Oh boy. Concrete thinking is a serious issue.
Don't submit a PR until you are ready for eyes on it. End of story. That is literally the entire point of a PR.
However, don't make damm mess with team process. Slightly odd process (and this is a small thing) that is followed is much better then situation where everyone follows different process.
This is level of conformity that is absolutely healthy.
I've been asked to review PRs that totally missed the mark, only to find out that there were already a dozen more branches relying on that PR's approval. I'm sure the context is different here, but that's an easy way to set yourself up for redoing weeks of work if you didn't get it right the first time.
One very common thing that I've seen is opening a PR just to double-check the way the PR looks on github. Ideally they mark these PRs with "WIP" either in the title or as a tag. It's implied that those PRs aren't ready for review.
If you weren't expecting critiques of your code while you were in the middle of developing it (perhaps you have placeholder code all over the place), it could be jarring and distracting.
This goes against the purpose of PRs, but at some point you have to put engineering culture flexibility above strict letter-of-the-law guidelines.
If someone didn't pick up on this phenomenon, I'd gently inform them. There would be no need for getting upset... unless they continued doing it in spite of the commonly agreed-upon behavior of the team.