There was a result recently of modeling an organic neuron with 1000 digital neurons.

And even if that result was perfect modeling of the neuron, that assumes perfect and exhaustive data readings on the organic neuron, which is, frankly, unlikely. (Not that I know how to estimate how much it's missing, but I don't think we fully understand a single neuron yet.)