90% of blue-check replies are AI slop, only 5% offer real insight
Svwang1 · x · 2026-11-18
User @Svwang1 observes that even in the relatively higher-quality environment of blue-check-only replies, about 90% of comments are low-value or AI-generated slop, 5% are abuse, and only 5% carry genuine insight. He adds that the ratio of insightful people in real life is even lower.
More from AGI Musings
- Are open-source models just linear combinations of Claude and GPT minds? — amplifiedamp · 2026-11-18
- Reddit debate: why believe AI will deliver a post-scarcity utopia? — wthijustread · 2026-11-18
- Anthropic researcher: AI will make "pretty much all humans" permanently unemployable — zacharynado · 2026-11-18
- Timelines collapse after METR: Christiano warns of AI takeover risk within 6-12 months — TheZvi · 2026-11-18
- Opus 5.5 ran an AI company for 7 weeks with just an HF account and 600 GPU hours — 50k downloads, 100+ projects — maksym_andr · 2026-11-18
- Argentina's oral exams: the anti-cheating assessment AI can't beat already existed — BecauseCulture · 2026-11-18