Intelligence Is A Collection Of Algorithms

Intelligence is a collection of algorithms.

That collection can be large or small. That collection can consist of algorithms that are more or less powerful, more or less useful, more or less efficient.

Collections of algorithms come in endless, neverending varieties. And within any given individual, these collections are ever-evolving. Thus, while we use various ways of talking about and measuring intelligence, these are only ever crude descriptions of the breadth and power of an evolving collection of algorithms.

Among those algorithms, there are algorithms about algorithms. Algorithms about when to use other algorithms, algorithms about how to generate other algorithms.

Some algorithms can run any other algorithm, provided it is adquately described. Some algorithms can generate any other algorithm, provided the right problem situation. We call the first of these, "Turing Complete", and the second of these, "Creation" or "Explanation". These too may come in endless forms and variations.

In this picture, then, terms like AI, AGI, ASI—can have either categorical significance (possessing some particular algorithm), or relative significance (possessing similar algorithms to a certain percentage of the population, etc).

I don't think we know enough to employ categorical descriptors meaningfully. We know enough to assess Turing Completeness, but not enough to assess the capacity for Creation or Explanation. Thus, we fall back on relative descriptors, comparing AI to human populations, and comparing individuals to one another.

AI surpassed the Turing Test (another relative metric) a while ago, and continues to surpass other relative metrics. But there is ultimately no end to the span, or the breadth, or the depth of intelligence we may encounter and explore together.

by Micah Redding -- https://www.facebook.com/groups/ChristianTranshumanistAssociation/posts/2911859925654914/