Drive-by contributions of all kinds suck for the most part, don't contribute if you're not going to do it properly. At some point there's got to be more than one person able to review PRs (if the project scope mandates it).

Linus Torvalds doesn't review all changes he merges, how do you become a comaintainer if not assisting with maintenance?

If you can't review, don't make a PR (or something).