What we mean by "intelligence," must also be considered. As they are, computers are most useful for solving problems where a solution is known, theoretically,

What we mean by "intelligence," must also be considered. As they are, computers are most useful for solving problems where a solution is known, theoretically, to exist. For example, you know there are articles on the internet concerning Javascript, and when you ask google to point you towards them, the relevance of the returned results is often viewed as a measure of the algorithm's "intelligence." But behind this intelligent response lies complex but defined probability calculations that are provably correct given a known algorithm. On the other hand, can you ever imagine asking the questions, "Computer, does this shirt look good on me?" or, "Computer, what is the most ethical course of action in this situation?" How a human responds to these questions helps indicate that human's level of intelligence. But different people could also disagree on the above questions and still be considered intelligent. To me, it doesn't seem to make much sense to invest in computers that are modeled after humans, as there are certain types of situations where even the most intelligent of these machines could be wrong as a matter of opinion. Rather, the more powerful computers become, the more sense it makes to have them tackle finite problem spaces more efficiently. Because of this, I don't really see the case where human intelligence and computer intelligence would or should converge. Put simply, it would NEVER make sense to give Skynet control of our nukes.