As an Anthropologist, magical thinking is a normal fact of life. Rather than dismissing it outright, our job is to look at its function and yes, rationality, for groups at hand.

Starting out with a story about actual chickens being sacrificed to ensure the harvest, this talk explores the prevalence of useful magical thinking in our own community. Using metaphors, or even personifications, doesn‘t make a person irrational. It‘s applying a principle implicitly onto a subject matter which works completely differently, that would be the problem. After all, unless you are a strict vegetarian, it‘s not the killing of a chicken as such you‘d object to, it‘s the idea that this act makes rain.

With LLMs, our public sphere has run into a problem where experts are at loss explaining a very complicated thing to a general public, which often lacks the basic terms with which to understand how this mechanism works. The instant personification of LLMs can lead to vast mismatches between their actual capabilities and what those stories imply. Rather than dismissing them outright, the question posed would be, what‘s the alternative?

The talk is intended to be a light-hearted overview of some examples of both useful and dangerous constructions used to simplify complexity. It aims to touch upon some of the mechanisms that should be heeded in order to be able to tell a better story.