There are papers and Postgres projects that attempt this kind of learning-based optimization, with some success. None are in widespread use. (One part, but certainly not the entirety, of the problem is that it's not just A/B, it's an exponential number of options that all could seem close to each other.)

You can and some databases can do this (e.g. Oracle).

Customers love it when their queries sometimes run a lot longer.