It's reasonable to say that current or near future AI can find novel mathematical results and develop applications from those. The idea of AI stepping from a g

It's reasonable to say that current or near future AI can find novel mathematical results and develop applications from those. The idea of AI stepping from a graph theory/combinatorics innovation to some new and useful algorithm isn't crazy.