Ghosts in the Machine and The Dark Psychology of AI Simulated Societies

Sometime in the late 1930s, a man in rural Alabama met another man in a cemetery at night. The second man waved a bottle in his face, told him he had been hexed, and told him there was nothing left to do but go home and wait. He went home. He stopped eating. Weeks later he was near death in a hospital bed with a chart that said nothing was wrong with him. What his physician did next involved a syringe, an elaborate lie, and a lizard.

In roughly 945, Prince Igor of Kyiv went back for a second round of tribute from the Drevlians and paid for it with his life. What his widow did next has been retold for a thousand years: four escalating acts of revenge that ended with a fortified city burning from the inside out, and a rival tribe erased as an independent political force. This episode traces the story of Olga of Kyiv from the death of her husband through the siege of Iskorosten, then follows the stranger part: how the woman responsible for it all was eventually canonized as a saint, venerated in both the Eastern Orthodox and Roman Catholic churches under the title "Equal to the Apostles," and named patron saint of widows and converts. Along the way: what social identity theory explains about why an audience roots for her, what honor culture research says about societies where retaliation functions as maintenance rather than excess, and why the concept of double deviance explains the particular shape her legend took. Olga stands as one of the earliest examples of the female antihero in Western storytelling, and the way her chroniclers handled that says as much about them as it does about her. The episode closes on the Primary Chronicle itself, compiled by monks in Kyiv a century after Olga's death, under the rule of her Christianized descendants, with every reason in the world to make the story land exactly the way it does. Where the history ends and the myth begins is not a question that has a clean answer. Hosted on Acast. See acast.com/privacy for more information.

In 2009, a team of psychologists published the results of a haunted house they built on purpose. Ghost sightings, sensed presences, cold spots, and unexplained dread had already been linked to infrasound and electromagnetic fields, so Chris French at Goldsmiths and architect Usman Haque constructed a featureless circular chamber inside a north London flat and ran seventy-nine people through it to find out whether a haunting could be manufactured from physics alone. It could not. Whether the machinery was running or switched off made no difference to what people experienced inside that room. What did predict a person's report was the person, and the single most powerful variable in the entire study turned out to be the ethics paperwork they had to sign before walking in. This episode starts with Vic Tandy, the engineer who found his laboratory ghost hiding inside an extractor fan, and traces the theory forward through Michael Persinger's God Helmet, the Swedish replication that undercut it, and the argument that never resolved before he died. Then it goes into what happens in a body left alone in a dim, curved, silent room for fifty minutes after somebody has told it that something might happen. Predictive perception, the misattribution of physical arousal, and why some people meet something in that chamber while the person after them gets bored. Eight percent of participants reported terror. Some of them were in the group where the room was doing nothing at all. The uncomfortable question this episode holds open is whether the traffic in a haunted place runs the direction everybody assumes it does. Wiseman's work at Hampton Court and the Edinburgh vaults suggests the buildings really are doing something. What that something is stays unsettled. Until next time, stay curious and stay strange. Hosted on Acast. See acast.com/privacy for more information.

DMT entities, machine elves, and the hyperspace breakthrough: what people report meeting under N,N-dimethyltryptamine, and why the reports match. A Johns Hopkins survey of 2,561 people found that most described a conscious, intelligent being, and that more than half of the atheists in the sample stopped identifying that way afterward. This episode covers Rick Strassman's University of New Mexico research, Terence McKenna's machine elves, the Hyperspace Lexicon, and the listeners who get locked out of hyperspace entirely.Show notes: A Hungarian chemist injected himself with a compound he had synthesized in 1956, and the first thing he wrote down was that there were creatures there. Thirty-five years later a psychiatrist at the University of New Mexico ran the first legal psychedelic study in the United States in two decades, and half his session notes involved his volunteers meeting somebody. In 2020, Johns Hopkins surveyed 2,561 people who had encountered an autonomous entity under DMT and found that most of them believed the being was conscious, intelligent, and still out there after they left.This episode is about the machine elves, the insectoids, the half-animal figures, and the vocabulary a forum full of strangers built to describe them. It's about why so many accounts converge when the people giving them have never spoken. And it's about the part nobody planned for, which is the travelers who show up at the door and get turned away.Along the way I get into predictive perception, Caputo's strange-face research, Shibutani's work on rumor as improvised news, and the liminality framework that explains why an experience with gatekeepers feels like an initiation. I also come back to the Lanmaoa asiatica episode, because the little people in Yunnan complicate the tidiest explanation available.The book behind a lot of this episode is Rick Strassman's DMT: The Spirit Molecule, and it's on The Strange Shelf along with Terence McKenna's True Hallucinations: https://www.amazon.com/shop/influencer-b7bd654b/list/1OX5UZT4WR668?ref_=aip_sf_list_spv_ons_mixed_dThe companion essay goes up this week on Substack and takes the Yunnan question further than I could take it here. Paid subscribers get ad-free listening and the full essays: https://substack.com/@psychstrangepodUntil next time, stay curious and stay strange.

Witch's familiars, psychic pets, and telepathic cats: this episode digs into the centuries-old belief that animals can sense things they have no ordinary way of knowing. From seventeenth-century witch trial confessions to a modern scientist's experiments on dogs that wait at the window, I trace how the same core claim has survived in different forms for hundreds of years. I start with the witch's familiar in early modern England and Scotland, including Elizabeth Clarke's 1645 confession under Matthew Hopkins From there I move through shamanic animal guides and the Norse fylgja, and what these traditions across cultures agree on even when they disagree on mechanism. Then it's J.B. Rhine's psi-trailing research at Duke University in the 1930s, followed by Rupert Sheldrake's experiments on Jaytee the dog and his ongoing dispute with psychologist Richard Wiseman over what that data actually shows. On the skeptical side, I get into the Clever Hans effect and ideomotor cueing, along with how cats may actually find their way home through scent-mapping and possible magnetoreception. I also spend some time on apophenia, grief, and why the story that includes you always feels truer than the one that doesn't. Toward the end I bring in the Earth Species Project's NatureLM-audio and what AI-decoded animal communication might mean for this whole question going forward. My own cat Sabrina makes an appearance too. The companion essay for this episode, which goes deeper on the witch trial records, is up on Substack now. Paid subscribers get that essay in full along with ad-free listening. Until next time, stay curious and stay strange. Hosted on Acast. See acast.com/privacy for more information.