Grok-3 just proved Riemann's hypothesis. We decided to pause its training to check its proof, and if the proof is correct, training won't be resumed, as the AI is deemed so smart that it becomes a danger to humanity.