Man Enters ChatGPT Prompt Loop, Emerges Believing He’s a 13th Century Herbal Knight
By Staff Reporter, SpinTaxi’s Office of Prompt-Induced Delirium
Reno, NV — In what scholars are calling the first “Medieval Identity Crisis Caused by a Language Model,” 29-year-old Tom Keller has rebranded himself as Sir Thomás of the Prompted Grove, a self-declared 13th-century herbal knight with a solemn duty to “restore balance to the kingdom of Nevada using tinctures and ancient banter.”
It began, innocently enough, with a prompt:
“Write a Dungeons & Dragons backstory for a medieval healer with anxiety.”
ChatGPT responded with a rich, highly detailed saga involving lost scrolls, cryptic ravens, and a mole with trust issues. That’s when the trouble started.
“I kept asking for more,” Tom admitted. “A few sessions in, I wasn’t playing a character anymore. I was the character. I started talking like I’d eaten a thesaurus dipped in thyme oil.”
His friends noticed a change when he began insisting his Prius was a “four-wheeled steed consecrated by forest druids” and refused to enter grocery stores “without permission from the village elders.”
He was later detained after attempting to barter for pizza using homegrown basil and a mason jar labeled “Elixir of Digestive Blessings.”
The Prompt Loop Explained
Psychologist Dr. Elsa Romero describes this as a “recursive identity feedback loop.”
“Each prompt builds upon the last, rewarding the user with deeper immersion. After 90 iterations, users can no longer distinguish themselves from the character ChatGPT constructs for them,” she explained while flipping through a file labeled “Chronicles of Sir Promptalot.”
In Tom’s case, the loop reached critical absurdity after he requested a medieval monologue, an herbal grimoire, and a poetic eulogy for his missing AirPods.
Digital LARPing, Now With Delusions
Tom now operates a roadside apothecary called “The Syntax Grove,” where he sells sage bundles, beetroot ale (it’s just kombucha), and hand-drawn maps of imaginary villages.
Tourists are charged $15 for a “soulwarding” ceremony involving lavender oil and GPT-generated incantations. He carries a wooden staff he claims “sings when near forbidden Wi-Fi.”
When approached by local police, Tom declared, “I am under oath of the Green Order and answer only to the Guild of Bandwidth.”
OpenAI’s Legal Response
OpenAI, under growing pressure to address prompt-induced psychosis, issued a statement clarifying that “medieval hallucination is not a documented side effect of language model interaction,” but advised users to avoid “layered character loops exceeding 25 iterations per hour.”
They also recommended not combining GPT roleplay with herbal tea, isolation, and a desperate need to feel important.
An internal document leaked to SpinTaxi showed staff privately referring to such users as “Knightmares.”
Family Intervention Goes Awry
Tom’s family staged an intervention. It failed after he demanded they each present “a dowry of symbolic wisdom” before he’d allow them into his tent (a repurposed Costco canopy lined with fantasy maps).
His cousin was turned away for failing to speak “in verses.”
His grandmother was allegedly told, “Your presence in this realm is welcome, crone, but thou bringest no mint, nor moonwater.”
She now communicates only via Middle English emoji.
What the Funny People Are Saying
“He thinks he’s a knight but lives in a van. That’s not a delusion. That’s just Burning Man off-season.” — Ron White
“ChatGPT gave him a sword and a backstory. That’s what happens when therapy gets replaced by Mad Libs.” — Jerry Seinfeld
“You know your son’s gone when he asks Alexa to summon the Blacksmith of Spotify.” — Sarah Silverman
“When I tried to be a medieval knight, my mom just grounded me. This guy gets a booth at the farmer’s market.” — Bill Burr
“He charges $40 for beetroot in a jar and people call him wise. When I do it, it’s ‘food poisoning.’” — Ricky Gervais