Well, maybe. Roland Piquepaille discusses the new 20 billion node neural network computing system at Artificial Development (site has essentially no content), which is intended to be the "first neural system to achieve a level of complexity rivaling that of the mammalian brain." With a thousand processors working away on 20 billion artificial neurons and 20 trillion connections, it does sound impressive. Piquepaille notes, however, that the company employs only programmers and mathematicians, no neurobiologists or cognitive scientists; we'll see if they can come up with something interesting.
A cursory glance at the human brain reveals a striking level of structural complexity. It is not enough to create a neural net that rivals the brain in number of nodes and connections. The end result of such a project may create interest, but it will be nothing like a thinking brain.
At the very least Piquepaille should consult someone such as Gerald Edelman before finalizing his design.
That's right. The human brain consists of many thousands of parallel neural nets, functioning simultaneously in competition and in concert with each other. Different structures of the brain are highly specialized, and the intricate connectivity of the different neural structures allows for higher order re-entry phenomena, and binding of multiple ideas and sensations.
I'm skeptical whether many of these details have been thoroughly considered by these researchers. Nevertheless, I'm curious to see what they come up with.