xAI's Grok 4.6 model card says it outright: "we don't secretly make your model dumber"
XFreeze · x · 2026-09-11
A blogger spotted that xAI's official Grok 4.6 model card includes the line "we don't secretly make your model dumber" — an open jab at AI companies that quietly nerf their models a week or two after release. The post also teases Grok 4.7 as coming soon and praises xAI's stated philosophy: either make models more efficient or help users get more out of their subscription.
More from Fun
- GPT-iterated browser racing game gets smarter rival AI and smoother visuals — TheMoonMidas · 2026-09-12
- Muse AI avatar mysteriously acquires The Orb — billyjhowell · 2026-09-12
- "My dad just smokes weed and uses ChatGPT" goes viral as a relatable AI meme — moonsandhues · 2026-09-12
- Perfect AI use case? Asking a model to merge 7 podcast episodes into one — billyjhowell · 2026-09-12
- "LLMs Cure Cancer" Article Goes Viral as Rob Leclerc Quips About Humor — robleclerc · 2026-09-12
- Paper-Tearing Comparison Video Shows a Year of Video-Gen Physics Gains — sabage27 · 2026-09-12