The problem that I see here in comparing 0s and 1s with a photo of a child is pretty obvious.
I was surprised the direction you headed in anticipating skeptics.
It just has to do with comparing unlike things. Or, more specifically, comparing arbitrary and non-arbitrary items as a basis for a particular view.
So what you end up implying, but don’t explicitly stand behind, is the idea that “different” means - more significant, more present, more beautiful or something. Your example, in short, infuses values.
I don’t think that is your intention, and believe me, I get scale.
The meaning of “different” in my thought exercise example does mean something aesthetic and, in the end, something with values. But that is a kind of ancillary point. The generalized rule does not bring values to the table. It just creates the space for difference as scale increases. What is the nature of that difference? Depends on the system in question. I don’t think I’ve seen anyone generalize it or explain it fully.
Yes I saw that my own example includes some “cheating,” so tried to anticipate these valid criticisms. You are 100% right that arbitrary (or context-free) information isn’t really comparable to meaningful or non-arbitrary information (communicated in context). Here the scale just offers an opportunity to do things in a different way
But even this imperfect example is enough to refute the original point I am trying to refute… which is the claim (made by some)that AI is not intelligent any more than a calculator is intelligent. I do not argue the opposite; only that the claim over-reaches. We cannot extend the one case to the other.
The problem that I see here in comparing 0s and 1s with a photo of a child is pretty obvious.
I was surprised the direction you headed in anticipating skeptics.
It just has to do with comparing unlike things. Or, more specifically, comparing arbitrary and non-arbitrary items as a basis for a particular view.
So what you end up implying, but don’t explicitly stand behind, is the idea that “different” means - more significant, more present, more beautiful or something. Your example, in short, infuses values.
I don’t think that is your intention, and believe me, I get scale.
The meaning of “different” in my thought exercise example does mean something aesthetic and, in the end, something with values. But that is a kind of ancillary point. The generalized rule does not bring values to the table. It just creates the space for difference as scale increases. What is the nature of that difference? Depends on the system in question. I don’t think I’ve seen anyone generalize it or explain it fully.
Yes I saw that my own example includes some “cheating,” so tried to anticipate these valid criticisms. You are 100% right that arbitrary (or context-free) information isn’t really comparable to meaningful or non-arbitrary information (communicated in context). Here the scale just offers an opportunity to do things in a different way
But even this imperfect example is enough to refute the original point I am trying to refute… which is the claim (made by some)that AI is not intelligent any more than a calculator is intelligent. I do not argue the opposite; only that the claim over-reaches. We cannot extend the one case to the other.
Folks have to remember that You are A.I. It is your data
Amazing stuff. A lot of Anderson's work reminds me of Weizenbaum's thoughts, another AI pioneer and published at a similar time.