Essay

Who gets to shape my attention?

Technofeudalism, advertising and the responsibility of the cloud

By Rishi

Editorial disclosure: A personal position developed from Rishi’s brief, drafted and source-checked with AI assistance. This is an essay, not original scientific research or an investigation of a particular abuse case.

Content note
Non-graphic discussion of unwanted sexual content and online exploitation.

I want to be able to spend time online without feeling that my attention has become somebody else’s territory. Advertising troubles me most when it tries to turn vulnerability into an opportunity: insecurity into a purchase, curiosity into repeated exposure, or a moment of distraction into another reason to stay.

That is the personal feeling behind this essay. It is a concern about the conditions under which we form desires and make choices. I want room to think before an interface decides what deserves my next reaction.

A useful question inside technofeudalism

Yanis Varoufakis uses technofeudalism to describe a proposed shift in economic power toward the owners of digital platforms and what he calls cloud capital. In his account, people and businesses depend on privately controlled digital spaces, while their activity helps enrich those who govern access. This is a political-economic argument, including the disputed proposition that capitalism has been superseded; I read it as a framework to examine, rather than an established description of every internet business. Varoufakis, 2024

The question it opens for me is who gets to set the terms of participation. If leaving a service means losing contact with a community, how much freedom does the exit button really offer? If reaching people requires adapting to a commercial ranking system, what happens to expression that is thoughtful but slow, difficult or unprofitable?

I do not need to settle the name of our entire economic system to take those questions seriously. My concern is concentrated power over the spaces where we meet, learn and speak. I want people to have meaningful choices about those spaces, including choices that do not require constant commercial performance.

What technofeudalism means to me here

I do not use technofeudalism as a slogan that turns every technology company into a villain. I use it as a question about asymmetry. A person may enter a platform to speak with friends, learn something or simply pass time, while the platform controls the ranking system, the commercial rules and much of the context in which that person is seen.

That asymmetry matters to me even if we continue to call the wider system capitalism. The same person can be a participant, a source of behavioural signals, an audience for advertising and a producer of the engagement that makes the platform valuable. My philosophical concern is not that influence exists; human beings always influence one another. It is that influence can become difficult to see, difficult to refuse and expensive to escape.

I therefore judge digital freedom by more than whether a button labelled “leave” exists. I care about whether refusal is meaningful while remaining connected to ordinary social life. I care about whether a person can understand why something is being shown, reduce unwanted targeting and continue participating without being repeatedly pushed back toward the same commercial objective.

What the evidence can say about cognition

There is evidence that the presentation of social advertising can change behavior. In two large field experiments, Bakshy and colleagues varied social cues attached to advertisements. Displaying a peer’s association with an advertised entity increased responses on average, with stronger effects associated with closer social ties. The measured outcomes included clicks and connections with the advertised entity. This supports a specific claim about social influence in advertising; it does not establish that every viewer is manipulated or that an advertisement causes lasting cognitive injury. Bakshy and colleagues, 2012

Attention also depends on interruption. A laboratory study by Upshaw and colleagues found slightly slower responses when a letter task was paired with smartphone notification sounds instead of control sounds. Error rates did not differ between the sound conditions. Some brain-response measures were consistent with greater recruitment of cognitive control, making the result more complicated than a simple story of a damaged attention span. The study tested notification sounds during a particular task, not advertisements or lifelong changes to the brain. Upshaw and colleagues, 2022

These findings give me reasons to care about the design of an interface while keeping the limits of the evidence visible. A change in clicking, a temporary distraction and a clinical disorder are different outcomes. Claims about permanent harm would need evidence directed at that question.

My ethical concern begins earlier than a diagnosis. I want advertising to be recognisable, dismissible and proportionate. I want meaningful control over targeting and a refusal that continues to matter after the next refresh. An invitation to consider something can respect a person’s agency. A system designed around wearing down that person’s refusal deserves scrutiny.

For me, the misuse of advertising includes disguising persuasion as an ordinary conversation, exploiting sensitive inferences and repeatedly pushing unwanted material into spaces where people did not choose to encounter it. These are the practices I want judged. I am not claiming that every platform, advertisement or user experience fits the same pattern.

Why does harmful content remain reachable?

The presence of unwanted explicit material raises a harder question for me. When large infrastructure companies make the internet work, what responsibility do they have for material that harms people?

Cloudflare is a useful example, but accuracy matters. Its published policy distinguishes its reverse-proxy and CDN services from products that host content. Seeing Cloudflare in a site’s delivery path does not, by itself, identify the server holding the original files. Ending a CDN service may therefore leave those files online elsewhere. Cloudflare says it forwards relevant abuse reports to site operators and hosting providers, while content hosted through its own products can be removed or disabled under applicable policies. Its account also emphasises proportionality and freedom of expression. Those statements describe the company’s policy; they do not demonstrate how well it handles a particular report. Cloudflare’s abuse approach

This helps explain why a visible infrastructure provider and a reachable page can coexist. It does not tell me whether anyone at a company knew about a specific item, when they learned about it, or what they did. I have not investigated a documented case here and cannot fairly present those details as known.

