Steps are certainly being made, but I don't think anyone in the field said any of the technology that exists today can be iterated upon to get even close to the kind of AI that would eventually become superintelligent. But yes, I do believe that with the amount of people working on these problems, we may very well "hit gold" and come up with something that does have the potential to become superintelligent, before we get there through brute-forcing means.
Why do people who are laypersons to machine learning continue to insist, despite what those in the field say, that we are learning nothing about the principles of general-domain learning from current developments? I think DeepMind is making genuine discoveries: they're certainly able to build things nobody was able to build before.