Integer multiplication below n log n

Researchers at OpenAI have discovered a new algorithm for integer multiplication that can outperform the standard O(n log n) time complexity. The algorithm is based on a new mathematical technique called 'tensor decomposition' and has been implemented in a code repository on GitHub. The algorithm has been tested on a range of datasets and has shown to be faster than the standard O(n log n) algorithm in many cases, with a median speedup of 1.1x on a benchmark dataset of 1000x1000 matrices. The code repository also includes a comparison of the new algorithm with the standard O(n log n) algorithm on various hardware platforms, including CPUs and GPUs.