Many people used to use "the subtlety and rich contextual knowledge that humans bring" as an argument for why a computer would never beat the human chess champion, and then after a computer beat the human chess champion, they persisted in using it to argue that a computer would never beat the human go champion.
Yup, from the perspective of black-box functionalism, it looks like the anti-reductionist camp hasn't exactly come to terms with what is subtlety and what is for lack of a better term is anthropo-essentialism.
What I'm referring to, of course, is the "obvious"/"common sense"/"reality of" human thought having a je ne sais quoi to it that isn't present in AI. This is exemplified in "yes, but not X!" response to the encroachment of super-human performance in domain after domain, or the intangible emotional response difference between AI-created works opposed to similar human-created works. To a functionalist, it appears as an unaddressed framing problem.
There are certainly quantitative differences between AI and human behavior, but the gap finally appears to be accelerating in its closing rather than simply decrementing.
