Rage farms

Read time: 6 mins

Hi Reader,

It’s been a couple of weeks now since I released my ‘Digital Habits Survey’ of the local area. A 10-minute read with stats, graphs, and references isn’t exactly a formula for virality, but that wasn’t the game plan.

Even so, the reception has been strong. It has reached the inboxes and desktops of teachers at every level, from nursery to secondary and university, and the local council, including the education councillor and the mayor. I was really pleased and surprised to hear from the addictions unit at the local hospital too. There are a few talks lined up at schools in the coming weeks.

So, in terms of strategy, it’s a good start. It’s led to quality conversations, with a bit of social media noise to complement.

In the last few weeks, I’ve also seen political content intensify. Not a coincidence, with snap elections called in Spain, midterms approaching in the USA, and protests and strikes in France and Spain.

It’s made me think about how many people read likes and views as support. Often they’re something else entirely. A lot of it comes down to how political messages are put together, and how social media algorithms are coded to make money.

In this email, I want to cover some basics, now that this has become an ever-present topic of conversation.

From the experts

Frances Haugen, former Facebook insider. She worked on Facebook’s civic misinformation team before leaking its internal documents. This line is from her evidence to UK MPs in October 2021. Zuckerberg rejected her claims at the time.

Liliana Arroyo, sociologist. She said this to the Ara in June 2017, when she taught at UB. She now directs Esade’s chair on socially responsible digital innovation

Steve Rathje, Jay Van Bavel & Sander van der Linden, PNAS (2021). They analysed 2.7 million posts. Posts about the political opposition were shared roughly twice as often as posts about the poster’s own side, and the effect was the same for left and right.

There is a digital asymmetry in what appears in our feeds.

Social media algorithms have one primary goal: maximise watch time and engagement to sell ads and satisfy stakeholders. The quickest way to keep a human brain glued to a screen is by triggering “high-arousal” emotions. In this case, moral outrage, urgency, and fear.

Three things are worth understanding about how that plays out in politics.

1. Who's making it

Some of the angriest political content in your feed is made by people with no stake in the outcome. Content farms, many of them overseas, produce it in bulk because outrage gets views, and views pay. AI has pushed the cost close to zero. Christophe’s video, linked below, follows the money.

2. The "Outrage Symbiosis" (Two-Way Engagement)

Algorithms cannot read human intent; they only see data spikes. When a political figure posts a controversial take, it triggers a massive wave of engagement from both sides:

• Supporters comment, share, and like.
• Opponents flock to the comment section to argue, quote-tweet the post with angry responses, or just watch the clip out of hate.

To the algorithm, a view is a view and a comment is a comment.

It reads this ideological warfare as “massively engaging content” and pushes the post out to neutral users. Ironically, users often act as a primary amplifier for the opposite side’s content simply by trying to fight it online.

So a big number under a post tells you very little about how many people agree with it.

3. Hyper-Integrated Alternative Ecosystems

Over the last 15 years, anyone who can get attention online has stopped needing mainstream media’s approval. Think of the most controversial figures of recent years: how rarely you see them write an opinion article, and how often you see them in a talking-head video.

The left has had its own historic moments online, usually driven by high-emotion, decentralised movements like #BlackLivesMatter or reproductive rights campaigns. Day to day, though, the numbers lean one way. Sitra, a Finnish research foundation, found most political content in young Europeans’ feeds was right-leaning. A large experiment run inside Twitter found right-wing parties getting more reach in six of the seven countries studied. Research linked below.

Neither study could say exactly why. Sitra’s researchers couldn’t rule out that it comes down to what gets produced in the first place, which brings us back to the content farms. What I’m sure of is that the digital colosseum rewards speed, conflict and simplicity, and whoever brings those gets the crowd.

We saw a version of this in Manresa these past few weeks. The first responses online to the devastating news in early September were videos of people talking to camera. The written proposals came a week later, by which time the feed had moved on.

When Sitra’s researchers read 1,719 political posts shown to young users in three European countries, two-thirds were opinion or entertainment. Only a quarter contained anything you could check. If you want to understand what’s happening, your feed is the wrong place to start.

There are ways to get involved without feeding the comments sections. Read the full proposal yourself. Go to a public meeting. Support local journalism.

Escape the algorithm

​

VIDEO: Who's behind the angriest content on social media? Content farms around the world profit from rage, misinformation, and AI slop. Christophe does an amazing job in finding out and explaining this in this video. Overseas content production hubs, the economics of attention, why anti-immigration content floods feeds, and demonstrates the low barrier to mass propoganda. In a 21 minute video. Videos like this are why I found YouTube fascinating.

FULL REPORT: Algorithms and democracy How social media shapes young Europeans’ worldviews.

RESEARCH: Algorithmic amplification of politics on Twitter - Based on a massive-scale experiment involving millions of Twitter/X users, analysis of political parties in seven countries, and 6.2 million news articles shared in the United States, this study carries out the most comprehensive audit of an algorithmic recommender system and its effects on political content.

If you got this far,

thank you for your attention, Reader. I know there are millions of other things you could be looking at right now.

If a section has particularly interested you, click the link of the corresponding section. It helps me prepare future newsletters.

If this was forwarded to you, you can subscribe here.

See you soon.

​Peace,

Has

​​​