Well actually it seems to say in the paper "the same neurons fire when you poke someone with a needle as when you watch someone being poked". So how can they be "mirror neurons"? Seems to me they could hardly be both, "mirror" and "direct" neurons, or at least the term "mirror neuron" does not make much sense if the neuron also acts directly.

Therefore the "software" theory seems much more plausible.

Simply, the brain models it's environment. That is all there is to it. A flower is a flower is a flower.