Maybe a bit early to be confident about it, but the more we see architecturally simple artificial neural nets able to perform tasks of human perception (speech, hearing, vision) and now language and primitive cognition, the less likely it seems that the brain is using a radically different or more complex approach to do these same things.
Evolution appears (as one might logically expect) to act in localized incremental ways. The result is that our body+brain is more of a modular design than a monolithic intractably complex one. We seem to be well on the way to understanding how these cortical functions of perception and cognition are working and what types of representations are being used, so once we refine our understanding of the other major moving parts such as thalamus (and thalamo-cortical loop) and hippocampus, it seems we'll be well on the way to understanding the overall architecture.
Evolution appears (as one might logically expect) to act in localized incremental ways. The result is that our body+brain is more of a modular design than a monolithic intractably complex one. We seem to be well on the way to understanding how these cortical functions of perception and cognition are working and what types of representations are being used, so once we refine our understanding of the other major moving parts such as thalamus (and thalamo-cortical loop) and hippocampus, it seems we'll be well on the way to understanding the overall architecture.