my second year calculus teacher would be heartbroken, if he didn't already die of a heart attack. i mean no disrespect to dr bose or his family and that was a long time ago. it's a good line. admit it.
like navier-stokes, this is again the kind of thing that you can solve with brute force with enough computing power and getting a counter-example is probably more accurately described as a reflection of the parallel computing power of the ai systems as functional mainframe super computers (something that is perhaps being downplayed) than it is described as advanced intelligence.
they really could have just written a simple program in basic with a few counting loops, fed it into the mainframe and waited for it to halt. doesn't halt? conjecture is true. halts? conjecture is false, print n1, n2...nk. the only reason you could do this now and couldn't do it 50 years ago is the difference in raw computing power. it's no advancement in the algorithm.
they'll have to rewrite a lot of calculus textbooks.