You are overestimating attention to detail of most devs
Most of these workflows appear to be much dumber than you think. There is a bunch of M1-native games on Steam that do not start because they are missing software dependencies. It seems that many devs simply have their "desktop" Unity project, click build and ship it off as it is. For newer Unity versions this will build both Intel and Apple Silicon code, but they apparently don't run any tests. So you end up with an Apple Silicon version that is nominally there but is utterly useless because third-party dependencies are Intel only. I run into this issue when trying to start 7 days to die: in the end one had to manually flag it to launch with Rosetta to work properly.
I'd... both agree and disagree. If a game's been plopped out onto Steam, then a dev would probably just hit the magic 'PORT TO MAC' button, slam it on Steam.
If a dev's taken time and attention to detail, then they probably would consider it and do one of the following:
a. Reject a Mac port. Maybe the game didn't do well. Maybe the dev team hates Tim Cook. Either way, won't happen.
b. Make a Mac port themselves. It could be customer demand. It could be an Apple engineer asking. Either way, the dev works on it and releases it.
c. Ask a third party to make a Mac port. Again, could be customer demand. Could be a porting company going "Hey, we did titles for you before, can we do this one?"
But if a dev releases a crap port, then that'll show in reviews. People'll see curators going "I BOUGHT THIS AND IT MAKES ET ON THE ATARI LOOK TRIPLE A!", and go "Yeah, not buying that."
...Or they'd buy it to see if it is actually that bad.