It's not a surprise that after brute forcing a problem with neural networks and generating a black box model it becomes much faster than on-the-fly computations

It's not a surprise that after brute forcing a problem with neural networks and generating a black box model it becomes much faster than on-the-fly computations. The post made me remember this article about predicting chaotic equations with machine learning with great accuracy: https://www.quantamagazine.org/machine-learnings-amazing-ability-to-predict-chaos-20180418/ https://www.quantamagazine.org/machine-learnings-amazing-abi...