First of all: I hereby apologize to orthonormal if s/he felt insulted or belittled or mocked by what I wrote. That was in no way my intention.

I'm not sure "sarcasm" is quite the right term, though. I really do think that, among human beings, doing mathematics well takes exceptional brainpower, and even most very intelligent people can't do it; I really do think that playing chess well has the same feature; and I really do think that that feature is why orthonormal expects doing mathematics well to be one of the last achievements of AI before general superhumanness.

If my last paragraph had said "I don't think that's very good evidence, though; you could say the same about playing chess well, and we all know what computers have done to that" then it would have been clearer and that might have been worth the loss of brevity. But I'm not sure sarcasm is the failure mode (if failure it be) here.