New on Substack: How to Beat the Algorithmic Fascist Beast
Rupert Lowe posted one word and got sixty million views.
Here's why that number is a lie, and what you can do about it:
X's algorithm does not read sentiment. It reads engagement.
Your angry reply is algorithmically identical to a supportive one. Every dunk, every quote-tweet, every "can you believe this" is rocket fuel for the content you oppose.
Peer-reviewed research in PNAS, ACM FAccT, and a MDPI systematic review of 30 studies all confirm X's algorithm amplifies right-wing content by default. The PNAS study found this replicated across seven countries.
A pre-registered experiment (Milli et al., 2025, PNAS Nexus) found the algorithm amplifies "emotionally charged, out-group hostile content" that users themselves say they don't want to see. The algorithm overrides your preferences.
Analysis of the 2026 open-sourced code found that mass blocking acts as a negative quality signal: "If many people block you, the algorithm treats it as a quality signal and reduces your distribution more broadly".
Based on these documented mechanics, mass blocking would collapse engagement rates, degrade algorithmic quality scores, and sever the pipeline that pushes far-right content into centrist feeds.
This applies to the whole network: Lowe, Farage, Reform, Restore Britain, Tate, Robinson, GB News, and every mid-tier amplification node in between.
The far right needs your outrage. Without it, sixty million becomes a fraction. The algorithm only runs if you feed it.
Block. Mute. Don't engage. Starve the machine. And more importantly, take action in the real world.