AI pundit realizes he can't tell real Chinese posts from AI slop via Grok translations
teortaxesTex · x · 2026-10-18
teortaxesTex admits he can no longer distinguish real Chinese posts from AI-generated ones: he only reads Grok translations, which gave him a stereotype that Chinese posts have an "autistic disagreeable robot" quality — until someone pointed out many such posts are blue-check AI slop farmed for traffic money. The quoted screenshot also contains a mini-debate on whether small models are made by distillation vs pruning/retraining, with MING arguing V4.1-level gains come from data mix ratios and rejection-sampling thresholds, and Teortaxes countering that OpenAI openly distills big models into small ones.
More from Fun
- Skipping the model, scrolling straight to the pricing 😌 — wateriscoding · 2026-10-18
- AI-generated buzzword calendar goes viral: every slot from 6 AM is corporate jargon — StewartalsopIII · 2026-10-18
- Heartbreaking: the AI bot in your replies just made a really good point — menhguin · 2026-10-18
- The "Crazy Anime Wizard" Scene Taking Over Communities — 8th_circle · 2026-10-18
- 'Claude: a factory for lil dudes' — the agent-spawning meme captures how devs use it — Aryvyo · 2026-10-18
- AI Recreation of Original Footage Draws Community Attention — mhu99 · 2026-10-18