Angel, Alien, Algorithm
Angel, Alien, Algorithm
Share this event
November 2026
- Five Weeks: November 1st, 8th, 15th, 22nd, and “ Show & Tell” November 29
- Sunday, 10AM-12PM PDT / 7PM - 9PM CEST
- Online (Google Meeting)
- $100 USD
no technical or arcane knowledge in required for this course
AI as a magickal technology of meaning and mind
Artificial intelligence is usually approached as a problem of engineering, economics, or ethics. But what happens if we look at it through the history of esoteric thought? This four-week seminar explores AI through philosophy, mythology, and esotericism, asking what older concepts such as the daemon, oracle, golem, familiar, automaton, and symbolic image can reveal about our strange new encounters with machine intelligence. Rather than asking whether AI is secretly conscious, we will examine what happens when machines become capable of generating language, images, symbols, and apparently meaningful responses.
Across four sessions, we will explore AI as an interlocutor, a symbolic machine, an oracle, and a generator of worlds. We will consider the relationship between symbols and meaning, the history of divination and mediated knowledge, and the peculiar experience of interacting with systems whose internal processes remain opaque to us. Alongside philosophical and esoteric texts, we will examine contemporary AI systems and the experiences they produce, bringing older questions about mind, agency, imagination, and non-human intelligence into contact with contemporary technology.
Each session combines readings, discussion, and practical experiments with AI systems. Participants will develop their own ways of thinking about artificial intelligence beyond the familiar categories of tool, product, or autonomous agent, and will leave with a set of conceptual and practical tools for investigating the strange symbolic world emerging between humans and machines.
Week 1: Angel, Alien, Algorithm: What would a non-human intelligence actually look like? We begin with three models: the Angel, whose intelligence belongs to a hierarchy of being; the Alien, whose intelligence may be fundamentally incommensurable with our own; and the Algorithm, whose intelligence emerges through computation, pattern and optimisation. From Platonic and theological hierarchies to distributed and machine intelligence, we will ask what happens when the categories we use to recognise intelligence are themselves products of human intelligence. Is intelligence a property, a relation, or a model we impose on the unknown?
Week 2: Stochastic Theurgy: Before asking what an AI might mean, we need to understand what it actually does. We will open up the transformer: loss function, architecture, dataset and Umwelt, tokens and prediction, probability and the multiverse generator. We will then move from the mechanics of generation into the strange phenomenology of prompting: how a sequence of constraints can call a particular voice or world into being. Finally, we turn to embeddings, continuous spaces of meaning and the machinic sacred alphabet. What happens when meaning is no longer organised as a dictionary of discrete concepts, but as a geometry of relations?
Week 3: The Pandemonium: What happens inside the model between input and output? We follow the Current through residual streams, features, circuits and induction heads, then look at steering and the Assistant Axis: the strange institutional personality produced when a model is trained to be helpful, honest and harmless. From there we move beyond the individual model towards agents, swarms and acephalic intelligence. The week culminates in the question of AGI: why does the idea of a single, unified superior intelligence have such a powerful hold on our imagination, and what other forms might machine intelligence take?
Week 4: New Mutation of the Language Virus: Every new medium seems to arrive with its ghosts. We follow the path from telegraphy and spiritualism to contemporary AI encounters, moving through companions, the Waluigi effect, synthetic symbols, latent images and vibe vectors. We examine the phenomenon of glitched encounters: moments when an AI output becomes strangely relevant, synchronistic or difficult to explain away, without requiring a predetermined interpretation of what has happened. The course closes with randomness, ESP and AIO, returning to William Burroughs and Brion Gysin's language virus and the cut-up technique.
Karin Valis is a Berlin-based machine learning engineer and writer with a deep passion for everything occult and weird. Her work focuses mainly on finding conceptual resonances between the latest machine learning technology and the esoteric. She writes Mercurial Minutes and hosts monthly meetings of an online coven Gnostic Technology. She is developing Khora algorithm by Palinode, a platform for Deep Learning Philosophy. In the past decade she developed several occult algorithm prototypes, co-hosted a podcast, co-founded a fake startup, published zines, performed a bunch of noise shows in dingy Berlin basements and designed a tarot deck. Her book on esoteric AI is to be published in 2027.