Viral screenshot shows AI live-factoring a rant's claims against scientific evidence
juliey4 · x · 2026-10-15
A viral screenshot shows a chat UI displaying the model running "Assessing a rant's claims against scientific evidence" in 10 seconds mid-conversation, while discussing claims like cotton absorbing up to 10x its share of VOCs. It's a striking example of models visualizing their reasoning as spontaneous fact-checking during casual chat.
More from Fun
- Watch Astra Draw Its Author With Fourier Epicycles — flowersslop · 2026-10-15
- The AI 'taste' debate is just the American Psycho business card scene — dreamwieber · 2026-10-15
- X users riff on Tim Cook and Alexandr Wang with 'business up front, party in the back' — roydanroy · 2026-10-15
- Safety researcher quips: one spy found in your AI company means there are more — dhadfieldmenell · 2026-10-15
- Debate: If 90% of Your Code Moves JSON and Writes UI, AI Should Take It — gdechichi · 2026-10-15
- "It's a small world": network map links LessWrong, MIRI, Ethereum DAOs and civic tech communities — ryan_t_lowe · 2026-10-15