Comment by thatgloomyguy

GPU per person or models working on less computing power