In this episode, we celebrate 50 episodes (!!!!!) of Bad Lit Friends by doing what we do best: yelling at the internet.
First, Kirk has actual writing-career news, and Meagan's book remains perpetually one chapter away from being finished, but this time she means it, seriously, no joke, fingers crossed.
Then we return to one of our favorite recurring nightmares: AI detection, literary witch hunts, and (what the actual hell?) the massively disturbing idea that writers should have to prove they wrote their own books. Meagan dives deep on a university preprint about AI-generated fiction, Pangram's AI detector, The Atlantic's reporting on H.M. Wolfe's Daggermouth, the implosion of Jerry Falade's seven-figure book deal, and the extremely convenient timing of Pangram's shiny new model and fresh pile of venture capital.
We also dig into what AI detectors actually measure, why genre fiction is particularly vulnerable to false positives, what "rare expressions" do—and absolutely do not—prove, (oh no, I used em dashes!) and why asking one probabilistic machine whether another probabilistic machine wrote your book might not be the forensic slam dunk the tech bros are trying to convince you that it is. Meagan also brings receipts from her own deeply semi-scientific Pangram experiment, in which (spoiler) actual AI-generated writing was confidently declared 100% human. Whoops.
This is only Part One. There is, unfortunately, so much more.
Welcome to Bad Lit Friends. We're glad you're here, even if none of us know what we're doing.
Bad Lit Friends was created and is hosted by Meagan Thompson-Mann and Kirk Rafferty.
Show email: [email protected]
Bluesky: @badlitfriends.bsky.social
Threads: @badlitfriends
Instagram: @badlitfriends
YouTube: https://www.youtube.com/@BadLitFriends
Website: badlitfriends.com
Show notes:
Stony Brook preprint: https://arxiv.org/html/2510.13939v4
First bubbling on Threads: https://www.threads.com/@thewellreadnurse/post/DbUNQgikTM5
The Atlantic piece: https://www.threads.com/@theatlantic/post/DbUB1ZClON-
Writers contacted by the preprint authors: https://www.threads.com/@bykenziejames/post/DbUi6pKFlAW; https://www.threads.com/@authorleighmiller/post/DbUfWoCDYzP
The supposedly telltale phrases sound awfully familiar: https://www.threads.com/@dianaurban/post/DbUVaMiAHSU
Reddit enters the chat: https://www.reddit.com/r/fantasyromance/comments/1v8jy52/comment/p07ke6k/?share_id=jWP7w6W-aaJ82WnwPzDSy
An article promoted by the author of the Atlantic piece which coincidentally was written by the person who wrote the Mia Ballard takedown: https://www.nytimes.com/2026/07/28/books/ai-bookselling-amazon.html?unlocked_article_code=1.1VA.TUP9.b3IGCF6Kum1h&smid=url-share
Falade’s deal falls apart: https://www.threads.com/@dianawallachauthor/post/Dbb5nvSCEV3
Publisher’s Lunch does cool shit: https://www.threads.com/@authorpenncole/post/Dbb3M0wjGTA; https://www.threads.com/@fatedlovesreturns/post/Dbb5pQwjz1t (screenshots included of article)
The Bookseller weighs in: https://www.threads.com/share/_nf25Ylhd/
Quotes from Falade: https://www.theinsneider.com/p/ai-scandal-rocks-publishing-world-and-hollywood-black-author-minotaur-macmillan-auction-uta-tv-show
Substack x Pangram: https://mediacopilot.ai/substack-ai-detection-pangram/
The woman who wants writers to prove it quote-threaded by Kirk: https://www.threads.com/@kirkraff/post/DbY1iCGmQug
