Tag
An experiment ran the Qwen 3.8 27B model for 63 hours on a RTX 3090 to attempt solving the Riemann hypothesis, demonstrating autonomous reasoning, self-correction, and no hallucination.
Allegations suggest OpenAI may have trained its Astra model on unpublished mathematical research, raising concerns about intellectual theft and potential large-scale misconduct in AI development.
Terence Tao discusses how AI is dramatically accelerating math research, reducing the years of education needed to contribute to the frontier.