Instead of using one-dimensional FFTs—the staple of all work on this problem since 1971—our algorithm relies on multidimensional FFTs.

To this end, [Ken Shirriff] set about reverse engineering the 8086’s multiplication algorithm. [Ken]’s efforts were achieved by using die photos of the 8086 chip.

DeepMind breaks 50-year math record using AI; new record falls a week later AlphaTensor discovers better algorithms for matrix math, inspiring another improvement from afar.

AI engineers develop an algorithm that will replace floating-point multiplication with integer addition to make AI processing more efficient.