I remember reading about two different types of goals for an intelligence. The gist was that the first type of goal is one that "just is" - it's an end goal for the system. There doesn't need to be any justification for wanting to achieve that goal, since wanting to do that is a fundamental purpose for that system. The second type of goal is a stepping stone, for lack of better words. Those aren't end goals in and of themselves, but they would help the system achieve its primary goals better.

I've forgotten the names for these types of goals and Googling didn't help me much. Is there a standard definition for these different types of goals?


Do you mean weak AI and strong AI? The former is roughly about pretending to be intelligent, ie do intelligent things, but without trying to work the same way an actually intelligent system would work. The latter attempts to replicate how an intelligent system works, so would require us to understand a lot more about cognition than if we just mocked up a quick little chatbot to try and compete the Turing Test.

  • $\begingroup$ I don't think the OP was referring to "weak AI vs strong AI" (but it's also possible). To me, it seems they are talking about "intermediate goals", "motivational goals" and "main goals" (I don't think this is standard terminology anyway) . $\endgroup$ – nbro Mar 1 at 21:23
  • $\begingroup$ @nbro Yes, it's a bit of a vague question. $\endgroup$ – Oliver Mason Mar 2 at 10:05

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.