I would have loved to see a big grant like this go into creating a public computer of some kind. The amount of compute required to work on cutting edge deep lea

I would have loved to see a big grant like this go into creating a public computer of some kind. The amount of compute required to work on cutting edge deep learning is starting to go beyond what any lab can afford. A "public research cloud" or something similar would help equalize the playing field for academics. We already have national supercomputers used by academics for all kinds of disciplines, and I don't see why ML should be any different.