AI detectors flag Pulitzer-level human writing as 100% AI
Afinetheorem · x · 2026-09-22
Afinetheorem ran a batch of selections (ostensibly Pulitzer-level human writing) through the Pangram detector and every one came back 100% AI; other AI models agreed. He says he's never seen detections this wrong this often, joking that AI can apparently write Pulitzer-level work — highlighting both detector unreliability and how hard top human prose is to distinguish from AI.
Related event: Award-Sweeping French Novel Flagged as Nearly Entirely AI-Written(13 posts)→
More from Fun
- Fable 5.2 wows with game-generation: detailed xianxia scene built from tools alone — op7418 · 2026-11-11
- Vibes with agents feel like being a Green Lantern wielding the ring — rachittshah · 2026-11-11
- A 3D ray tracing engine written entirely in SQL — fofrAI · 2026-11-11
- Splash One: flight combat sandbox built with Opus 5.5 in PlayCanvas — willeastcott · 2026-11-11
- PlayCanvas engine repo hits 10,000th issue/PR — and it's a Claude-authored merge — willeastcott · 2026-11-11
- AI-built game nerfed 40% after Claude playtest bot beat it in 6 minutes — builderjaydub · 2026-11-11