Should machines think like humans?