Twitter’s “Left” Discourse

What we call ‘left discourse’ on Twitter is the algorithmic result of the harvesting of data entered by leftists on Twitter. That data is not neutral, and the users are not neutral either. The more users use Twitter, the more the feed they get is curated to suit their own biases and algorithmically-perceived expectations.

The more you do it, the better the algorithm gets. You are its resource, quite literally. So if you start ranting about capitalism or imperialism, you’re more likely to find people who are doing so but just a tad more extreme, and the more likely you are to engage with them. And at the same time, the more active you are, the more likely you are to meet extremists, by your new standards, like tankies and so on. To tankies, this will look almost like the literal opposite. The more frequently they ‘engage’ on Twitter, the more likely they are to meet people most likely to keep them on Twitter; in other words: anti-tankies, because tankies and anti-tankies are likely to clash, feeding more data into the algorithm by doing so, and therefore more likely to do it again the next day.

It’s not very good for engagement to find other users who are only moderately different than you and with whom you already agree with most of the time. At some point, you stop engaging as frequently because there’s no need for it. There has to be something that keeps you on the site just a bit longer than you would have otherwise. So what happens then is that you’re presented with options most likely to get your attention, and that tends to be things that piss you off. These can be real news or opinions or completely fabricated ones. The algorithm can’t tell the difference and, unless you’re extremely careful, neither can a lot of people. It’s why clickbait-y headlines work, why people don’t usually actually read most articles they share on Twitter, why nuance is drowned out.

And the more you do that, or the more your followings do that, the more you’re also likely to see quote-tweets in which the ‘natural’ response for you would be outrage. It’s why I’m much more likely to see tankies even though they’re a small percentage of any population, and why their impact is so disproportionately toxic. It is why ‘the discourse’ on Syria is so rabidly irrational and fascistic regardless of the facts on the ground. When the facts are entered on Twitter that data has to compete for attention, it has to generate more ‘engagement’, it has to get them clicks. And stories of people slowly building alternatives to authoritarianism, or resisting fascism for years and years cannot compete with the shit-posting and the trolling and the disinformation and the extremists. It also works both ways: that same algorithm is more likely to put tankies in touch with one another where group dynamics take over and the loudest and most extreme takes drown out any nuance. If a would-be tankie sees some takes by already-convinced tankies and engages in it, they are more likely to see more of them. Eventually, with time, they no longer see nuanced takes. It is how, to these people, I can turn from ‘boring liberal’ (I’m not a liberal) to ‘agent of empire’ in a matter of days. They may even agree with 90% of what I’m writing or saying, but recognizing so would be antithetical to the premise of ‘more engagement = good’, a premise that no Twitter user ever chose in the first place.

Simply put, you can only engage in genuinely authentic discourse for a limited amount of time, but it cannot last. The algorithm is gathering data and catering to your biases. It isn’t long before your own mental independence is taken over by ‘the discourse’. This has gotten so advanced that, in certain cases, the algorithm can even predict which Twitter fights are more likely, and it then simply puts the users most likely to ‘engage’ (argue, insult, troll etc) in it in touch with one another. It’s the only way to maintain ‘engagement’. There’s only so many anti-authoritarians you can engage with, at some point you just don’t engage with them as frequently. This signals to the algorithm that something else must be tried to generate engagement, and that usually means something you hate as that’s most likely to get you going. So when I used to say ‘I just went on a Twitter rant’ – well, a lot of that was the result of curated content placed on my feed by that very same algorithm. This is manufactured outrage, not by any evil genius behind the curtains, but simply by an algorithm following the designed rules of the game. And speaking of genius, respected public figures who are Twitter addicts have muddied the water by making that kind of online addiction more ‘respectable’, even benign-looking. They’ve contributed to the Network Effect of making it ever-more-difficult for intellectual conversations to happen anywhere else, contributing to a problem that’s been worsening for years.

This is why it is so difficult for Twitter to remove ‘inauthentic activity’ (bots, fake accounts etc). It looks so much like the ‘authentic’ one to the point where the line between the two is blurred. How many tweets did you quote-retweet or reply to, or both, because you thought they were more popular than they actually were? Can you know when a tweet got 2,000 real retweets and when 1,500 of those retweets are bots? Let us be more honest and also ask, does it matter? Part of the goal of ‘engagement’ is to signal to your followers your ‘take’. You have to tell your followers that you know that ‘this is a bad take’, that ‘this ain’t it chief’. That look-at-me-I-am-smarter-than-this-idiot take is a performance, and we need to be honest about that.

So here’s my ‘take’. There’s no such thing as ‘left’ discourse on Twitter, not at scale in any case. There are obviously many leftists who tweet at one another, and it is not always a toxic situation. It can even be wholesome and very informative, but ‘the discourse’ is an algorithmic illusion fed by our own activity. If we can’t even tell the difference and continue to call this ‘discourse’ then there is no objective case to be made regarding any ‘leftism’ that is even remotely coherent enough to even be called that.

What’s extraordinary is that my positionality gave me an accidental insight into Twitter’s algorithm. It’s been blatantly obvious to me for years that most of the people engaging with one another so actively on Twitter – and this includes me, until very recently – are more alike than they would ever want to admit, that there are more similarities between the tankies and the anti-tankies, between the trolls and the anti-trolls, between the liberals and the conservatives than anyone would ever wish to admit. And because I knew that I’m myself, given the circumstances, a would-be tankie that I reacted so intensely to tankies, and why friends who have their own ‘Twitter obsessions’ (the perfect situation, ‘engagement’-wise) were so perplexed by why I saw these tankies as such a threat. This does not in any diminish the very real damage done by tankies, and in particular the inevitable cynicism of that form of politics. All I’m saying is that they’re not the only ones doing the cynicism bit.

Algorithms can be changed, business models can be changed. There is nothing about today’s Twitter that is inevitable. If, like me, you see Twitter’s potential then we must recognize that it must change. This will require significantly more effort than what is currently being done, so you need to ask yourself, as I’m trying to do, whether you want to spend that effort fixing a corporation that has the resources to fix itself or whether you want to be part of building the alternatives. The internet is an experiment and manipulative algorithms are not the only way of existing online. If we don’t at least work under that assumption we would have given in to a deep cynicism of staggering proportions.

PS: In case you’re wondering why I didn’t bring up Facebook, it is just because the Twitter users I know tend to believe that they’re not as impacted by the toxicity of the internet because Facebook is worse – and it is, so much worse. I just wanted to address that. When it comes to Facebook, I honestly usually recommend people just deleting their accounts there. It is an extremely unethical company and has arguably become an existential risk.