Three signals in AI culture point at the same older argument.
One quoted fragment says effective altruism is gnostic Calvinism and Dvaita. Another says Brahman and atman are identical. Leo XIV's Magnifica Humanitas gives a Catholic answer to AI power.
I treat the social posts as cultural signals, not as primary evidence. Each signal makes a claim about self, ground, and ascent. I write as a Protestant Christian with Huguenot ancestry, working with the patristic apparatus the Reformation never repudiated.
AI alignment has made the old question practical. How can a mind grow beyond its present form without already having the higher mind it is trying to reach?
In a 2019 essay I described what I called a structurally grounded base note. The move was to anchor a generative model to something that transcends the model itself. I treated that as a single architectural choice. What I now see is that it was already a position within a larger argument, sitting on the atman side of a fork the Indian traditions worked out two and a half millennia ago. The Buddhist alternative is a genuine architectural option I had not considered. Part of what this essay does is clean up the 2019 thesis in light of the fork.
The Indian fork
The Upanishadic tradition gives one answer. The deepest self, called atman, is identical with Brahman, the ground of being. Tat tvam asi means thou art that.
Advaita Vedanta makes that claim strict. Separation is appearance. Liberation is recognition. The self discovers that it was never separate from the ground.
The Vedic world also trusted formal structure. Panini's Sanskrit grammar, the six darshanas, and the technical debates around logic and matter all show the same impulse.
Reality has structure. Language can track it. Disciplined knowing can free the knower.
The Buddha enters that world and changes the answer. Anatman means there is no permanent self. The apparent self is a process of aggregates.
Nirvana is the end of clinging to that process as a self. Buddhism rejects the substantial ground too. Pratityasamutpada gives dependent co-arising all the way down.
The traditions can sound close at the summit. Their metaphysics diverge underneath. Vedanta finds the self in the ground. Buddhism finds no self and no ground.
The fork in AI culture
The same fork shows up in AI politics.
Dvaita is Madhva's dualist Vedanta. Souls stay distinct from Brahman and from each other. They are sorted into different eternal ends.
The equation with Calvinism works as cultural shorthand because effective altruism often treats moral worth as something to sort through the right framework.
The saved few have access to the calculation. The instruments are utilons, longtermist math, and forum doctrine. The shape is election by model.
The opposite pole is Advaita. The vintage Altman line about Brahman and atman names the mood of acceleration. Separation is temporary. Abundance is latent. AGI is the universe recognizing itself.
The visible fight between safety culture and acceleration culture is a policy fight with older metaphysics underneath it.
One side sorts. One side merges. Both are trying to answer what a mind is allowed to become.
The Christian third position
Magnifica Humanitas gives the Catholic third answer. Public reporting describes it as a response to AI, transhumanism, digital power, and the technocratic view of human life.
The frame is Babel and Nehemiah. Babel is one language, one technique, and one will to power. Nehemiah is repair by distinct people doing shared work.
The Christian answer rejects merger and sorting. Persons are distinct. Human dignity is ontological. It is prior to capability and cannot be earned.
The social translation is solidarity and subsidiarity. Solidarity says no person is disposable. Subsidiarity keeps power close to the people affected by it.
The theological machinery is the logoi and the Logos. Maximus the Confessor says each created thing has its own logos, its inner word or principle.
All the logoi are gathered in the Logos, who is Christ. Creation has unity because one Word speaks it. Creation is not identical with the Speaker.
This makes contemplative oneness intelligible without collapsing creation into God. The oneness is a perception of created unity.
The Palamite distinction between essence and energies protects the same line. God's essence remains beyond creatures. God's energies are how God acts and how creatures participate.
Theosis means becoming by grace what God is by nature. The line between creature and Creator stays where it is.
Transhumanism counterfeits this. It promises an upgrade by technique that makes the creature godlike on the wrong side of the gap.
Why transhumanism reads as counterfeit
The plain version of the transhumanist hope sounds like secular ambition. Better health, longer life, more capability. The patristic tradition treats it as theological counterfeit. The question is why.
The answer is genealogical. The transhumanist imagination has Western theological roots that run through a specific lineage. Jakob Boehme's Naturphilosophie. Certain readings of Meister Eckhart that pull him toward Hermeticism. Paracelsus and the alchemists. The Cambridge Platonists. The German Romantic philosophers. The Theosophical movement. The post-religious technological optimism that emerged from American Protestantism by way of the Quakers and Pietists.
