At a glance, this paper seems extremely questionable. It appears to take an offhand comment by Kahneman in a Lex Friedman podcast regarding the limits of AI and interpret it as asking whether there is a problem that humans can solve but machines cannot -- which is definitely not a reasonable interpretation IMO as in context Kahneman is talking about the limitations of present AI systems, and computability is a totally different question. The paper's claim to have found such a class of problem would mean the Church-Turing thesis is false, and so it is quite surprising that the Church-Turing thesis and existing discussion around it does not appear to be mentioned or cited. Anyway, the paper constructs at some length a non-computable problem and then declares, "Nevertheless, a human can certainly solve this problem by creative work." While I haven't gone through the construction in-depth to see whether that's reasonable, this certainly doesn't come across as a very rigorous proof. Also, in the introduction, it makes the bold claim that "digital computers ... are able to compute with finite sequences of rational numbers only," (and suggests that analog and quantum computers lack such limitations) which is certainly false as can be seen by programs such as Mathematica that easily solve many systems involving irrational and transcendental numbers.