Well yeah, but my statement is true anyways IMO. Deep neural networks are unchanged since the eighties I guess, and again IMO it’s not they who need/can have a breakthrough, doubly especially since nobody seems to be working on anything else than “bigger datasets” and profiling and such.