> What about the old idea of replacing one neuron at a time with a simulation? This sounds disturbingly like replicating a live database by copying one bit eve

> What about the old idea of replacing one neuron at a time with a simulation? This sounds disturbingly like replicating a live database by copying one bit every 162 ms.