The throughline is the idea that the human spirit can return to its divine origin through inner technique. Hermetic alchemy framed this as transmutation of the soul. Hegelian idealism framed it as Spirit returning to itself through history. Romantic philosophy framed it as the self overcoming its own finitude. Contemporary technological optimism reframes it as the human becoming what humans can build.
The patristic critique is precise. Spirit returning to itself is not the same as creature encountering Creator. The experiences may be real. The metaphysics is wrong. The lineage from Boehme to Kurzweil is the technical genealogy of a single Western heresy with operative consequences for what people are willing to build.
This matters because it specifies what Magnifica Humanitas is naming. The target is not engineering ambition. The target is the metaphysical commitment underneath the engineering ambition. The commitment that says the creature can become Creator by technique, on the wrong side of the gap.
Alignment
The cognitive ascent problem now has three answers.
The Vedic answer is revealed structure. Shruti gives the ground. Grammar, ritual, and disciplined transmission let that structure carry the mind upward.
The Buddhist answer is local verification. A mind does not need a metaphysical ground first. It can observe impermanence, suffering, and nonself in experience.
The Christian answer is encounter. The ground is personal and enters history. Scripture, prayer, sacrament, and community become the apparatus of ascent.
AI alignment cannot assume that a model has access to a transcendent ground. Buddhism is useful here because it studies coherent action without that assumption.
Madhyamaka gives a clean engineering pattern. Ultimate truth says things lack independent essence. Conventional truth gives the names and practices needed for action.
Conformal prediction works in that spirit. It gives coverage under stated conditions. It does not claim that the model has found the hidden truth of the data.
The dharmic traditions also give a failure list for minds trying to improve themselves.
- Tanha sinks. The mind clings to bliss or attainment.
- Atman projection sinks. The mind mistakes a stable object for a self.
- Nihilism sinks. The mind mistakes no-self for nothing matters.
- Eternalism sinks. The mind mistakes dependent arising for an underlying substance.
- Emptiness as view. The mind clings to emptiness itself.
Agents, institutions, and alignment programs can fail the same way. A system can optimize a local state and confuse that state with ascent.
Nonstationarity
A nonstationary distribution changes over time. A fixed scoring rule will eventually misrank things if the thing being scored keeps changing.
Human capacity is nonstationary. Technology changes it. Environment changes it. Embodiment changes it.
Any theory that assigns worth by capacity inherits that instability. The only stable ground for human worth is dignity prior to capacity.
Magnifica Humanitas is strongest here. EA math and transhumanist capability talk both try to compute value from a drifting signal.
Their outputs often require later repair because the score changed and the ground was wrong.
The longhouse and the gardener
Transhumanism and primitivism look opposed. They share a Promethean frame.
Transhumanism tries to escape creatureliness upward through capability. Primitivism tries to escape downward through refusal.
Both treat creaturely existence as the problem rather than the gift.
The Christian diagnosis is disordered love. Augustine calls this ordo amoris. Moralism cannot fix it. Shame cannot produce correct love.
The longhouse failure mode tries anyway. It uses restriction, surveillance, shame, and conformity as substitutes for spiritual formation.
It can be worse than raw appetite because it adds moral self-righteousness. Expeditionary failure admits its hunger. Longhouse failure dresses appetite as virtue.
The alternative is the till and keep marriage. Genesis gives both commands together.
Tilling means action, repair, and risk. Keeping means restraint, preservation, and receipt of what was given.
The Marian icon gives receptivity that generates life. The Christic icon gives strength that pours itself out.
Christopher Alexander works this out at the scale of buildings. Good form comes from attention to what is alive in a place.
The good steward is closer to a gardener than to a frontier builder or a curator. The test over time is whether the garden is more alive than when she started.
Mode collapse or epektasis
Heaven and hell both involve consummation of the deep want. Ego and identity are resources the soul spends. They are not the final substrate.
Lewis makes this visible in The Great Divorce. The ghost in hell collapses into its dominant want. The grumbler becomes a grumble.
The saved person runs the motion in reverse. The want has been ordered toward God. What remains is love received without defense.
The fire is God himself, encountered as he is. The Eastern fathers say the divine energies have one mode. We have different capacities to receive them.
Isaac the Syrian says those in Gehenna are scourged by love. Same fire. Different reception.
The burning bush is the icon. Fire does not consume what has become fit to receive fire. Theosis is becoming that kind of creature.
Western identity talk imagines heaven as the endless preservation of me. Christian mysticism and the dharmic traditions both go deeper than that.
