If you can drivechain a BitVM, you'd greatly reduce how costly and slow it will end up being.

If I understand it, BitVM is too slow and data inefficient as-is for any application other than as a means of enforcing a very simple contract that is just outside the current means of Bitcoin (ie: simple but needing Turing completeness). Higher-level commits, especially 32-bit values, would greatly expand BitVM's possible applications.

As of now, building logic circuits on Bitcoin is cool, but it is only a start with very little application.