Data Council SF, April 2017

In this 2017 talk, I discuss for the first time my concerns that the ML Community would eventually run out of training computational power.