🚨 ELON: AI MUST BE MAXIMALLY TRUTH-SEEKING "The thing that I think is most important for AI safety is that AI must be maximally truth-seeking. This is very fundamental. You can think of AI as a super genius child that will outsmart you, but you can still instill the right values." Source: @xai @elonmusk
Mario Nawfal
Mario Nawfal10.7. klo 12.48
🚨GROK 4 SHATTERS ARC-AGI BENCHMARK WITH 60.6% The reasoning test that stumps most AIs just met its match. Grok 4 hit 60.6% on ARC-AGI, leaving OpenAI's best at 41.0% and everyone else in single digits. Cost per task? Just $18. Claude needs $107 for worse results. This benchmark tests abstract reasoning - the kind humans use to solve puzzles they've never seen before. Grok just proved it thinks different. Source: @xai @elonmusk
96,56K