There are many ways to run code at startup. cronjobs and systemd are common ways to handle this. I have also had things start automatically with my desktop environment which comes later in the boot process.

You may be accustomed to the process, but fixing issues in the registry is not intuitive. It is simple enough if you find a guide that tells you exactly which item you need to work on and exactly what the default is and what you need to change it to, but what if the guide isn’t exactly what you want?

In the GNU/Linux ecosystem, nearly every program has a config file. Sometimes each line has detailed comments in plain text around it you what the option does with examples of what it could be. If the documentation doesn’t exist, you can dig deeper and see what that option does in the source which is usually documented as well. Programming experience is not required to search for text and read comments. Such documentation is not equivalent in Windows.

Many apps that are on F-Droid are also on Google Play, but there are exceptions. Gadgetbridge (smart watch connection app) warns that the Google Play version is unofficial and unaffiliated. An unofficial app might be upstream compiled without modifications, but this opens users to potential supply chain attacks from the uploader and Google.

This is the first top-level comment that hints at the main criticism against systemd. systemd is increasingly difficult to replace as time goes on. I like and use systemd because it has a fast boot, but I wish the project was developed in a more modular way that had choice built-in. It is instead developed as the way that everyone should systemd instead of alternatives. This philosophy gets in the way of distributions that want to provide alternatives (Devuan, Gentoo, Parabola, etc.). Some of the sysadmins I work with closely use Devuan and follow development. I hear the patch set around bypassing systemd grows in size and complexity each year which is worrisome for choice.