I agree, I find algorithms daunting, I don't inherently dislike them. I recognize the potential of recommendation algorithms as revolutionary, as long as they are open-source and people have the freedom to choose.

I think that their (FAANG Rec engines) downfall is that they were designed not for personal value or rich communication - but solely for attention and advertising.

Algorithms are just simple transformations applied to a group of things. Not evil, or wrong (or right) - just instructional workers to do a predefined computation.

The main risk I see is self bubbling accidentally. And it not being noticeable as obvious. Like I may accidentally filter out all crime, and then not realise my street has had several recent break-ins and I should be extra careful. Or I could filter out all trump content and wake up with him as president again without knowing there even was an election.