Is Super Optimization a Way to Improve Compiler Development?

The idea of super opnization as a way of helping people who are building the cor infera structure, seems to me very compelling. The otherthing you describe, i worry about just because in terms like saying, oh, you can have a cheap compiler for your domain specific language done this way, a cheap but very slow compilation. Maybe our's long optimization process of going and trying to find the actual best code sequenceshin compromises the practicality of the fruit. But then perhaps there is a way to idely, i don't call it partial valuation, but partially valuate that formula, or that constraint system to get more efficient compilation when you decide on the pair of sourceang and