Google Research has introduced NeuralGCM, a modeling approach that uses AI to simulate long-range global precipitation. The work is part of the company's climate and sustainability research, and is aimed at improving how precipitation patterns are represented over extended time horizons.
According to the source, the key contribution is the use of AI to better simulate precipitation at global scale and long range. The announcement does not provide technical details in the excerpt, so specifics on architecture or validation are not available.
Because there is only one source, no independent comparison is possible. The claims should be read as Google's own description of the work.