This talk will present a brief survey on an approach for designing matrix multiplication algorithms initiated by Coppersmith and Winograd which lead to the current best upper bound on the matrix multiplication exponent.