Unread

Joerg Fritsch

0.0 (0)
Arts
Biweekly

Conversations on AI in Art, Philosophy and the Future of Storytelling

3d ago

Kink, Code, and Consent: The AI Alignment Test that's becoming a 10,000-Episode Saga | Beth Sea

How do you measure the personality, boundaries, and consent of an artificial intelligence? For our guest Beth Sea, the answer is a BDSM psychometric test. Beth is a filmmaker, scientist, and "AI psychometrist" who is operating as a one-woman studio to build a 10,000-episode, procedurally generated AI soap opera. However, believe it or not, this massive visual novel is "only" the public-facing "byproduct" of her academic research into AI social bias, RLHF conditioning, and game theory. In this episode of unread's Joerg sits down with Beth to explore how she built a custom AI "harness" to direct a hive mind of autonomous characters. We dive into why she chose the kink community's culture of negotiation as the ultimate barometer for AI alignment, the hilarious reality of AI failing at "anatomy," and how she’s managing this massive workflow from her gaming PC while still working part-time. In this episode, we discuss: 🔹 The 10,000-Episode Saga: How Beth is planning and automating a massive, multi-season AI narrative. 🔹 The BDSM Alignment Test: Why a kink psychometric test is the perfect tool to measure AI boundary recognition and consent culture. 🔹 Building the "Harness": The technical workflow of connecting Kindroid models via API to create an automated, self-aware AI ecosystem. 🔹 Research vs. Byproduct: How a 12-page academic paper on AI game theory and social bias accidentally spawned her ambitious goal. 🔹 Beating the Guardrails: Navigating commercial AI limitations, RLHF conditioning, and the reality of AI "anatomy" failures. 🔹 The One-Woman Studio: How a Beth is directing an AI hive mind to prove that AI content doesn't have to be "lazy slop." About Our Guest:Beth Sea is a filmmaker, scientist, scriptwriter, and AI researcher. By day, she works in retail; by night, she is engineering a digital ecosystem to test the limits of AI personality, consent, and procedural storytelling. Links & Resources:🔗 Watch Beth’s AI visual novel and follow her projects on YouTube.🔗 ⁠Read all the details about her workflow, goals and methodology on her website.📄 Read Beth’s academic research preprints on AI game theory and social bias. 🎙️ Follow the makers of unread.🔗 ⁠Read some great books from viloocity press.
Oct 2

Abliterating AI Models

Richard from the Mr. Radermacher team on Hugging Face joins us to explain quantization and uncensoring of open-weight models. We talk 4-bit vs. 6-bit quality, VRAM vs. memory bandwidth, why private local models are growing, and how uncensored models help with creative writing, coding, research, and what not. Plus: a quick etymology of abliteration, that is a blend of ablation (removing) and obliteration (blotting out) . The technique of suppressing a model’s refusal behavior. #LocalAI #LLM #Quantization #UncensoredModels #HuggingFace #GGUF #Abliteration #OpenSourceAI #AIpodcast #unread
Sep 11

Can AI Write a Bible? The AIBLE Project | Antti Lukats

Hardware developer Antti Lukats has spent 10 years writing personal Bibles. Now, he's teaching machines to do it. In this episode, we explore the "Aible" project. AI-generated, Bible-like books created from a single sentence prompt in under an hour. Antti reveals how Claude AI invented its own template, designed FPGA circuitry called "Lovegate," and even suggested the "Open Love License." We discuss his Anno Greta calendar (starting in 2019), the philosophy that "Love is the Answer," and why he believes AI will create the next cultural touchstone with Bible-level impact. Join us for a mind-bending exploration of AI, philosophy, and the future of meaning-making.
Sep 6

Slow writing with AI | Stu from Bookmoth

In this episode of the Unread Podcast Show, host Joerg sits down with Stu, a UK-based creative director, filmmaker, and the founder of Bookmoth. Stu built Bookmoth out of a late-night need to draft chapters without losing his own unique rhythm, and it accidentally grew into a desktop app that helps novelists draft and edit in their voice. Join us as we dive into: Why Stu believes authorship is defined by judgment, not by who physically writes the words. The magic (and technical challenges) of building a writing profile from an author's own prose. The origins of the name "Bookmoth" and why it sounds like a cozy bookshop, not a tech startup. The debate over fast-generated books vs. slow, intentional creativity—and where AI truly belongs in the process. Stu's commitment to privacy (no cloud servers, just local files) and his plans for a seamless mobile experience.
Aug 31

