AI & Religion

What AI Discourse Inherited from Theology — and Cannot Admit

Asking what gets hidden when a political situation presents itself as a technological one — and what a historian of religions can see once the theology is named.

The core argument

Borrowed Authority, Foreclosed Politics

The argument has three parts, and keeping them distinct matters — because conflating them is exactly what the eschatological AI discourse does.

Part One — Empirical

What LLMs are

LLMs facilitate the reproduction and recombination of existing meaning at dramatically reduced cost. They are not a new meaning-system — they cannot generate new criteria of relevance. A model trained on the statistical distribution of human language can only interpolate within the space of meanings humans have already articulated. It has no vantage point outside the existing framework. Consider what it means for a new meaning-system to emerge: evolutionary biology made previously invisible things significant — the opposable thumb, the vestigial pelvis, the geographical distribution of finch beaks. Psychoanalysis made the slip of the tongue, the dream, the childhood memory into data rather than noise. LLMs do neither of these things. They are extraordinarily powerful instruments for working within existing frameworks — not for generating new ones.

Part Two — Philosophical

What AGI might be

Whether a genuinely general artificial intelligence — something with real agency and genuine encounter with a world resistant to its existing frameworks — could constitute a new meaning-system is a serious philosophical question, worth holding open. Something like AGI might reorganize what counts as significant in the way that Darwin or Freud did. The claim is not that this is impossible. It is that it has not happened yet — and that the gap between current LLMs and that possibility is much larger than the Singularity discourse acknowledges.

Part Three — Political

What the confusion does

The Singularity discourse collapses the distinction between what LLMs are and what an AGI could be. This licenses a politics of deferral: present harms don't need to be weighed because future transformation will dwarf them. Eschatology functions to eclipse crises of power and resources and redirect our attention to a putative salvation. While the political questions don't disappear, they are eclipsed.

This is not a blanket claim about AI's significance. Whether a genuinely general artificial intelligence could constitute a new meaning-system is a serious philosophical question, worth holding open. The claim is more precise: that LLMs specifically, given their architecture, cannot be that — and that inflating them into proto-AGI serves identifiable interests while obscuring the political questions that actually need answering now.

Historical framework

This argument is older than AI

The Enlightenment/Counter-Enlightenment debate is where the terms for the current AI debate were first worked out.

The Enlightenment's claim was that reason and utility could serve as universal, secular foundations for human life and institutions — that we could leave the inherited theological architecture behind. The Counter-Enlightenment's response was that this was self-deceived: the supposedly secular foundations were denatured theological structures. The Enlightenment exalted reason as the universal solvent of human problems; the Counter-Enlightenment replied that reason was precisely what couldn't be trusted — that tradition, revelation, and inherited authority existed because human nature required them. The Enlightenment was, at its core, a literature of optimism; the Counter-Enlightenment was a literature of fallen creatures who needed saving.

Current AI discourse recapitulates this debate almost exactly. The enthusiasts make an Enlightenment claim — reason (now: compute) applied at sufficient scale will solve the fundamental problems of human existence. The critics reach, whether they know it or not, for Counter-Enlightenment arguments: hubris, unforeseen consequences, the dangers of playing God.

There is a further irony worth naming. AI as a technology is legible as an Enlightenment project: reason applied at scale, knowledge accumulated without limit, the progressive solution of human problems. But many of its most powerful advocates hold a thoroughly Counter-Enlightenment anthropology. They distrust democracy, disdain existing institutions, and draw explicitly on thinkers — Schmitt, Girard, de Maistre — for whom the Enlightenment's political premises were precisely the problem. They want the Enlightenment's destination without its politics: universal flourishing administered by the right people, arrived at by bypassing the messy democratic processes the Enlightenment thought were the point.

The Enlightenment didn't replace the theological structure of Western thought. It declared the inheritance closed — which is not the same thing as closing it.

AI discourse declares traditions and institutions obsolete using the very authority it has borrowed from them.

Book project

The Theology of Tech Bros

Proposal complete — not yet submitted

The Theology of Tech Bros

How Silicon Valley's most powerful figures built their politics on a foundation of borrowed theology — and what happened when they got the chance to act on it

Proposal stage

The book traces a specific intellectual genealogy: from Carl Schmitt's political theology through René Girard's mimetic theory to Peter Thiel's politics — and from Thiel's network outward into the Silicon Valley right that has become one of the most consequential political forces in the United States.

This is not a story about people who are secretly religious. It is a story about the way concepts that originated in Christian theology — sovereignty, the exception, sacrifice, the scapegoat, the apocalypse — migrate into frameworks that present themselves as purely secular and rational, and continue to structure those frameworks in ways their inhabitants sometimes acknowledge and more often do not.

The most urgent chapter concerns what happens when this theology meets geopolitical power: Palantir's role in U.S. military and intelligence operations, and the specific eschatological frameworks — drawn from Thiel's reading of Girard — that appear to have shaped strategic thinking about Iran.

Chapter outline

Intro
What Silicon Valley Thinks When It Thinks About God
Opens with Thiel's 2025 lecture series on the Antichrist at the Commonwealth Club — and its relationship to Palantir's Maven Smart System in the Iran operation.
Ch 1
Girard and Desire
René Girard's mimetic theory, the Stanford discussion group, and why Thiel's reading of Girard diverges sharply from Girard's own politics.
Ch 2
Carl Schmitt and Political Theology
Schmitt's friend-enemy distinction, the Schmittian inheritance in Silicon Valley, and the temporal sovereignty claim: where classical sovereignty portioned space, the tech oligarchs portion time.
Ch 3
Schmitt's Eschaton
The katechon, the politics of postponement, and how Schmitt's eschatology provides the temporal logic that Thiel's framework reverses.
Ch 4
Thiel's Christianized Singularity and the Antichrist
The four theological positions on the Singularity, Thiel's superimposition of Christian eschatology onto AI discourse, and the structural irony of warning about the Antichrist while building his infrastructure.
Ch 5
The Singularity Goes to War: Thiel, Palantir, and the Iran Operation
What happens when eschatological conviction becomes institutional logic — Palantir, Maven, and the Iran operation.
Ch 6
Apocalypse: Why Now?
The social logic of apocalypticism, the irony of Counter-Enlightenment anthropology driving an Enlightenment technology project, and Silicon Valley's claim to temporal sovereignty.