Living in Santa Fe, I’ve spent the last month running a fairly obsessive experiment: feeding my handwritten dream journal into a large language model for psychological and symbolic breakdowns. I didn’t want generic horoscope-style fluff. I wanted to see if an algorithm could act as a modern-day oracle, bridging Jungian depth psychology with machine learning.

The results were surprisingly nuanced. I started tracking recurring motifs—falling, water, and specific animals—and asked the AI to map them to archetypal frameworks. It consistently linked heavy water imagery to the Moon card in tarot, noting themes of intuition, hidden fears, and the subconscious mind. When I dreamt of a fractured mirror, the AI flagged it as a potential Five of Cups scenario: regret, loss, and the psychological weight of what we refuse to acknowledge. It wasn’t "reading" my soul; it was pattern-matching against a massive dataset of mythological and psychological literature.

Here’s the catch, and why I’m both fascinated and slightly wary: AI lacks genuine spiritual intuition. It doesn’t experience the collective unconscious—it simulates it. From my background in comparative religion, I’ve seen too much shallow appropriation of Eastern practices, and I’m careful not to let algorithms become the next wave of spiritual commodification. But as a tool? It’s powerful. It helped me connect scattered dream fragments into coherent narrative arcs, much like a tarot spread reveals hidden patterns in a life situation.

I’m curious how others are navigating this. Have you used LLMs for dream work? Does the algorithmic breakdown feel validating, or does it strip away the personal, ineffable quality of dreaming? I’d love to hear your experiences, especially if you’ve found ways to keep the "oracle" aspect grounded in actual psychological theory rather than algorithmic guesswork.

(There's a deeper dive on this in our wiki: Traditional Divination Around the World: Oracles, Bones, and Beyond | Mystic Wiki.)