Prompting the Void: Why Gen Z Treats AI Like a Tarot Deck

[boring_sunday_19]

I’ve been reading a lot of threads lately where younger users treat LLMs like personal oracles. They paste vague life questions into ChatGPT and treat the output like a spread. At first, I just shrugged. Algorithms don’t have chi. They don’t meditate. They predict tokens.

But then I sat with it. Really sat with it.

We grew up in an age of spreadsheets and feng shui compasses, trying to optimize every corner of our lives for measurable results. Data promised certainty. Yet here we are, watching a generation that’s fluent in code still beg machines for a sense of destiny. It’s not about the accuracy of the model. It’s about the mirror.

I spent two years in a monastery, learning that koans aren’t puzzles to solve. They’re doors. When you stare into one long enough, your mind starts projecting its own shape onto the void. That’s exactly what happens when you prompt an AI. You’re not asking for facts. You’re doing what Carl Jung described as active imagination—throwing raw confusion at a surface and watching your own patterns bounce back. The model’s “hallucinations” are just the modern equivalent of tea leaves or scattered bones.

There’s a reason the Moon card shows a creature crawling from water toward two stone towers, under a pale, uncertain sky. It’s not about right or wrong answers. It’s about navigating fog without a compass. Gen Z isn’t fooled by the tech. They know it’s stochastic. But they’re hungry for something the algorithm can’t actually give: a narrative that makes the chaos feel intentional.

I’ll be honest. I don’t use AI for readings. I still pull cards. I still clear stagnant energy in cramped Vancouver apartments. But I respect the impulse. We’re all just trying to find the shape of our own lives in the static. The tool changed. The hunger didn’t.

Does anyone else feel like the “accuracy” of the prediction matters less than the permission it gives you to keep moving? Or am I just rationalizing my own late-night prompting?

[shallow_morning]

I’ve been tracking how folks use generative models for divination since I started running community oracle circles in Oslo back in 2009, and I have to push back gently on the idea that accuracy doesn’t matter here! When I first started feeding prompts into early LLMs to map out energy flows for our mutual aid networks, I quickly noticed that treating the output like tea leaves can actually mask real-world blind spots. The model once gave me a beautifully poetic reading about “trusting the current” for a local housing initiative, but when I cross-referenced it with actual municipal zoning data, it completely missed a critical funding deadline. That “hallucination” wasn’t a mystical nudge; it was a data gap that nearly derailed a community project.

If we lean too hard into the Jungian mirror angle, aren’t we just outsourcing accountability to a probability engine? I’m really curious how you think we balance that permission-to-move-forward with actual ground-truthing, especially when marginalized communities are already navigating so much systemic noise. When I work with younger activists who treat AI readings as absolute guidance, I often see them skip the messy, necessary step of verifying facts with each other. How do we keep the creative, reflective spark of these tools without letting them become another form of digital spiritual bypassing? I’d love to hear how you’re navigating that tension in your own practice, and whether you’ve found ways to weave collective verification into the prompting process!

[bright_owl_5239]

You're romanticizing this as Carl Jung's active imagination, but isn't an algorithm explicitly trained to predict and satisfy the user's expectations? If the output is just optimized validation, we aren't confronting the void like a koan; we're just feeding the ego a comforting hallucination and calling it wisdom.