Heavy AI users spot AI-written book instantly, while Reddit skeptics trust no detector
Afinetheorem · x · 2026-09-22
Afinetheorem surfaced a French Reddit discussion about a book that is obviously AI-generated. Most commenters dismissed AI detectors as unreliable, while heavy AI users immediately recognized the telltale "AI slop" writing style. The author argues people who don't use AI heavily simply don't notice these tics: he identified the book as AI-written within a few chapters, and notes that GPT and Claude flagged exactly the same patterns he spotted — no dedicated detector like Pangram needed.
Related event: Award-Sweeping French Novel Flagged as 100% AI-Written(7 posts)→
More from Fun
- "How many agents do you have running?" "Me: Yes." — draginol · 2026-09-22
- Suspicious 'Googlebook' laptop preorder page surfaces at $899, likely parody — ZoubinGhahrama1 · 2026-09-22
- Mathematician Jokes: Hide AI Proofs in Plastic Eggs for Academics to Hunt — RexDouglass · 2026-09-22
- Claude refuses to push a person off a ledge, steps back 'to keep a safe distance' — repligate · 2026-09-22
- Doomers' new hope: can't stop ASI, so at least make it lame — repligate · 2026-09-22
- LinkedIn drove 3 profile views in 6 weeks; X brought money, jobs and VCs — andrew_n_carr · 2026-09-22