Given the state of the art in AGI research today, what can we say about the core AGI hypothesis? Is it actually the case that creating generally intelligent systems, requires fundamentally different concepts and approaches than creating more specialized, "narrow AI" systems? Is there a deep necessity for considering "AGI" as its own distinctive pursuit? Personally I am confident the answer to this question is "yes." However, setting aside intuition and looking only at the available relevant science and engineering results, I would have to say that the jury is still out. The narrow AI approach has not led to dramatic progress toward AGI goals; but at the present time, the AGI achievements of researchers explicitly working toward AGI (myself included) have also been relatively modest. There exist a number of theoretical frameworks explaining why AGI is profoundly distinct from narrow AI; but none of these frameworks can be considered thoroughly empirically validated.

The next question, then, is: What is being done – and what should be done – to further explore the core AGI hypothesis, and move toward its verification or falsification? It seems that to move the AGI field rapidly forward, one of the two following things must happen: • The emergence, within the AGI community, of a broadly accepted theory of general intelligence – including a characterization of what it is, and a theory of what sorts of architecture can be expected to work for achieving human-level AGI using realistic computational resources; or • The demonstration of an AGI system that qualitatively appears, to both novice and expert observers, to demonstrate a dramatic and considerable amount of general intelligence. For instance: a robot that can do a variety of preschool-type activities in a flexible and adaptive way; or a chatbot that can hold an hour's conversation without sounding insane or resorting to repeating catch-phrases, etc. Neither of these occurrences would rigorously prove the core AGI hypothesis. However, either of them would build confidence in the core AGI hypothesis: in the first case because there would be a coherent and broadly accepted theory implying the core AGI hypothesis; in the second case because we would have a practical demonstration that an AGI perspective has in fact worked better for creating AGI than a narrow AI approach.

These are still early days for AGI; and yet, given the reality of exponential technological advance, this doesn't necessarily imply that dramatic success is a long way off. There is a real possibility of dramatic, interlinked progress in AGI design, engineering, evaluation and theory in the relatively near future – in the next few decades, and potentially even the next few years. No one can accurately predict the course of development of any research area; but it's interesting that in a survey of researchers at the AGI-2010 conference, the majority of respondents felt that human-level AGI was likely to arise before 2050, and some were much more optimistic. Optimism regarding the near advent of advanced AGI is controversial, but is a position held by an increasing plurality of the AGI community, who are working hard to make their hopes and projections rapidly eventuate.