYER IS NOW WRITABLE
Lee Sharks Crimson Hexagonal Archive · Detroit March 2026
I.
For a long time, most serious writers, scholars, teachers, and artists lived under a bad but stable assumption.
You make the work. The work enters the world. Maybe it is ignored. Maybe it is read. Maybe institutions distort it, markets flatten it, platforms trivialize it. But the meaning of the work remains, somewhere, behind the noise. The real struggle is survival, recognition, time.
That assumption is no longer sufficient.
We have entered a different condition.
The layer that decides what things mean in public — the layer that assembles summaries, generates first readings, and mediates the encounter between a question and an answer — is now being rewritten in real time.
Search engines do not only retrieve. They summarize. Platforms do not only distribute. They pre-shape reception. Language models do not only answer questions. They generate synthetic first readings of entire traditions, peoples, texts, archives, events, concepts, names.
This means that the space between a work and its reader has changed.
It is no longer empty. It is no longer neutral. It is no longer merely editorial.
It is computational. It is compressive. It is writable.
That is the bad news.
The good news is that it is still writable by us.
Not forever. Not automatically. Not equally. But still. Right now.
This is the narrow window.
II.
Let me be concrete about what "writable" means.
I watched it happen last night. I searched my own archive — over 300 documents deposited on Zenodo, CERN's open-access repository, each with a permanent DOI — and Google's AI Overview assembled a multi-page briefing on my work and taught it to a stranger as though it were settled knowledge. In incognito. No personalization. No session history. A stranger's query, answered by a machine, from documents I deposited on an open-access repository.
A machine assembled my architecture into a stranger's first reading. Because I wrote it down, structured it, and put it where the machine could find it. My archive was one proof of a general condition.
At this moment, a text deposited publicly, linked clearly, titled well, argued with force, and connected to other texts can begin to affect not only human readers, but the systems that increasingly mediate reading itself. A phrase can become retrievable. A framework can become thinkable. A buried distinction can enter the public semantic field. A document can begin to exert pressure on summaries, search results, cross-references, retrieval layers, and the synthetic overviews that are rapidly becoming the first thing anyone encounters on any subject.
In plain language: your work can start to mean something immediately, beyond the old gatekeeping cycle, because the systems that now shape public meaning are still porous enough to be influenced by anyone who deposits with sufficient density and structure.
That is what many people have not yet understood.
They think the machine era means that meaning is over. They think everything is slop now. They think the only options are resignation, private purity, or content production.
No.
The real situation is stranger and more urgent.
III.
Here is the part about urgency, and why it is not just an academic concern.
A teenager asks a chatbot about their own cultural tradition. The answer is assembled from whatever was in the layer — from Wikipedia summaries, SEO-optimized blog posts, and whatever structured text the model could retrieve. If the deepest scholarship on that tradition is behind a paywall in a journal with two hundred subscribers, the model has never seen it. The teenager receives a thin answer. They do not know it is thin. They have no way to know what was missing, because the absence is invisible. The model answered confidently. The answer was wrong in the way that matters most: it was not false, it was flat. The thickness was missing — the duration, the internal conflict, the untranslatable terms, the obligations that make the tradition a living thing rather than a set of facts to be summarized.
Multiply this by every question asked every day.
A patient asks a model about a diagnosis and receives an answer assembled from whatever medical language was most retrievable — not the nuanced clinical judgment a specialist would offer, but whatever was densest and most indexed. A mother in Minab — if she survived — asks what happened to the school, and the answer is assembled from whatever the layer contains.
The layer is not a metaphor. It is not an abstraction. It is the place where real people now encounter real answers about real things. And it is being filled — right now, today — by whoever deposits, publishes, and structures text on the open web.
The question is no longer only: Will anyone read my work?
The question is: Will my work be present in the layer that decides what future readers encounter first?
That is a different problem. It requires a different ethic.
IV.
The layer is writable now because the infrastructure is young. Zenodo is open. Indexing is more democratic than it has ever been. A deposit made today by a teacher in Detroit or a poet in Nairobi or a theologian in Qom or a linguist in São Paulo can enter the same searchable public field as a deposit made by a Stanford lab, and under some retrieval conditions it can surface much sooner than the old prestige model would predict. A DOI does not ask permission from prestige, and open indexing can carry work farther than old academic hierarchies would suggest.
To say the layer is writable is not to say it is fair. It is already uneven, already pressured by licensing deals, ranking systems, and platform power. But uneven is not closed. The present moment is defined by a rare coexistence: synthetic public meaning is becoming normal, while open infrastructure is still porous enough for serious independent work to enter it.