I also need to be precise about what I mean by harmful explicit content. Consensual adult sexual expression, unwanted exposure, intimate images shared without consent and child sexual abuse material involve different questions of consent, rights and harm. Calling them all unhealthy would conceal the distinctions that a serious response needs.

I care particularly about exploitation, lack of consent and exposure that people cannot meaningfully refuse. I want companies to explain what happens when they receive a credible report: who can act, whether it reaches that party, and how a person can challenge an inadequate response. Cloudflare’s reporting process includes categories for child sexual abuse material and non-consensual sexually explicit imagery. The existence of a reporting channel is a starting point for accountability, not proof that every report is resolved. Cloudflare’s reporting process

At the same time, I would be troubled by a small number of infrastructure companies becoming unaccountable judges of everything people may say. My demand is for clear responsibilities, effective responses to substantiated harm and procedures that can be questioned. Protecting expression and taking abuse seriously both require that work.

Infrastructure power is still power

I do not think “a CDN is not the origin host” should be used as a shortcut for either accusation or absolution. Technically, the distinction matters. Ethically, it does not end the conversation.

A person reporting exploitation or unwanted exposure may not know which company controls the origin server, the domain, the caching layer or another part of the delivery chain. Large infrastructure companies are often easier to identify than the operators behind a harmful site. That gives me a simple expectation: if a company cannot itself remove the underlying material, it should still make the path to the responsible party understandable, preserve an accountable reporting process and explain the limits of what it can do.

That is different from asking an infrastructure provider to police every lawful page on the internet. I do not want private infrastructure firms to become invisible ministries of speech. I want power to be matched with legible responsibility. The larger and more indispensable the intermediary becomes, the less acceptable it is for an ordinary person to be trapped between technical layers that each say the problem belongs somewhere else.

The kind of space I want Regard to become

I want Regard to give people room to describe what they feel, including frustration with powerful companies. A personal feeling deserves to be heard without being dressed up as a scientific result. A factual allegation deserves evidence even when it grows out of an understandable feeling.

That is the editorial standard I want us to practice: identify whose perspective we are reading, distinguish observation from inference, link the evidence, protect private people and make corrections visible. Review should help a writer become clearer and more accountable. It should leave room for disagreement and uncertainty.

A publication cannot certify that a person’s inner experience is true. It can ask whether the words faithfully express the position being offered, whether factual claims have support, and whether the piece respects the people it discusses. I want that care to remain available to writers who are unfamiliar with academic language or who think in Tamil as well as English.

What “verification” should mean on Regard

I want Regard to be open to people who want to share a considered thought, not only to professional writers. But openness should not mean that every statement receives the same status.

A personal essay should be allowed to remain personal. Its perspective can be attributed to its author, its revision history can be preserved and any factual claims it relies on can be checked against sources. That is a form of verification I trust because it verifies what can actually be verified. It does not pretend that a platform can inspect someone’s mind and certify that a feeling is metaphysically true.

The same rule should protect people who are being discussed. A serious allegation about another person or organisation should not become “verified” merely because the author feels strongly about it. Evidence, attribution, uncertainty and the right to correct the record matter more when a claim can affect someone else.

For the same reason, I do not want an NFT on Regard to become a badge saying this opinion is true. If Regard creates an edition, the token should prove something narrower and more honest: that a particular public revision, with a particular content digest and disclosed provenance, was approved by its author and issued under stated edition terms. The blockchain receipt can help prove the edition event. It cannot prove the truth of every sentence inside the work.

That distinction is important to me because I want the technology to preserve authorship rather than replace judgment. Public reading should remain possible without a wallet. Minting should remain optional. An edition should happen only after the author approves the exact revision, license, supply, artwork and provenance being attached to it. If the text changes later, the old edition should remain identifiable as the old revision instead of silently inheriting the new words.

This is how I want open publishing to work on Regard: speak freely, identify the kind of claim you are making, show evidence where evidence is possible, disclose meaningful assistance, and preserve the exact work you chose to publish. Verification should increase accountability without pretending to manufacture certainty.

My hope is simple. Let the reader finish a thought. Let the writer speak without turning every vulnerability into a product. Let the companies carrying our words explain the power they exercise and the responsibilities they accept. I want the internet to make more room for people to regard one another with attention they have freely chosen to give.

References

  1. Yanis Varoufakis (2024). The Age of Cloud Capital. Persuasion. Link
  2. Eytan Bakshy, Dean Eckles, Rong Yan and Itamar Rosenn (2012). Social Influence in Social Advertising: Evidence from Field Experiments. ACM Conference on Electronic Commerce. doi:10.1145/2229012.2229027
  3. Joshua D. Upshaw, Carl E. Stevens Jr., Giorgio Ganis and Darya L. Zabelina (2022). The hidden cost of a smartphone: The effects of smartphone notifications on cognitive control from a behavioral and electrophysiological perspective. PLOS ONE. doi:10.1371/journal.pone.0277220
  4. Cloudflare. Cloudflare’s approach to abuse. Link
  5. Cloudflare. Reporting abuse. Link

Citations were checked against the publisher of record. Where a claim is not backed by a reference it is marked in the text as interpretation or hypothesis.