Over reliance on algorithms has degraded the user experience to the point that the average user is drowning in ragebait and extremist politics, because they drive up engagement. Just like a toddler, algorithms don’t discriminate between good and bad attention, so everything that gets clicks is thrust forward. Now, you could hope to train the algorithm to show you only postive things, but engagement is engagement and the algorithm curators often engage in rage farming, where your feed is injected with things that are likely to enrage you.
You can avoid this by installing an RSS reader, going to your favorite sites, and manually adding a RSS feed. Now, your reader has things that you manually selected, with the added bonus of having a content pipe free of malicious interference. You can also divide topics in a way that you can avoid certain themes and news until you decide to engage them.
I have a list of all my subreddits as RSS feeds. Did some text transforming to add the RSS links and used an OPML generator to make the file. I was not adding 600+ subreddits individually, lol.
I’ve collected loads of RSS feeds, from Congress (bills and other happenings, etc.) to the NY Times, and Science Daily with their topic feeds. GitHub has a few OPML files, though some of the feeds are out of date. Tumblr used to have a way to export your followed blogs as OPML, but that broke at some point in 2020. Mastodon has RSS feeds for every profile, but it’s a pain to collect, as their CSV export outputs the address in the wrong format for the feed.
I use Feedbro on Firefox, and QuiteRSS & RSS Owlnix on my Windows desktop. I also use Podcast Addict on my phone for my podcasts and keep a copy of the OPML file in my RSS reader as backup.