The short idea: Mathematics will split into two somewhat different domains: applied math will become the fabric of all engineering and sciences, where theories and algorithms will be built and verified automatically and on-the-fly. It will become a fundamental infrastructure. Pure math will transform into "high math," addressing the most essential questions of all time, such as P != NP. This will focus on a few central, high-level questions and will be driven mostly by human interests, requiring huge resources and centralized investments. Physics is a good model: we have applied physics with many practically relevant questions and tasks, while the part of physics concerned with the deepest questions of reality - particle physics and astrophysics - has little economic relevance but requires huge investments.