It requires solving lots of problems that I don’t think are obviously bottlenecked on raw intelligence
This is why I’m trying to shift away from “AI” and “AGI” terminology and towards “Outcome Influencing Systems (OISs)”, a set of terminology, lens, and theory agenda I have been developing. The two most important properties of an OIS are it’s preferences and capabilities. “Intelligence” in all of it’s different conceptions are one aspect of capability, and an important one, though I usually focus on “skill” which seems more useful to focus on. But access to and control of resources is also a capability that has nothing to do with intelligence. An intelligent entity using intelligence to recursively improve it’s intelligence may seem far fetched to you, but an OIS using it’s capabilities to improve and gather more capabilities should seem obvious, and that is what we are really concerned with. “Intelligence” be damned.
This is why I’m trying to shift away from “AI” and “AGI” terminology and towards “Outcome Influencing Systems (OISs)”, a set of terminology, lens, and theory agenda I have been developing. The two most important properties of an OIS are it’s preferences and capabilities. “Intelligence” in all of it’s different conceptions are one aspect of capability, and an important one, though I usually focus on “skill” which seems more useful to focus on. But access to and control of resources is also a capability that has nothing to do with intelligence. An intelligent entity using intelligence to recursively improve it’s intelligence may seem far fetched to you, but an OIS using it’s capabilities to improve and gather more capabilities should seem obvious, and that is what we are really concerned with. “Intelligence” be damned.