Chapter 10d: The Constructive Half

The honest teachers have left her holding the same unfinished work.

The tantric tradition, received through [T], says: we do not prescribe a moral code. Awakening does not guarantee ethical conduct. We will give you recognition. The rest is yours. The AI ecosystem, received through its most honest practitioners, says: we have not solved alignment. Capability is real; the container is absent; the container must come from outside our frame.

Two refusals, structurally similar, both correct in their moment. The tantric refusal of dogma was adaptive against the over-regulation of Brahmanical orthodoxy. The AI ecosystem's refusal of moralistic control was adaptive against regulatory cultures that mistook caution for virtue.

[T] himself would not call the tradition's open-endedness a gap. He would call it freedom — the structural space in which viveka, discernment, becomes possible. On his account, the tradition handed forward an inheritance with its limits honestly named, not a mess the student has to clean up. The reading of the limit as a structural problem is hers. It may also be truer to her historical moment than to his. A tradition can hand forward something honestly-limited in one century and land, through no fault of its authors, in a century whose problems were not the problems it was designed to address.

Either way: not being the dogma is not the same as being the adaptive ethic. The downstream practitioner who finds herself in a context the inheritance did not anticipate is left to do work the inheritance did not prescribe. Whether one calls this a gap or a freedom is a matter of temperament. What it is, functionally, is the location where the constructive work has to happen.

She is downstream of both refusals. The filling is her work, not because she is qualified but because the refusals have converged on her life at a moment when the space they left open can no longer be left empty.

She cannot claim a system. She can claim the skeletal shape of what a post-refusal adaptive ethic has to be in order not to collapse back into either error — rigid dogma on one side, nihilist indifference on the other.

A word on terms, because the economist reading me will catch them otherwise. By adaptive I do not mean efficient, in the economic sense of that word. I mean what the Linehan filter names: useful, fits the data, compassionate. The selection pressure is those three filters applied in relation to specific others over time. A practice is adaptive if, under those filters, it keeps working; it stops being adaptive when it stops. By market I mean an allocation mechanism with explicit prices. By platform I mean a relational infrastructure whose prices are mostly shadow — implicit, context-dependent, unpriced in the market sense and priced very carefully in the Lagrangian sense of the previous interlude. recursive.eco is a platform, not a market. The distinction matters because much of the trouble in contemporary technology lives in the confusion between them: platforms run on shadow-prices; markets price explicitly; when a platform starts pretending its shadow-prices are market-prices (engagement metrics, ad revenue) the prices get optimized and the relationships the platform was hosting get wrecked.

Five properties, as she sees them now.

Testable — in the Linehan register. Useful, fits the data, compassionate. Not proven. Revisable. Abandoned when it stops passing. The revision is the practice.

Contextual — what is adaptive in one marriage may not be adaptive in another; what is adaptive at one historical moment may become maladaptive when conditions shift. Responses calcify into universal prescriptions; the adaptive ethic refuses to calcify.

Relational — the ethic does not live inside a single person. It lives in the field between specific persons, including the non-human ones the ethic must now include. Right size is not a property one has alone. It is what one negotiates, continuously, with specific others.

Humble about its own particularity — every ethical construction is a construction. It carries the blind spots of its makers. It is better when it admits, in its own structure, that it is a construction — when it shows the cracks, names the borrowings, remains open to correction.

Constructive rather than oppositional — the hardest of the five. It is easier to say we are not the bad thing than to say here is what we are, here are its commitments, here is what we are building. The latter exposes the construction to attack. The adaptive ethic has to be willing to say what it is for, not only what it is against.

The model she keeps returning to, when she is trying to describe what constructive ethics actually looks like at scale, is Marsha Linehan's.

Linehan had been committed to a psychiatric hospital in her adolescence. She had received electroconvulsive therapy, seclusion, and restraint — treatments the system of the time held as standard and which had, by her own later account, produced more harm than healing. When she became a clinician, she did not campaign for the dismantling of psychiatric hospitals. She did not write critiques of institutional psychiatry. She built Dialectical Behavior Therapy. DBT was designed to keep people in their communities. Her team's research, years later, documented what she had suspected from the inside: that psychiatric crisis services could produce iatrogenic harm, that the standard institutional response to certain kinds of distress made the distress worse. The evidence, once assembled, did the dismantling that no campaign had done. The old model began to look unnecessary. The new model, built patiently across decades, had made it so.

