Comment by deepnotderp

New hardware could greatly reduce inference and training costs and solve that issue