No Ads. No Surveillance.

Most of the internet you use every day is funded by advertising. That is not a secret, and it is not, by itself, a scandal. But it is a design constraint with consequences that run far deeper than the banner ads you scroll past. When a platform makes its money by selling your attention to advertisers, every design decision — what you see, when you see it, how long you see it — becomes downstream of a single imperative: keep you looking. The feed is not built to inform you. It is built to hold you.

This is what engagement optimisation means in practice. The system learns what keeps people on the platform — what makes them tap, scroll, react, share, argue — and surfaces more of it. The content that wins is not the content that is most true, most useful, or most important. It is the content that is most engaging. And the things that engage people most reliably are not the things that serve them best. Outrage engages. Fear engages. Scandal engages. Provocation engages. The carefully reported, the nuanced, the merely accurate — these often do not engage at all, at least not in the way the algorithm measures.

It Influences What You See

This is not an abstract concern, and it is part of why Veridonia exists. You might expect a war to be exactly the kind of content an engagement-optimised feed would amplify. It captures attention like nothing else. But war also reveals where the real loyalty lies: not with engagement, but with the advertisers who pay for it. When Russia launched its full-scale invasion of Ukraine in February 2022, social media was not a distraction. It was a lifeline. People used Facebook, Instagram, and Twitter to coordinate evacuations, locate family members, document war crimes, and tell the world what was happening in their cities. The stakes could not have been higher.

And yet, within weeks, something shifted. Posts about the war began losing reach. Independent Ukrainian newsrooms — some of them the only source of local information after Russia destroyed physical infrastructure — found that their audiences could no longer see their reporting in Facebook and Instagram feeds. One of them, the most popular local news site in Slovyansk, saw an 80% drop in its Facebook audience; at least 31 newsrooms were affected. Content that had been shared widely — footage from the front lines, reports of civilian casualties, pleas for help — was being shown to fewer people. The visibility dial turned down, with no explanation offered and no mechanism to contest it.

The practice had a name, even if platforms avoided using it publicly. When internal Twitter (now X) documents were released in December 2022 — what became known as the Twitter Files — they confirmed that the platform had built and routinely used what employees internally called "visibility filtering": mechanisms to suppress what users could see, at varying levels, without ever informing them. Those documents focused on political content moderation, not advertising pressure specifically, but what they proved is that the mechanism exists and was deployed at scale. Twitter executives had previously denied doing this. The documents showed otherwise. Once you know the tool exists, the question is only when and why it gets used. The practice is commonly called shadow banning, and it means that a user's content can remain technically published while being rendered effectively invisible.

The reasons behind the suppression of war content, specifically, are harder to pin to a single internal memo. But the converging evidence points in a clear direction. War content is disturbing, polarising, and commercially toxic. Advertisers do not want their brands displayed next to images of bombed hospitals and fleeing civilians. The advertising industry has a term for this: brand safety. When the invasion began, advertisers pulled spending from news content at scale. One publisher reported CPM rates on conflict coverage falling by roughly 20% even as traffic surged. The Global Alliance for Responsible Media issued formal guidance to help brands "navigate brand suitability in time of war". When a platform's revenue depends on keeping advertisers comfortable, the incentive is not to show you what matters. It is to show you what is safe. Safe for the brands, not for you.

This is not a Ukrainian problem. It is a structural problem that a war made starkly visible. The same dynamics apply to any content that is important but commercially inconvenient: political unrest, public health crises, corporate misconduct, environmental disasters. Anything ugly, anything that makes advertisers nervous, anything that threatens the smooth, palatable information environment that advertising requires. The system does not need a conspiracy to suppress it. The incentives do the work on their own.

Answering the Wrong Question

Step back from the specifics and the deeper issue becomes clear. An engagement-optimised, ad-funded platform is answering a fundamentally wrong question. It is answering: what will keep users on the platform long enough to see more ads? That is a coherent question — well-defined, measurable — and the algorithms that answer it are, by most technical standards, extraordinarily good at their job.

But it is not the question you came with. You came to be informed. You came to understand what is happening around you. You came because your attention is limited and you want to spend it on something worthwhile. The system is optimised for a different customer — the advertiser — and you are not the customer. You are the product. Your attention is what is being sold. This is not a fringe critique. It is the business model, stated plainly.

Privacy as Collateral

There is a third dimension to this that deserves its own weight.

To serve targeted ads effectively, a platform needs to know as much about you as possible. What you click on, how long you look at it, where you are, what you search for, who you talk to, what you buy, what you almost bought, what time you wake up, what makes you anxious. The business model does not merely tolerate surveillance. It requires it. The more precisely a platform can model your behaviour, the more valuable your attention becomes to advertisers, and the more the platform earns.

This is not a side effect. It is the engine. The result is an architecture of data collection that treats your private life as raw material, extracts behavioural predictions from it, and sells those predictions to the highest bidder, at a scale encompassing billions of people, largely without meaningful consent or understanding. That is not a technical oversight. It is a business model that has decided your privacy is an acceptable cost of doing business, and has not asked you whether you agree.

The Solution

The alternative is not complicated in principle, even if it is difficult in practice. Put the humans at the centre of the equation. Build a system where the people using it are the people being served, not a proxy audience whose attention is being sold to someone else.

This means, above all, committing to monetisation methods that do not distort the content. If the way you make money does not depend on maximising engagement, you do not need to optimise for engagement. If you do not sell ads, you do not need to collect surveillance data to target them. If your revenue comes directly from the people you serve, your incentives align with theirs rather than against them.

For Veridonia, that means we will never use advertising. The platform will be funded through methods that keep the relationship between the system and its users clean: subscriptions, donations, or some combination, depending on what proves sustainable and what suits the project as it develops. If it turns out that subscriptions introduce their own distortions, the model will move toward non-profit funding. The point is not to commit to a specific revenue mechanism forever, but to commit to never letting the funding model compromise what the feed shows you. Public benefit content — community feeds, the core of what the platform does — will always remain free to access. Only private benefit features may carry a cost.