Now I can see how this might be a problem for code libraries, infrastructure system etc but the only reason to worry about AGPL in an end user application context is the desire to repackage and sell commercially. Aka do the Amazon thing.
Prompt is here: https://news.ycombinator.com/item?id=37611571
--
More seriously and charitably though, affero is vaguely scary because there's a (perhaps misunderstood, but I can understand) feeling that if you cooperate with it in any way in the backend and then a value-added product of that interaction makes it to a user of yours then you're up for copyright infringement.
This isn't a reason not to release under affero, nor is it a bad thing that entities producing non-free software need to take care what they leverage, but I can see how the decision could be made to avoid using agpl style software whereever possible to err on the side of caution.
A good example is ghostscript. At what point does an app converting PDFs to images (to display thumbnails, for example) as a small part of its workflow become an aggregate of that free software? According to the faq it ends up being a fairly subjective question of form and intent. Syscalls or static linking? Etc.
What is often misunderstood about Affero?
As I could not reply to your other thread as you were banned
The best I can do is direct people to the GNU FAQs on the topic