The brain has redundancy at the level of neurons: it is quite resilient against diffuse neuron loss, and in case of localized damage, unless the affected area is large or includes key regions such as the brainstem, impairment is often limited to one or a few functions, and in some cases it even reorganizes to transfer the lost functions to other areas, partially recovering them.
However, there is no expectation that the brain has redundancy against the loss of an information storage medium that is used in all neurons.
If you destroy half of your collection of DVDs, the information in the other half is still intact. If you destroy every odd-numbered track on all of your DVDs, instead, most of the remaining data will be too fragmentary to be of any use, even if the number of bits you destroyed is the same in both cases.
There can be a lot of redundancy within neurons as well. Just because you find causally relevant chemical densities that predict neuron states doesn't mean that there aren't other chemical densities that also predict those same states.
Question:
His response:
Link