The difficulty I'm having with this article is that many people, when discussing intelligence, invoke some specific capability, X; for Deutsch, X seems to be "creativity", for Penrose, X may or may not be the ability to grok the truth or falsehood of an arbitrary statement as if one has the personal cellphone number of God. Then, they go on to assert "computers (or whatever) cannot (currently) do X, therefore we need to create a grand theory of X to create Artificial (General) Intelligence" (or "computers cannot do X, therefore A(G)I is doomed").
Unfortunately, they never seem to provide proof that X exists, or is true. Sure, you can provide a plethora of examples of creativity, but the plural of "anecdote" is not "data", the plural of "data" isn't really "proof", and I don't know what creativity is well enough to tell what I do commonly (or even occasionally) is creative and a forward-chaining formal logic thingy printing out a novel, true formula isn't creative.
Ultimately, that is why I like the Turing test---I don't have to understand all sorts of magical X's, all I have to do is to give you the benefit of the doubt if you don't have to have a bunch of truly arbitrary conditions on what you do.
>Unfortunately, they never seem to provide proof that X exists, or is true
I think we can forgive Deutsch for this, because he's claiming that creativity is part of an unsolved philosophical problem. Which means we don't know how to think about it yet. We'll know when when we do know how to think about it because there'll suddenly arise answerable questions, proofs, definitions, and whatnot.
Unfortunately the Turing Test can't cut through the philosophical wrangling because, for example, imitating a human being successfully is not the same thing as evidence of thinking. Without an explanatory theory we wouldn't even know how to interpret the relevant evidence
Unfortunately, they never seem to provide proof that X exists, or is true. Sure, you can provide a plethora of examples of creativity, but the plural of "anecdote" is not "data", the plural of "data" isn't really "proof", and I don't know what creativity is well enough to tell what I do commonly (or even occasionally) is creative and a forward-chaining formal logic thingy printing out a novel, true formula isn't creative.
Ultimately, that is why I like the Turing test---I don't have to understand all sorts of magical X's, all I have to do is to give you the benefit of the doubt if you don't have to have a bunch of truly arbitrary conditions on what you do.