For me the really unreasonable change was the app testing requirements on non-corporate developers. Having to get 20 users to beta test an unlisted Android app for two weeks before getting it on the store is not a reasonable thing to require for hobby projects. I'm not sure I even know 20 Android users well enough that I'd feel comfortable asking for that level of engagement from them.
It's a particularly bad policy to launch with existing developers grandfathered out, because the policy probably looks really successful to start with due to the difference in new developer vs. old developer populations -- the entities who are right now making most of the quality apps aren't affected. What's being affected is the pipeline of new developers, but the effect of killing that pipeline won't become obvious for years.
When making LearnTheWords[1] I had to wait 6 weeks for approval from Google. They weren't happy with how I documented what this testing process was like. I had to wait 3 rounds of 2 weeks between submissions, writing ever more kafkaesque descriptions of the insights i gleaned from the definitely-not-paid-for test users that i'd required.
I wasn't expecting it to be easier to launch on iOS than Android, but here we are.
It's crazy to have to pay someone to do this, particularly because Google don't want you to use paid testers.
The idea was to stop spammy apps, i believe. But they've really thrown the baby out with the bathwater here, making it really hard for small-scale innovation.
It's a particularly bad policy to launch with existing developers grandfathered out, because the policy probably looks really successful to start with due to the difference in new developer vs. old developer populations -- the entities who are right now making most of the quality apps aren't affected. What's being affected is the pipeline of new developers, but the effect of killing that pipeline won't become obvious for years.