Because human beings aren't mechanisms, and when you impose a mechanistic paradigm onto humans, invariably it returns worthless results. Take for example the notion of trying to superimpose computer models on the human brain. A couple of the ideas gain a bit of traction, but ultimately, the analogy breaks down completely. Even with the notion of "sentient" algorithms on the horizon, nobody has shown how the human brain is governed by algorithms, let alone where its central processor is. For memory we at least have some clues.