> It seems unlikely a few neurons would be more capable than what we can program a CPU to do. Our brain cells are very good at many things we can't duplicate t

> It seems unlikely a few neurons would be more capable than what we can program a CPU to do. Our brain cells are very good at many things we can't duplicate to the same level in software today (e.g. processing real time info for patterns and learning). Even a single neuron can identifying real time patterns (Purkinje cells at the cell level, and in other cells internally local synapse activity patterns are identified and calcium signals forwarded to other parts of the cell). > Unless we don't fully understand what the neurons will do Scientists have a long way to go to understand how neurons function. For example, it was only relatively recently that they accepted that tunneling nanotubes between neurons (and other cells) happen in living brains (as opposed to just in cultures). Now they are trying to understand what these tubes do functionally, why do they dynamically generate on short time scales, why do some of them have electrical gap junctions (typically used for forwarding electrical signals without spiking), and a million other questions.