Open AI releases the solutions to over 400 "big" open problems on GitHub
So, apparently most of these problems are really big problems in their subfields at the least. A lot of landmark breakthroughs in one day. The biggest one is apparently the Quasi Riemann Hypothesis problem. Some other examples are - Hilbert's tenth problem over Q, Hodge Conjecture for abelian varieties, algorithmic breakthrough for matrix multiplication, Spacetime Penrose inequalities from relativity. With this rate, it seems like we'll achieve super intelligence in math very soon just like chess. This also has a bunch of big problems from Theoretical Physics. Crazy times.
Pure math as a domain will probably turn into a bit like literature departments where ppl are just having fun exploring the beauty of math, writing expositions and not really "researching" in the sense it's done now. Jobs adjacent to math say financial modelling, cryptography, applied maths might almost completely vanish.