I think the application of neural networks on quantum computers will be interesting due to the potential to explore probabilities and decision making that isn't possible on current processing units. Who knows if any headway will be made even with that tech, but certainly don't think we are close with what we have now. For reference I remember the hopeful saying this exact same time frame back in 1989 and every year since (and they were saying similar hopes back in the 70s as well). I err on the side of pessimism in the sense that the optimists have consistently been wrong.