A new study from the University of Cambridge demonstrates a significant breakthrough in AI-powered weather forecasting. The research introduces a machine learning model that can predict precipitation patterns up to 12 hours in advance with greater accuracy and 10,000 times faster than traditional physics-based simulations. The model was trained on decades of historical radar data from the UK and US, learning to identify the complex evolution of storm systems. While the system excels at short-term, high-resolution forecasts, it is designed to complement rather than replace existing methods for longer-term predictions. The technology could improve severe weather warnings and has potential applications in agriculture, transportation, and disaster preparedness. Read the full article at https://technologyreview.com/2024/05/21/1093035/ai-weather-forecasting-model/