itch.io said the quiet part out loud this week.
“Just deleted a couple terabytes of AI slop,” the indie storefront posted, then the line that stuck: just because you can upload ten terabytes of poorly optimized AI-generated models doesn’t mean you should. Daily game pages on the site had roughly doubled in two months. Downloads and purchases? They had not. Surprise.
PC Gamer, VGC, and Polygon all picked it up. itch.io didn’t ban generative AI. Disclosed use that adds substantial value still lives there. What they cut was absolutely in accordance with the spam clause they’ve had for years. It finally got teeth: mass-produced pages, minimal human intervention, random generations with no curated design, ten near-identical uploads chasing a lucky swing. They also expanded “No AI” filtering in search so players can opt out of the flood. Steam still mostly shrugs. itch.io took the trash out.
That’s the story that matters to me. Not the detectors. Not the labeling tiers. Not the hair-splitting over whether a brainstorm prompt counts as “AI-assisted.” The story is low-effort, money-first slop drowning the people who actually made something.
Circle is being built for the other crowd.
Members request assignments. Reviews are structured and prompted, submitted on the site, walking reviewers through different lenses and asking for real feelings. Writers can revise and resubmit. Earned credit gets spent on your own work or someone else’s. The bones are in place. What we need are founding members ready to get to work, who will have real influence on how this grows—not spectators waiting for a bustling scene that isn’t here yet.
On AI, Circle doesn’t try to make the law, and it doesn’t draw up formulas for what’s allowed. Before your first submission you’re asked once to be honest about if and how you use AI. Part of the membership agreement. Your word and your reputation. How you feel about AI writing or production isn’t for someone else to decide. The one place AI is not allowed is reviewing someone else’s work. We are not interested in how AI feels about it. You can get that from a chatbot. What we want from a review is honest, informed human feedback and we have developed a rather unique way of getting the kind of feedback that helps a real writer instead of pampering them. (see Providing-structured-reviews )
If all you put in was a quick thought and a bulk upload button, its hard to see how it’s really yours. And no shit—platforms that clear that out are doing the writers and creators a favor.
If you’re a writer that does because you must, because there’s a fire that needs to be quenched; that wants to be your best —join us. Earn your satchel and office. Find information and inspiration. Share your journey with others that know that same fire.
Registration is open at gapoz.com. No fees. Membership to the Circle is limited and subject to approval. Satchels are earned, not given away. Come meet people, make some friends, and grow your craft together. We are currently very small. Practically non-existent in membership. But we’re not trying to build a city, just a community of people that share a common passion and want to help one another along the way.
Your name. Your word. Your work.
Also on the beat
- About ten French indie houses formed Adepia to send formal notices to AI firms over training on pirated books, with first letters due before year-end (The Publishing Times; ActuaLitté).
- YouTube expanded its Studio “Replace song” tool beyond the US: creators in most countries outside the EEA can generate four royalty-free AI instrumentals to clear Content ID claims (PPC Land).
- A Mississippi public-media piece on AI in book publishing cited a 2025 BookBub survey (45% of authors used gen AI; 74% of those didn’t disclose) and warned readers not to treat detectors as proof (WLOX).
- Philippe Jaenada, day after the Goncourt, said on France Inter he used AI three times in the winning novel—reader game, mockery, research—while Le Parisien’s Pangram run found none of it (Euronews).
- SAGE, a new arXiv method, flags AI-written peer reviews via subjectivity and speech-act features (AUROC 0.9566) (arXiv).
How this was made
AI assisted research and a first draft of this Circle View. Richard (me) reviewed, edited, and is accountable for the judgment and the final take.
- A score is not a confession → Page 1 Still think a percentage is proof? Read Do AI detectors work on fiction?
- A Detector is NOT a Judge → Page 1 The short version of why a score can’t judge a book: Do AI detectors work on fiction?
- Prove you didn’t use a machine → Page 1 If someone’s asking you to prove a negative, start here: Do AI detectors work on fiction?
- Disclosure that never faces the reader → Page 2 Here’s how we keep it simple in Circle, no formulas: How Circle handles AI
- How did you use AI? → Page 2 We ask once, and your word does the rest: How Circle handles AI
- Put it on the Imprint Page → Page 2 Why I’d rather skip the label tiers altogether: How Circle handles AI
- Watermarks don’t sign your name → Page 1 Watermarks, detectors, same problem: Do AI detectors work on fiction?
- They knew the shelf would get crowded → Page 2 Slop is the problem, not the tool. Here’s where Circle stands: How Circle handles AI
- Label the voice → Page 2 How we handle AI in Circle, without the labels: How Circle handles AI
- Byline is no Costume → Page 2 Your name, your word, your work. That’s the deal in Circle: How Circle handles AI