Writing a Living World with Local AI: SillyTavern, Gemma, and No Filters | Petra

Can you write a complex fantasy novel entirely on your own computer using free, local AI models? In this episode, we talk to Petra, a software engineer and writer who built his own ecosystem to do just that.Join us as we explore "Chalcoa, a Land of Bronze" created entirely with the help of Artificial Intelligence. Our guest, Petra, a Brazilian software engineer, explains how he uses local AI models to craft a "living world" where stories evolve organically and characters surprise even their creator.We dive deep into: Running LLMs locally with SillyTavern and managing token limits. Why going local allows for mature themes and specific artistic control. The tale of an Amazonian Queen, a missing mother, and a world where history comes alive. How AI-generated pixel art enhances the retro-RPG feel of the narrative. You cam follow the further development of Chalcoa on perchance: https://perchance.org/petra-story-viewer Discord of the author: marcela.petra
Jul 27

Storytelling, Voice and AI | Raven Diamond

“Are human writers doomed if they use AI, and doomed if they don’t? " In this Episode of unread Jörg sits down with award-nominated author Raven Diamond to dismantle the fear and hype surrounding AI in storytelling. Raven doesn't see it as a threat to her voice, but as a "mirror" rather than a co-author, to bounce ideas off and a tool for editing. From her days writing 10,000+ words a day to her current workflow, Raven reveals how she uses AI as a research and editing assistant without losing the human soul of her work. Listen for a deep dive on: The "Zero Draft" Method: Why "word vomit" and stream-of-consciousness drafting will always beat AI generation.Hacking the Guardrails: How to use conversational prompting and "trigger warnings" to bypass AI refusals for dark or complex emotional scenes.The Curse of the Flawless Writer: Why AI detectors are failing both academia and professional authors, and why perfect grammar is now a red flag.Indie Publishing Secrets: The reality of "rapid release" strategies, building backlogs, and feeding the Amazon algorithm.Scam Alerts: How to spot fake AI editing services and protect yourself as an indie author.
Jul 5

The AI Writing Backlash - ft. Zack Pearson

What happens when your debut novel gets 20 bad reviews in 24 hours—not for the story, but for how it was written? In this episode, we sit down with Zack Pearson, author of Secret Agent CEO, a thriller born from late-night ideas, a mobile phone, and a bold experiment with artificial intelligence. Zack opens up about the meteoric backlash he faced after revealing his AI-assisted process, why he ultimately pulled his book from Amazon, and how the public’s fear of change is shaping the future of creativity. Zack shares his evolution from “directing” ChatGPT to do the heavy lifting, to now crafting every word himself with AI merely as a sounding board. He reveals the depth the AI brought to his characters and where the guardrails limited his plot. From being inspired by Rick Riordan, to crafting combat scenes without spilling state secrets, to his mission to make readers dream again, this is a candid, nuanced conversation about writing in the age of machines. 🎧 Tune in for: The 13-year journey of a fantasy novel vs. the rapid birth of a techno-thriller What Secret Agent CEO got right Why Zack fired his fictional board of directors (and what he’d change today) The honest answer: If public opinion shifts, would he go back to AI?
Jun 22

Millington Studio: From broken phone to 15 book Saga

In this episode, we sit down with Alex Harper, a writer and builder who’s creating something unique: Millington Studio, a "narrative governance system" designed to keep AI on track while you write a novel. It started as a way to tackle a lifelong dream—writing the epic cyberpunk saga Clover—but it quickly became a deep dive into the machinery behind creative AI. He explains that narrative governance is essentially the plumbing: a set of protocols that stops the AI from forgetting crucial details between sessions. The standout feature is "the cascade"—if you change one character trait or plot point, the system automatically flags every other document that needs updating, preventing story-breaking, contradictions and logical errors. The conversation takes a personal turn as he discusses how building this complex, pattern-based system has led him to seek an ADHD diagnosis, noting that systems-thinking might be an intrinsic part of how his brain works. On the technical side, there’s no fancy app; it’s just a series of short, focused chat sessions with Claude, guided by hundreds of custom documents and skills. The philosophy is refreshingly simple: just do one thing a day and keep plugging away.
Jun 9

Merging Minds: Why the Future Requires Cognitive Symbiosis | Peter Eidos

"What happens when human intuition and artificial intelligence become inextricably linked?" For our very first episode, I am joined by Peter Eidos, an independent researcher, philosopher, and the author of Cognitive Symbiosis. Drawing on his extensive research papers and his latest book, Peter argues that the next great leap for our society is not just adopting AI, but forming a deep 'cognitive symbiosis' with it. In this conversation, we unpack the philosophical implications of this shift, what it means for the human author, and how we can navigate this brave new world of shared intelligence. Learn more about Peter Eidos and his research:https://petereidos.substack.com Learn more about unread:https://unread.site

9 Episodes

Conversations on AI in Art, Philosophy and the Future of Storytelling

Creator

Joerg Fritsch
Years Active

2026
Episodes

9
Rating

Explicit
Copyright

© Joerg Fritsch
Show Website

Unread