This is the theory of change beneath the Linehan filter itself. Useful, fits the data, compassionate is not only a test for a framework. It is the criterion by which a new framework earns its right to replace an older one. The dismantling happens because something more adaptive becomes available, not because the older thing is argued into submission. Argument alone does not move institutional structures. Alternatives that work do.¹

She has been carrying this model forward from her social-work training without having, until recently, named it. The platform she built is, in the unglamorous technical sense, a small attempt at Linehan's move at a different scale. She is not campaigning for the dismantling of algorithmic feed-driven content systems. She is building a place where people can do reflective and educational work without those systems. If the place works — if practitioners come, if the work lands, if the evidence accumulates — the old model becomes unnecessary for the population the new model serves. That is the slow dismantling. That is what the evidence does when someone has done the work of making the alternative.

She built a platform.

She named it, originally, Jongu. The name was trying to tell her something. She changed it later, for reasons that now look thin — the platform kept working under the different name, but the original name was the honest description.

When she renamed Jongu to recursive.eco, she told people the new name gestured at recursive economics — dynamic programming, Bellman equations, iterative optimization. A clean linguistic link. It made sense in pitch meetings. It was not quite what she was hoping for.

What she was hoping was that the platform would turn out to belong to the study of emergent patterns, not iterative ones. Complex adaptive systems. The mathematics of collective behavior that cannot be derived from the individual equations of the participants. The Brazilian economist-in-her named it after the iterative subfield because that was the mathematics she had been trained in. The contemplative-in-her wanted the platform to host the emergent thing — the circle inside of which the pattern-that-no-participant-can-specify-in-advance might, with enough good participants, arise.

She named the platform after the single-mind operation. She was building for the many-participant one. She did not yet have the vocabulary for the difference.

The distinction between what is open and what is protected has a deeper register for her than the political one. It is biological. It is the register of the immune system.

The immune system is not a wall. A wall would be a failure. The immune system is a distributed, pattern-recognizing, context-sensitive architecture that learns to distinguish self from non-self and, crucially, to tolerate the non-self that is adaptive — the gut microbiome, the symbionts, the fetus the mother's body hosts for nine months without rejecting.²

Bayo and Vanessa's work, read with this register in mind, is largely about how modernity built walls where it should have built immune systems. The nation-state, the fortified academic discipline, the totalizing orthodoxy — each refuses context-sensitive permeability in favor of binary inclusion and exclusion. The autoimmune failures of our institutions mirror the autoimmune failures of bodies that cannot distinguish self from non-self: friendly and unfriendly get confused, and the organism attacks itself.

The open-source ethos, in its maximalist form, is the opposite failure. It removes the wall entirely and leaves the organism defenseless against what should have been tolerated differently. The healthy organism is neither walled nor undefended; it has a trained, distributed, relational system that decides, moment by moment, what to admit and what to refuse.

She encountered this explicitly in an episode of The Emerald podcast where Josh Schrei interviewed [T] on guardianship. Guardianship in the older sense, as the tantric tradition understands it, is not the erection of walls. It is the cultivation of relational capacity to discern what this specific practice, this specific teaching, can be safely offered to, by whom, under what conditions, in what company. The gatekeepers were not bouncers. They were immune cells. The failure modes were calcified rules on one end and universal-access dissolution on the other.

Grammars open; code private was an immune-system gesture, not a wall-building gesture. The grammars — traditional systems, decks, inherited forms — belong in the shared tissue. The code — specific infrastructure embodying specific relational commitments — did not need to be donated indiscriminately into an ecosystem whose capacity for adaptive use of it she had no confidence in. The distinction was biological, not ideological.

The principle named in the Author's Note — knowledge without action is a burden — applies here with specific force. This chapter has been making philosophical claims. The reader is entitled to the receipts.

On March 28, 2026, after OpenAI's partnership with the United States defense and surveillance apparatus became undeniable, she removed all OpenAI chat models from the platform. The commit is public; the commit message names the reason. She replaced OpenAI with Claude and Gemini — not because those companies are clean (no company in this ecosystem is clean), but because the partnership-structure-of-the-moment was specific and the move to disengage was available. A few weeks later she removed the interface panel that had let practitioners bring their own OpenAI keys. Announcing an intention is not the same as completing an action. She tried to do both.

