← All links
Theme
Using automatic theme
Back to top

Article notes

The Closure of the Internet

By Arctotherium

Read the original

Listen

Audio version

A direct reading of the notes, with clickable timestamps throughout the article.

Total length: 11:07
11:07 remaining

In one sentence

Arctotherium argues that, beginning around 2015, major platforms and a network of governments, media organizations, NGOs, fact-checkers, academics, and intelligence-linked actors transformed the open Internet into a centralized system of ideological moderation, algorithmic suppression, deplatforming, and infrastructure control.

Overview

The essay defines two phases: 2015–2019, when the technical and conceptual infrastructure of moderation was built amid the migrant crisis, Brexit, Trump’s election, and the rise of the alt-right; and 2020–2022, when those systems intensified during COVID and the 2020 election. Its central narrative is one of rapid escalation: vague categories such as “hate,” “toxicity,” “misinformation,” “malinformation,” and “online safety” replaced narrower rules against threats, illegal activity, or harassment extending offline. The author then surveys YouTube, Reddit, Facebook, Twitter, Amazon, Wikipedia, Google Search, cloud providers, payment processors, open-source projects, forums, and comments sections. The essay concludes that 2022 brought only a partial thaw, not restoration, and that LLMs may reproduce and centralize the worldview formed during the closed period.

Core ideas

“Closure” is broader than ordinary censorship

The author uses “closure” to describe a system-wide change in the default conditions of participation: less open access, more opaque rules, algorithmic downranking, demonetization, deplatforming, doxxing risk, and dependence on centralized chokepoints. The important change is not only that particular posts disappeared, but that users learned to avoid certain topics and associations.

The escalation pattern

The essay’s slippery-slope model runs from clearly extreme targets to increasingly mainstream subjects. It begins with neo-Nazi sites and violent or plainly offensive material, then moves toward immigration debates, anti-feminism, transgender skepticism, COVID origins, lockdown opposition, vaccine skepticism, and mainstream Republican figures. The author treats the broadening of categories as evidence that the original safeguards lacked a limiting principle.

A cross-platform institutional shift

The same transition allegedly occurred across otherwise different services. YouTube expanded hate-speech rules and borderline-content suppression; Reddit moved from a narrow threat-based standard to identity-based hate rules; Facebook adopted fact-checking and deboosting; Twitter created trust-and-safety structures, hidden blacklists, and expanded protected categories; Amazon used removals and “Do Not Promote” measures; Google Search privileged approved authorities; Wikipedia restricted sources and editors.

The para-state

The essay’s most important explanatory concept is the “para-state”: nominally private NGOs, media organizations, academics, fact-checkers, foundations, regulators, and intelligence-linked bodies that exercise state-like influence without being directly bound by constitutional speech protections. Platforms are portrayed as responding to government pressure, regulatory threats, advertiser campaigns, and NGO blacklists while retaining plausible deniability.

Chokepoints matter more than formal ownership

The usual private-company defense—platforms may choose what to host—is presented as inadequate because network effects created quasi-monopolies. App stores, cloud providers, domain registrars, payment processors, search engines, and social platforms collectively controlled access, distribution, revenue, and even the ability to operate a website. In this environment, “just build your own platform” is not a meaningful alternative.

Fact-checking cannot safely become a censorship authority

The essay distinguishes correcting claims in public from suppressing them through centralized fact-checkers. Even accurate organizations can make correlated errors because they share funders, source hierarchies, institutional assumptions, and certification networks. If their judgments determine visibility or access, mistakes cannot be corrected through open contestation—the normal mechanism by which public knowledge improves.

The chilling effect exceeds the visible purge

Bans and removals are only the most measurable part. The author argues that opaque enforcement, doxxing, loss of employment, debanking, and fear of guilt by association deterred people from researching, publishing, collaborating, or even publicly interacting with dissidents. This damaged the pipeline of new writers and experts more than the direct removal of prominent figures did.

Intellectual ecosystems became lower-quality and more isolated

Before closure, heterodox creators could interact with subject-matter experts, hobbyists, and adjacent communities. Network analysis and guilt-by-association discouraged such cross-pollination. The author believes this converted political dissent into an intellectual ghetto: more insulated from criticism, less technically competent, more meme-driven, and more vulnerable to grifters and extremism.

Practical takeaways

Caveats and counterpoints

Questions worth revisiting

Return to this when…

Return to this essay when examining the history of platform governance, the rise of “misinformation” and “online safety” frameworks, the relationship between private infrastructure and state power, or the epistemic risks of centralized AI assistants. It is especially useful as a map of the author’s argument and of incidents worth independently verifying—not as a neutral account of the period.

References

  1. Original The Closure of the Internet