WORDS.

BLOGS, ARTICLES, AND POSSIBLY MORE.

02 OCTOBER 2023

ALGORITHMS: WHY I'M FINALLY GIVING UP SOCIAL MEDIA

Beep boop. Would you like to see a hate crime? No? How about a different hate crime?


A photo of three people wearing hoodies and Guy Fawkes masks in a room with multiple computer monitors.
A group of archetypal internet trolls (or hackers, they're apparently interchangeable in the stock photo universe). Photo by Tima Miroshnichenko on Pexels.

Just a quick warning: While I’m not going to say anything groundbreaking, I’ll be discussing some unpleasant topics, including racism, fascism, and other -isms that are popular amongst the dregs of society. If that’s not your cup of tea, avert your eyes and perhaps look at something a bit nicer.

Every so often, a popular media outlet such as the BBC will publish an exposé on one of the popular social media platforms. They’ll do something along the lines of creating an account pretending to be a 12-year-old, then see how quickly they’re shown posts promoting male supremacy or eating disorders, and collate all their findings and show off what they’ve found, to the surprise and shock of absolutely nobody who has ever used any social media platform, which is basically everyone. Following this, they’ll pat themselves on the back for a fantastic bit of reporting and repeat the same process in [roll dice] months time. Meanwhile, big cheese representatives from the offending social media will face a “grilling” from Parliament, which leaves us all begging MP Darren Jones to never cook chicken.

It goes without saying that nothing comes from any of this, and nothing is ever done. Why would it? After all, nobody would make any money from changing things. It would be silly of me to think writing about this could change anything, but I’m going to power through regardless because a lot of my thoughts about it all have been eating away at me, and, well, it’s my website.

I’ve used almost every social media platform out there in one way or another, starting with Facebook way back in 2008. Here’s something to get nostalgic about: remember logging on to Facebook or Twitter and only seeing posts from people and pages you actually follow? What a crazy time! It’s almost as if these platforms became so huge in part because of this specific feature. Fast forward 15 years to when the X (also known as X, formerly known as Twitter) rival Threads was released. At launch, the only feed you could view showed you “suggested posts” from people trying far too hard to be funny, with a ratio of about 30 suggested posts to 1 post from someone you followed. Perhaps that’s why it died a death about five days later.

I’ve never met a single person who preferred an algorithmic feed to the old chronological feed of accounts you choose to follow, which, to my knowledge, doesn’t exist on a single platform anymore. Instead, we’re presented with abominations like TikTok, where they take the phrase “throw shit at a wall and see what sticks” very seriously. They’ll base what you are initially shown on your age and gender; for example, if you’re a man, you’ll get shown football clips, because apparently, those bizarre and nonsensical gender stereotypes are still alive and well in the year of Our Lord two thousand and twenty-three.

This is where it gets ridiculous. If you don’t like what you’re being shown, you can click a “not interested” button, which TikTok assumes you simply misclicked unless you hit the same button on the next 100 videos of the same topic, at which point the algorithm will, if you’re lucky, stop showing you that topic as often. If you watch a video for more than a second before continuing, TikTok assumes you like that video and subsequently shows you more of it. If you press the like button, that acts as a sort of super-like, and you’ll be shown much more of that topic. If you ‘save’ a video, that topic is pretty much all you’ll be shown until you do the same with videos of other topics.


A screenshot of a TikTok video by a user called Podcast Clips. The top half of the video is a man from a podcast, the bottom half of the video is footage from Grand Theft Auto 5. The video caption reads: "Do tou agree? #whateverpodcast"
“Do tou agree?” I don’t care! I don’t care about this!

What content does the algorithm throw at you? Popular videos, of course. How does it measure popularity? Not through views, but through “engagement”. Every time someone likes, comments on, saves, or sends their friend a video, that counts as engagement. These metrics are all weighted differently, with the number of comments being one of the bigger deciding factors. Therefore, videos with more comments are far more likely to be shown to more people, which results in more comments, and it all snowballs out of control. I should know; when one of my TikToks got a comment, it always resulted in a boost of views, which bred more comments, and soon enough, there was a video of me playing a Disney game with millions of views and thousands of comments.

“That’s fantastic!” you might say, and yes, that worked out well for me personally. But you should take into account that the way this algorithm works is no secret. People can easily game the system to become more popular. All you have to do is produce content that gets people talking. So, what gets people talking the most?

Controversy!

I understand the scope of “controversy” can range from innocuous things like an American saying “Fish & chips sounds disgusting” to something far more serious, but generally, if you post something controversial, you’re guaranteed to get comments disagreeing with it, alongside others agreeing with whatever point the post is trying to make, who will often reply to the people complaining. This usually ends up with comment threads made up of dozens of replies full of insults and general vitriol being thrown about the place. There’s also a chance the video is fairly neutral in and of itself but is practically gagging for comments from those exposed to high levels of lead as children, ranging from not-so-subtle coded language to remarks which, if said by a talking dog, would result in it being put down.