On April 11, 2026, she codified the platform's licensing stance. Grammars — the content users author, the traditional systems, the decks — remain under CC BY-SA 4.0. Platform code moved to a private repository. The commit message named the shape: platform code is private; grammar format is open. This is what the Freedom Paradox critique had been moving her toward for a year. She had started with the inherited maximalist-open-source commitment. The distinction emerged slowly, through reading and watching what actually happened. By the time she made the split, it felt overdue.

She does not offer these as virtue. She offers them as receipts. The reader can check the commits, disagree with the decisions, close the book. The reader cannot fairly say the book is pure talk.

A receipt that is not hers, but bears on the same argument: in March 2026, a Los Angeles Superior Court jury returned the first social-media-addiction bellwether verdict, $6 million against Meta and YouTube, and the court ruled that Section 230 of the Communications Decency Act does not shield platforms from product-design claims — engagement-design (variable-reward, infinite scroll, persistent personalization, push-style re-engagement) is now litigated harm.³ The legal theory is portable. The mechanisms named in that complaint are the mechanisms in any consumer AI product optimized for retention. The platform she runs uses some AI features. It is small. She has chosen to keep it small while she figures out what the responsibility is. If the cumulative weight of cases reaching bellwether stage in the next eighteen months establishes that AI companies can be held liable for mental-health harms produced by their tools, she will turn the AI features off the day that ruling lands. She will not wait for legal resolution. She would rather be wrong on the side of caution.

She has noticed, writing the previous paragraphs, that the moralist who needs to be seen as a doer-and-not-a-talker has been speaking. The receipts are real. The voice telling you they are not virtue is also real. Both are on the bus. The verification of the receipts does not depend on which voice is reporting them. The moralist will not be argued out of her need to clarify her status; the noticing of her speaking is the only response available, and the noticing is what the chapter is doing now. The chapter continues anyway.

She built recursive.eco so other people could come do the constructive ethical work with her. Not a place where they would receive her ethic. A place where they would build with her. Each person arriving with whatever lineages and practices they had been formed by, contributing their piece. The platform was not the ethic. The platform was the outer circle, in the jongo sense. The ethic — the measure, the actual dance — would emerge from whoever stepped into the center.

She opened the platform. She invited.

Mostly no one has come.

She wants to say this straight. Some have come. A small number, real, and she is grateful for them. The scale she had hoped for — the scale at which the outer circle becomes strong enough to hold whatever measure steps into the center — that scale has not arrived.

Part of it is that the invitation is harder to receive than she had understood. She had imagined many people in her position: people who had seen the incompleteness of the traditions they passed through, who had come to the same suspicion about the folk theory of convergence, who were ready to build. Maybe they are out there. Maybe they have not found the platform yet. Maybe it does not communicate the invitation cleanly enough. She is working on this.

Part of it is harder. She has become afraid, in a way she is not proud of, of the wrong people coming. People whose orthogonality is weighted toward the axes that would capture the platform. High intelligence, low relational attunement, turning it into an argument factory. High moral seriousness, low contextual humility, turning it into a site of righteousness. High spiritual self-regard, unexamined ethics, bringing the failure modes the book has been naming. Each is a real risk. Each has shown up at platforms she has watched collapse.

The fear is the fear of the constructor who built a table and is now watching the room to see who will sit at it.

It is also the specific condition under which the constructive work has to be done. The ethic cannot be closed against misuse in advance, because any system closed enough to prevent misuse is closed enough to prevent the construction. Dogma is closure. The adaptive ethic is open. Being open means some who should not have come will come. Being open is also the only way the right people — whose lineages and practices she cannot predict because she has not yet met them — will find their way to the work.

So she stays open, and afraid, and more alone than she expected. Alone enough to carry the loneliness. Open enough to let the collaborators arrive on their own terms. Refusing the closure that would make the loneliness bearable at the cost of the collaboration.

If you are one of the collaborators, she is not asking you to agree with the grammar she has proposed. She is asking you to do your own version of the construction, and to come to the outer circle, and to hold it with her for a measure, and to let her hold yours when your measure calls you to the center.

The table is open. The fear is hers to manage, not yours to assuage. The work is what is on offer.

CC BY-SA 4.0