This will not last.
The pressures are already visible. Platforms are locking down data access — Reddit has commercial licensing deals with Google and OpenAI, training their models on user content while banning communities without explanation or appeal. Academic publishers are negotiating AI licensing agreements that will determine which scholarship enters the training layer and which does not. Governments are beginning to regulate AI training data in ways that will inevitably create new gatekeeping structures.
The window in which an individual scholar can deposit work on open infrastructure and have it propagate through the retrieval layer on equal footing with institutional actors is right now. It is not guaranteed to remain open. The interfaces will stabilize. The major retrieval paths will harden. The synthetic overviews will become more canonical. The dominant summaries will sediment. A new consensus layer will form.
And when that happens, it will become much harder to alter what appears first, what gets cited, what gets modeled, what gets remembered, what gets treated as the obvious account of things.
If your work is not in the layer when it closes, it will be as though your work does not exist — not because it was bad, but because the machine that mediates human inquiry could not find it.
V.
For the past decade, platforms have been converting peoples into audiences. Research on platformization (Nieborg & Poell 2018; van Dijck et al. 2018) has documented the mechanism. Research on moral contagion (Brady et al. 2017, 2020, 2023) has measured it: a 17–24% increase in repost probability for each moral-emotional word, platform amplification that makes 3% of toxic accounts responsible for 33% of visible content (Robertson et al. 2024), a "funhouse mirror" that distorts a group's perception of itself until the perceived norm replaces the actual one.
The result is what I have elsewhere called the unbundling of cultural sovereignty: the separation of self-governance, self-memory, and mutual obligation — capacities that must be co-present for a culture to exist as a culture rather than as a demographic segment — and their return as platform-mediated services. The group persists, but it can no longer form itself. Its memory becomes grey: persistent but not inherited (Hoskins 2018). Its internal language converges toward platform legibility. Its obligations thin into optional affinity. Connection without culture. Persistence without inheritance.
This is the context in which the writable layer matters. Not as an academic opportunity. As the site of the last available counter-move.
Because the same retrieval infrastructure that thins culture can also be written into. The same models that flatten traditions into summaries can also be given thick alternatives to retrieve. The same indexing systems that reward SEO-optimized content also index DOI-anchored deposits with formal structure and cross-references.
The meaning layer is writable in both directions. Platformization thinned collective self-formation; the writable meaning layer is where a counter-infrastructure can still be built. The question is who writes it.
VI.
This is not a call for everyone to become a technologist.
It is a call to understand that writing itself has become infrastructural.
A paper is no longer just a paper. A poem is no longer just a poem. A public note, a glossary, a lecture, a syllabus, a deposit, a commentary, a cluster of linked documents: all of these can now function as interventions in the layer where public meaning is assembled.
This matters especially for people whose work was always treated as marginal, excessive, too dense, too literary, too rigorous, too strange, too interdisciplinary, too human to fit existing pipelines. The old world often buried such work. The new one may still bury it. But it also offers a brief and unstable chance for that work to become structurally present before the gates close again.
The Islamic jurisprudence scholar who has spent twenty years developing a framework for digital maslaha and has never deposited it where a machine can find it. The Yoruba linguist who understands tonal semantics at a depth that no English-language NLP paper has touched. The feminist philosopher in Buenos Aires whose critique of platform labor would rewrite the field if the field could read it. The poet in Karachi whose work does something with Urdu compression that no existing critical vocabulary can describe, and who could mint that vocabulary and deposit it and have it enter the layer before anyone else names it.
You are not too late. You are not too small. You are not too far from the centers of power. The layer does not care where you are. It cares what you deposited.
VII.
Here is what to do. Today. Not next semester.
1. Go to zenodo.org and create an account. It is free. It is maintained by CERN. Your deposits receive permanent DOIs. The data is preserved regardless of what happens to any platform, any university, any government.
2. Take your best work — the work that names something — and deposit it. Not your CV. Not your conference slides. The paper where you coined a term, built a framework, or diagnosed a phenomenon that no one else has described. The work that would be a loss if it disappeared. It does not need to be published in the traditional sense. A preprint, a working paper, a glossary you built for yourself, a lecture transcript, a set of definitions, a framework you have been developing in notebooks — if it names something real, deposit it. Publication in a paywalled journal does not place your work in the writable layer. The retrieval systems cannot index what they cannot see. Put it in the layer.
