The world’s leading artificial intelligence companies are turning to advanced mathematical problems as a measure of progress, using unsolved questions to test their systems in the competitive race to build more capable AI models.
The shift was underscored recently when a University of Cambridge undergraduate used OpenAI’s most advanced model to solve a specific instance of the “Erdős problems”, a series of famous maths calculations, many of which remain unsolved.
The breakthrough follows a string of high-profile milestones, including gold-medal performances by systems from OpenAI and Google DeepMind at last year’s International Mathematical Olympiad and International Collegiate Programming Contest.
您已阅读16%（705字），剩余84%（3798字）包含更多重要信息，订阅以继续探索完整内容，并享受更多专属服务。