Quora – If we ever achieve AGI, how likely is it that it will stem from machine learning algorithms specifically? Are other branches of AI more likely to achieve AGI or will AGI likely be a combination of several different branches?
The simplest human being capable of speech is a system of at least 15 levels of sets of complex adaptive systems with many sets of heuristic priors deeply tuned by genetic and social evolution. Not 15 systems, 15 levels of systems – thousands of sets of tuned complex systems all interacting.
An AGI that achieves anything near human intelligence (or beyond) is going to be at least that complex.
It must involve thousands of interacting systems of context sensitive checks and balances on various sets of context sensitive valences.
It is going to be complex beyond the ability of any mind to understand in detail, and we are now getting some reasonable maps of the sorts of strategic territories that it must embody at each new level of complexity.
AGI is likely to be a combination of every branch of knowledge known, and is likely to rapidly explore many more far more quickly than we humans have.
It is seriously, mind numbingly complex.