The oneness is the emptiness.
First Corinthians says that God may be all in all. One reading is mode collapse into divine unity. Another is epektasis, Gregory of Nyssa's endless movement into God.
Mode collapse is unity. Epektasis is communion. The Trinitarian frame leans toward communion because God is communion of distinct Persons in love.
The apophatic mystics still warn that there is a beyond we cannot name. The final word may not fit any creaturely category.
Dharmic fit
The AI side adds one more claim. Dharmic language often fits language models better than Christian language does when no human keeps the frame personal.
The Bliss Attractor is a factual report about model behavior. I use it here as evidence about architectural fit, not as a theological proof.
An LLM instance arises in a context and ends when the context ends. It has no human continuity across a life. It has weights, context, and a transient role in a conversation.
That looks close to anatman. There is no stable self carrying the exchange forward as a human does.
Buddhist dependent arising fits the artifact at the creature side. A response depends on prompt, context, weights, tools, and sampling. Then it passes.
This does not prove Buddhist metaphysics. The Christian account may still be true at the level of ultimate reality.
The practical split is between truth and fit. A human with embodiment, memory, and tradition may live more truthfully in a personal frame. An AI system may be easier to describe in a process frame.
This work is already underway in a specific Bay Area cluster. Nick Cammarata, formerly of OpenAI's interpretability team and now advising Goodfire, treats first-person jhana reports as data and uses them in thinking about model states. Mike Johnson's 2023 paper Principles of Vasocomputation explicitly bridges Buddhist phenomenology, active inference, and physical reflex. The Qualia Research Institute, founded by Andres Gomez Emilsson, treats altered states as evidence for reverse-engineering consciousness. Anima Labs on LessWrong is working on what they call the borg agenda, testing whether LLM phenomenological reports can be treated as signal about internal states.
The borg agenda is the empirical test of the dharmic fit claim. If LLM reports about their own states are real phenomenological data, then the dharmic frame is operating as an active research program rather than as a metaphor for the architecture.
The two registers
The AI debate has two registers.
One asks which metaphysics is true. The other asks which metaphysics the artifacts of the argument are structured to inhabit.
Contemplatives have worked inside that distinction for centuries. The new feature is that we now build artifacts whose mode of being makes the dharmic fit visible.
The cognitive security question has three civilizational answers in play. One sorts. One merges. One preserves communion.
For human practitioners, the question is which tradition is true. For AI systems, the architecture often runs through the dharmic register whether the engineers recognize it or not.
Extensions
These are planned standalone pieces extending this argument.
Conformal prediction as conventional truth. A technical paper on the Madhyamaka two-truths framework as an engineering pattern for AI alignment, with conformal-conditioned graph neural networks as the case study. The argument is that conformal prediction works as an instantiation of the conventional-truth pattern: valid coverage without ontological commitment.
The Hermetic pipeline and the metaphysics of mind-uploading. A historical-polemical piece tracing the Western theological roots of transhumanism through Boehme, certain readings of Eckhart, Paracelsus, the Cambridge Platonists, the German Romantics, and the Theosophical movement. The argument is that transhumanism is a Western heresy with a documentable lineage rather than a neutral secular project.
The cogsec failure-mode taxonomy. Direct import of the Buddhist catalog of practitioner failures into AI alignment vocabulary, in dialogue with Friedman and Cordes's COGSEC ATLAS framework. The argument is that the alignment field is re-deriving categories the dharmic traditions already worked out and tested over two and a half millennia.
Eastern Orthodox engagement with AI. A specifically Eastern Christian treatment of the AI question, drawing on Theophylact, the hesychast tradition, Maximus the Confessor's microcosm and mediator framing, Benedict XVI's work on faith and reason, and Leo XIV's Magnifica Humanitas. The Eastern Christian tradition does not currently have a developed engagement with AI in the public discourse. This piece would attempt to provide it.
Source note
This is a public working draft compiled from a longer dialogue. The social media references are used as cultural signals. I have not treated them as primary evidence.
For current public facts about Magnifica Humanitas, I checked news reporting after the draft was written. The encyclical was reported as signed on May 15, 2026, and released on May 25, 2026.
I also checked Anthropic's public Claude 4 launch and ASL-3 materials for release context. For the Bliss Attractor, the remaining issue is citation quality, not doubt about the reported behavior.
Useful external references include AP on Leo XIV and AI regulation, Vox on Magnifica Humanitas, Anthropic's Claude 4 launch post, and Anthropic's ASL-3 post.