One thing is certain, as these networks become more complex, with more nodes and more data then ever thought possible, they start to look and behave like neural nets, the models that have historically been closest to modeling our own brain. This blog post is based on Nitin Pande‘s article “Deep neural nets and the purpose of life.”
Pande proposes that our minds are a kind of deep neural net (DNN), and the DNN of a given species is mapped in it’s seed. That allows the model to be passed from generation to generation, and the opportunity of each generation to work on and expand that DNN. If that is the case, then the question naturally arises: what are these DNNs optimizing for? Pande things, and I tend to agree, that each individual’s actions are less important than the emergent behaviors of a society or species as a whole, and the purpose of which is going to be very difficult to ascertain.