The Abundance Mirage — Cezar Babin

The Abundance Mirage: why the road to hell is paved in unchallenged assumptions

June 2, 2026 · Note · 660 words · 3 min read · 100% human<br>In the whole wasteland of AI discourse, there is perhaps no more dangerous proposition than "Abundance is universally good", a statement that is almost too tautological to be challenged. To illustrate my point, I'm going to walk you through a thought exercise in designing a dictatorship from the ground up without employing any obvious mechanisms for oppression.

We'll start by disassembling freedom of press, a pillar of a well-functioning democratic society. Using any physical or legal power mechanisms is out of the question. So would be bribing or cutting or interfering with media actors behind the scenes. In fact you'd have to assume that any second order or third order or nth order mechanism of intervention would eventually be exposed. Interference has to be significantly more subtle, to the point of being ingrained in the de-facto modus operandi and incentives of the system. For example: market forces.

There aren't many people that will dispute that for the press to be independent, it has to be self-sustaining. It follows that publishers, like any other economic agents, have to sell a product that people want and be able to charge more than it costs to produce it. There are only two good ways to monetize content: direct-to-consumer, and advertising. And while you can't directly influence who buys the content or intervene in the advertising relationship, you could very well influence the medium at a higher level, with the "intention" of actually promoting freedom of press.

Journalism, just like anyone else in the economic medium, is competing for a scarce resource: attention. So anything that makes attention more scarce, or less economically viable, would have detrimental effects on monetization, and consequently freedom of press. Sure, this is not exactly a recipe for dictating a prescriptive agenda, but it does muffle out the message for those that do try to get the real message out there. This, for example, explains how promoting and expanding the universe of publishers may seem like a good thing on the surface, but incredibly detrimental to the quality of information that gets to the average consumer.

A blunt, unsophisticated dictator would go for an obvious, liberty-infringing mechanism, yet the more effective tactic is to use existing and unchallenged beliefs to subvert the system by forcing its limits. Every system is ultimately constrained by a bottleneck, and independent journalism is constrained by human attention.

AI-powered abundance is similarly implied to be a widely accepted "good thing" on paper, but assuming that it will universally remove scarcity across the board is a chimera. Technology has always been about making something massively easier to produce, and the last 25 years of internet-powered transformation have consisted of catalyzing an oversupply of goods with the not-so-subtle goal of monetizing the abundance with a routing layer built on top of it. Abundance at one layer almost always implies scarcity at a different layer of the stack.

The other more concerning dynamic at stake is a tradeoff imbalance. End-users are the primary beneficiaries of AI abundance, with minimal tradeoffs, while the real tax is being paid by the businesses who lose value capture mechanisms due to the centralization of distribution. On a long-enough horizon, lack of competition will also result in a degradation of services provided to consumers, in theory. In practice, as we've seen it play out with the Google anti-trust lawsuit, when high-quality service can only be provided at scale, it is much harder to argue about the hypotheticals of a de-centralized world. The most likely outcome is a world where AI labs still get to capture a significant portion of the pie, with a minimal tax that indicates that society tried to solve the problem and the tax is as good of a compromise as we can get to without degrading service.

While bottlenecks in distribution are being discussed in pockets of the internet, the details are extremely fuzzy. The only way to get ahead of this problem is to call out specifically:

What are the exact mechanisms through which the wider ecosystem of products and services hurt by AI can protect themselves?

A new intellectual property framework that allows the idea → capability conversion cycle to survive in the face of "datacenter countries".

Financial mechanisms that liberalize access to capital participation (vs labor-only participation).

← All notes