Exactly this.

Read about the Hawthorne effect: https://en.wikipedia.org/wiki/Hawthorne_effect and how people change their behaviors when they are being observed. Being free of observation is vital to being able to think your own thoughts without outside influence.

If the problem is with the usage framing the hypothetical adversary as a country's law enforcement, pretend you live in a cyber North Korea and have a cellphone. The idea of an adversary is just a means of thinking about the problem. You want to build a system so that they can't capture your flag (your flag being some digital information that you want private).

Whatever the opposite of TL;DR is:

It isn't intended to be read as 'do this to avoid law enforcement so you can do crimes'.

When thinking about security/privacy (same thing), you don't know what kind of attack you may eventually have to defend against, maybe you have a partner who has decided to stalk you and so they screen read your PIN or a strong arm robbery where they'll try to use your phone to access your bank.

Instead of trying to imagine every single possible scenario, you imagine one model scenario. In this model scenario, the adversary has every possible capability that is available and your goal is to keep your flag safe, or be able to pass a flag between two people without it being seen, or various other scenarios (which are themselves just model problems of types of system that you need to secure).

This hypothetical adversary, in order to have these capabilities in real life, would be the equivalent of a sovereign nation with unlimited funding and access to all technologies that are possible (and some that are only hypothetical). This description fits one country pretty well and so, as shorthand, people often just write 'the feds'. I guess they could also write 'Eve' but that is a specific adversary in one kind of scenario and not the general Adversary.