A new artificial intelligence weather forecasting model developed by researchers has demonstrated superior accuracy compared to conventional physics-based systems in global medium-range weather predictions. The AI model, trained on decades of historical weather data, can generate a 10-day forecast in under a minute on a single desktop computer, while traditional supercomputer-powered models take hours. In head-to-head comparisons over a recent three-month period, the AI system showed reduced error rates in predicting key atmospheric variables like temperature and humidity at various altitudes. Researchers emphasize that AI models complement rather than replace traditional methods, offering rapid scenario analysis for emergency planning. The technology could democratize access to high-quality weather forecasting for regions lacking supercomputing resources. Read the full article at: https://technologyreview.com/ai-weather-model-outperforms-traditional-forecasts