Now, I can’t bear the thought of that guff going unchallenged, especially when these people are so brazen in their comments, so I’m definitely guilty of getting into these slanging matches. I remember when I was an impressionable 16 year-old, I could have easily been influenced by some of these videos and comments, and I very nearly went down the dark path of radicalisation that so many young people have fallen victim to in the past few years. I do try my hardest not to just throw insults, but to offer decent and well-thought-out arguments that would have swayed a younger me all those years ago. The problem is sometimes, comments are so vile, hypocritical, or plain nonsensical, that there’s really nothing you can say to change their minds, and I always regret just insulting someone, but I’ll get so mad that I feel almost compelled to do so.

Of course, the vast majority of TikTok’s moderation is automated, with another glorious algorithm that is so easy to trick, you can simply l33t speak or use emojis and not get caught. I’ve seen neo-Nazis get off scot-free because they’ll use two lightning bolt emojis instead of typing “SS”. You can have a swastika symbol (I’m talking about the literal Nazi flag, for any readers well-versed in making bad faith arguments) as your profile picture, and chances are you won’t get penalised. The algorithm never takes context into account either, so a comment that, in a vacuum, isn’t offensive (take an all-too-common example of “black people”), under a video that makes the comment extremely racist in context, won’t get deleted. These examples I'm giving are merely the tip of the iceberg; I genuinely don’t feel comfortable putting verbatim (or even euphemistic) instances of horrific comments I’ve seen on my website. I'm sure many of you reading this will know what I'm talking about because it's often just a click or two away.

I only report comments and posts that I’m adamant breach the terms of service, primarily because the process is such a pain to go through, but only around 10% of the comments I report are eventually taken down. Meanwhile, I’ve had comments removed by the algorithm mere seconds after they were posted because I dared to call someone a “clown”.


A screenshot of a TikTok page. It reads: "Comment removed. This comment violates our Community Guidelines. Comment details: "you have pronouns in your bio you clown" Posted on 09-29 at 20:44. Violation reasons: Community Guidelines. Your content is against our Community Guidelines. The guidelines include rules and standards for using TikTok that apply to everyone and everything on our platform. See C... See more"
TikTok doesn’t like you pointing out the stupidity of people who “hate pronouns users”.

So I gave up on TikTok, deleted the app, and spent more time on Twitter, until I realised that platform works in exactly the same way. Endless tweets from people you don’t care about who, for whatever reason, bothered to pay $8 a month, are constantly shoved in your face, and an entire section dedicated to trending topics which are almost always controversial, because again, controversy produces engagement.

Okay fine, let’s bin Twitter and try Instagram. However, with the sudden popularity of short-form videos, Mark Zuckerberg has entered maximum overdrive in his quest to turn everything he owns into a TikTok clone, meaning Instagram now focuses on “Reels”, which are mostly reposted TikToks. Again, these are shoved down your throat on the homepage, interspersed between posts from accounts you actually want to see.

Sure, there are a few platforms that may seem promising, but they have the same issues. I’ve already mentioned the issues plaguing Threads, and while there’s also the invite-only (as of publication) Bluesky, I'll be fairly bored deep in the cold, cold ground before I use another Jack Dorsey product (his brainchild trending topics arguably got the ball rolling for this whole mess). Hell, even Reddit has started paying users for “popular” posts, which will inevitably lead to people, in search of a quick buck, producing more of the same polarising posts that all the big platforms suffer from.

This obsession with “engagement” being used as a synonym for “popular” is exactly how we ended up with the likes of Laurence Fox and Andrew Tate in the limelight. How has nobody come up with a better way to measure popularity? Or have they, and they’re just being ignored because implementing the changes would take time and money that they’ll never recoup?

I will say, for the most part, deleting the apps from your phone and only using the desktop versions of the platform(s) of your choice is a happy medium for those unwilling to go cold turkey, as the more sinister and addictive elements of these platforms are somewhat nullified by the desktop experience. There are also various extensions available for download to help you out. This is the stage I’m currently at, but I still find myself being presented with things I never want to see on a regular basis, and for what? Who is benefitting from me seeing posts mocking and belittling the most marginalised groups in society? Certainly not me. I just want to see weird-looking dogs, ridiculous dinners, and whatever the hell this is.


An Instagram post by user epicjoko. A meme featuring Wallace and Gromit, with the caption "careful Gromit, there may be horny milfs in our area". The caption to the post reads: "OMG 🙈🙈🙈"
Imagine the look on the faces of edgelords when they see a post like this and realise you can be funny without using slurs.OMG 🙈🙈🙈” indeed.

Perhaps, if you have to spend the majority of your time on a platform finding ways to avoid certain content, it’s just not worth using the platform.