In the quantum space there are some interesting parallels with what happened in the early computer developments back in the 1940s and the like. What you saw happening in both of those areas is that people would build specific machines for specific purposes. And that's why you saw the explosion of different kinds of machines and so many different companies back in that time frame, because people would build machines specifically for counting or whatever. But then people realized: "Wow, there is generalization here." The same thing is happening in the quantum space. We are seeing that applied to crypt-analysis problems, to problems that are in effect NP-complete, not to say that we are ever going to solve those, because theoretically we can't. We are going to apply lots of computational power and apply statistic means to it. So that's a different mode of computation. But as we start to understand how to actually manipulate these things, you see that community starting to replicate at the quantum level or at the light level, being able to build machines that look computationally very much like what ours are today. So I think that will address the hardware side of things, but on the software side of things, at least for the foreseeable future, the average developer will continue to build what I speak of as small scripts, algorithmic scripts, in the context of a sea of languages. Now, there are other ways to break that metaphor: your notion of systems learning from one another, or learning about themselves, the idea of some degree of autonomousness. And we're seeing that happen in those systems for which we reasonable well enough understand the principles, that we can then build enough intelligence in the lower level pieces to do so. To have that kind of general intelligence, I think we're still a ways off. Yet I will mention now that the notion of the Turing test is: can we have a device that replicates the sort of human experience? And there's a competition over here in that space, and they're getting better and better. We aren't quite there yet, we might be at a point where we can emulate someone of mediocre intelligence, like a vice-presidential candidate or something. We aren't quite there yet for the average human.