roytang.net Posts Photos Archives About

Jeremy Keith wrote about the cage of algorithmic recommendations:

But that’s not true. You don’t ask “the algorithm” for a recommendation—it foists them on you whether you want them or not. A more apt metaphor would be that you walked by a record shop once and the store worker came out and followed you down the street, into your home, and watched your every move for the rest of your life.

What Peter describes sounds great—a helpful knowledgable software agent that you ask for recommendations. But that’s not what “the algorithm” is. And that’s why it feels like a cage. That’s why it is a cage.

This got me thinking about my own social media use and how I use algorithmic recommendation engines:

  • Spotify: This is where the algorithm most affects me I think. I regularly use the Discover Weekly and Daily Mix, algorithmically generated playlists that depend on my previous music tastes. A caveat: my music tastes can be a bit monotonous and hasn't really changed during the years, so that might be why I enjoy these.
  • Youtube: Most of my Youtube use is intentional - I go there to look for a particular video. I do sometimes open the homepage and click one of the recommendations there, and I will sometimes leave a Youtube window on autoplay so it goes to the next recommended video. But almost all of my YT consumption is entertainment-focused - usually about videogames, movies, anime, that kind of thing. I don't watch political or opinion stuff, except for the occasional John Oliver episode, which seems mostly informative anyway. I don't think YT has ever recommended me any other shows like John Oliver for some reason.
  • Netflix: I rarely use the recommendations, usually when I open Netflix it's specifically for watching something specific. I do on occasion browse through their "for you" lists, but rarely play anything from them.
  • Facebook: I don't use Facebook, and neither should you.
  • Twitter: I do use Twitter actively, but my Twitter experience is strictly curated without algorithmic recommendations; I don't see other people's likes and the timeline I scroll through is always strictly chronological.

I think the key to healthy consumption of algorithmically-recommended content is to restrict it strictly to entertainment purposes. Avoid political or opinionated content, since that will likely send your algorithmic recommendations down an echo chamber from which you may never recover.


Related: Matt Ruby's newsletter issue "When there’s shelf space in wrong, someone’s gonna fill it" talks about how the platforms are incentivized to promote misinformation:

It’s about incentives. These platforms are designed to create tribal echo chambers where pre-existing opinions are validated. That’s the best way for them to deliver us our dopamine hits in order to maximize attention and profits. In turn, creators have realized being wrong is an easier path to money/fame/followers than being right. After all, disinformation on Facebook gets six times more likes and shares than factual posts.


Recommendation engines are really only useful when you're browsing the internet in passive consumption mode, basically the modern equivalent of channel-surfing back in the day. It's when you're basically just scrolling through a feed to pass time in some way.

The antidote to passive consumption is active consumption of curated content, curated in the sense that you purposely choose the content you consume instead of mindlessly scrolling through a feed determined by some algorithm. This generally means seeking out specific content sources and browsing them in a content-neutral manner (typically chronological). RSS and newsletters are typically good sources that don't rely on algorithmic recommendations.


Related: Robin Rendle, responding to Jeremy Keith's post, writes about how RSS is the promised land for this sort of thing:

One of the neat things about RSS I’ve noticed is the number of times I walk away from it feeling awe instead of anger (basically 100% of the time). I think that’s because there’s no money in it, there’s no incentive to profit off my rage.

Aside from browsing chronological feeds, it's also nice to browse the internet purposely - by seeking out specific content. For example, you can seek out content based on human-curated recommendations from friends and other trusted sources. Another thing I like to do is to visit and explore personal websites and keep following links that look interesting (especially if there's a blogroll! I love blogrolls! I should update mine.), basically relying on my own judgment instead of a recommendation engine.


Unfortunately, it's unlikely that active web-browsing becomes the dominant form of internet usage among the masses - for the simple reason that passive consumption is much, much easier as a time-waster. Which means of course that those people who prefer it will also prefer algorithmic recommendation engines because algorithms won't actually run out of content to promote to you. If you're waiting somewhere and just need to kill time, it takes zero effort to open your Facebook feed and keep scrolling through your feed, and Facebook will happily toss unlimited ads and memes and videos and whatnot there as long as you keep scrolling.

RSS and other active browsing methods might be the promised land for those of us who don't want to be caged in by algorithmic recommendation engines, but it's gonna be an uphill climb to get them into widespread use. I'll be happy when that day comes, but I'm not holding my breath.

Sept. 8, 2021, 2:07 p.m. / / blog / #tech-life / Syndicated: mastodon twitter / 895 words

Last modified at: Sept. 8, 2021, 3:52 p.m. Source file

Referenced by