> (In the common case that one matrix is known ahead of time,) our method also has the in- teresting property that it requires zero multiply-adds. These results suggest that a mixture of hashing, aver- aging, and byte shuffling—–the core operations of our method—–could be a more promising building block for machine learning than the sparsified, factorized, and/or scalar quantized matrix products that have re- cently been the focus of substantial research and hard- ware investment.`

This is not at all what modern gpus are optimized for.