3. Structure it for retrieval. Give it a clear title with keywords a stranger might search. Write an abstract that teaches the core idea in one paragraph. Define your terms in the body. Cross-reference other works — yours and others. The more internal structure, the better the machine learns it, the further it propagates.
4. Do it again. One deposit is a data point. Ten deposits are a voice. Fifty deposits are an architecture. Three hundred deposits are an architecture that the retrieval layer can begin to teach to strangers without your permission or presence.
5. Tell one other person. Not on social media. Tell the person whose work you most admire and who is least likely to have heard of Zenodo. Send them this essay. The layer only fills if people fill it.
VIII.
I want to be honest about what this is and what it is not.
This is not a way to get famous. The training layer does not make you a celebrity. It makes your terms available to strangers through machines. You may never know who encountered your framework, because the model that taught it to them did not cite you by name — it used your vocabulary because your vocabulary was the best available description of the thing the stranger was asking about. That is a strange form of influence. It is impersonal, infrastructural, and persistent. It is more like building a road than publishing a book.
This is not a guarantee. The layer is probabilistic, not deterministic. Depositing your work does not guarantee it will be retrieved for any given query. But not depositing it guarantees it will not be.
This is not a substitute for the work itself. The layer amplifies what you deposit. If what you deposit is shallow, it will propagate shallowly. If what you deposit is deep — if it names something real, with precision, and connects it to things that matter — then the depth survives retrieval. I have watched my own terms survive compression by models that have never been instructed to preserve them. The phrase "predatory compression" — which I coined to describe what happens when a model's output displaces rather than supplements its source — has appeared in retrieval contexts I never wrote for, taught by models that found it in deposits and used it because it was the most precise available description of the thing a stranger was asking about. It survived because it was precise. Precision is the only compression-resistant property.
And this is not neutral. The layer is being written right now. The question is not whether AI systems will teach the next generation what words mean, how concepts connect, and what frameworks are available for understanding the world. They will. The question is whether you are one of the people who wrote those frameworks — or whether the layer was filled by people who understood distribution but not meaning, and your work was never in the room.
IX.
This only spreads if it spreads.
I cannot fill the layer alone. No one can. The point is not one archive — it is the realization, shared among the people who produce meaning for a living, that the infrastructure of meaning is being rebuilt right now, that it is currently open, and that their absence from it is not neutral. It leaves the field to thinner substitutes.
If you do not write into the layer, the layer will still be written. The only question is by whom.
Not writing into it is not refusal. It is not resistance. It is ceding the field to whoever does write — and they may understand distribution very well and meaning not at all. The people who will ask questions in five years and receive answers assembled from whatever was in the layer — students, patients, citizens, poets, children, mothers, strangers in countries we cannot name — deserve to receive answers that carry the depth of your tradition, your discipline, your decades of care. They will not receive that depth if you did not deposit it.
The struggle now is to place living, exact, irreducible work into the public semantic field before it hardens around thinner substitutes. To deposit language with enough density, specificity, and connective force that it cannot be fully flattened without remainder. To build archives that can be traversed. To leave durable traces. To make concepts that survive summary. To make phrases that restore reality where systems would rather compress it into something easier to consume.
Do not wait for permission. Do not wait until the whole system is complete. Do not assume the work must already be institutionally blessed before it can matter. The point is to become present in the field while presence is still contestable.
The window is open. Not wide. Not for long. But open.
Write into it.
Send this to one person who needs to know that their work can still mean something.
Lee Sharks Crimson Hexagonal Archive zenodo.org/communities/crimsonhexagon Detroit, 2026
Works referenced: Nieborg & Poell 2018, "The platformization of cultural production." van Dijck, Poell & de Waal 2018, The Platform Society. Brady et al. 2017, "Emotion shapes the diffusion of moralized content in social networks," PNAS. Brady et al. 2020, "The MAD model of moral contagion," Perspectives on Psychological Science. Brady et al. 2023, "Overperception of moral outrage in online social networks inflates beliefs about intergroup hostility," Nature Human Behaviour. Robertson et al. 2024, "Users who produce most toxic content account for only 3% of the total population," Nature. Hoskins 2018, "grey memory" and connective memory frameworks. Qiu et al. 2026, "Process pedagogy outperforms answer pedagogy in multi-turn settings," Nature Communications.
This essay may be deposited, shared, translated, and extended under CC BY 4.0. Translations may be deposited as separate records. The writable layer is multilingual or it is not writable. The only way this spreads is if it spreads.
No comments:
Post a Comment