They thought it was impossible, but now it has succeeded: here is the twice as accurate weather forecast

Meteorologists struggle with accurate weather forecasts beyond two weeks due to atmospheric noise. Researchers at MIT developed a machine learning method that corrects existing models, doubling the accuracy of forecasts for two to six weeks ahead. This advancement allows for more reliable predictions of extreme weather events, aiding disaster preparedness and resource allocation. The open-source model is cost-effective and accessible for various sectors, including energy and agriculture. Kenya's meteorological service has shown interest in using the model for better monsoon predictions and improving agricultural yields.