The current goal/trend seems to be to mimic the human brain in software/algorithmic form, but that standard and comparison might be part of the problem. It will certainly be a massive task to accomplish. Maybe we need to think about neural networks and the like as more of a colony of bees or farm/hill/whatever of ants. A bunch of individual units that work together in ways that appear magical/mysterious to us to achieve complex results. Plus, nature usually figures things out in a much more elegant way than we do.