I believe one of the problems is extreme GEO, and I notice it when I search for technical information or solutions to problems with devices or software to do things; I find dozens of identical articles on different sites, often on the blogs of companies selling the solution, disguised as lists of the "5 best ways to" where there are indeed many tools, but only theirs has the link and is always first or best or fastest. Setting aside the fact that GEO for me is the ever-necessary SEO with a new name, we are truly witnessing the capitulation of the great promise of the internet: information at the click of a button connected to each other for a seamless and limitless browsing and discovery experience. And I know well that this sounds like a nostalgic old boomer's speech.

The Problem of Impeccable Mediocrity

The temptation is strong and calls to us loudly. AI generates correct, structured texts that can be published in minutes, but they are identical to thousands of others. Formally perfect, strategically banal. This is what I call Impeccable Mediocrity: output that contains no errors but also lacks value. No overly original perspective, no truly recognizable angle, no personal signature. As I am writing in my upcoming volume "Strategic Metaprompting".

Those who accept the first coherent output and publish it are fueling the phenomenon that the Dead Internet Theory describes: interchangeable content that no one reads, written for algorithms that reward quantity and indexed by crawlers that do not care about value, at least for now. It’s a vicious circle: the more generic content there is, the more noise grows, the harder it becomes to stand out, the more tempted one is to produce even more content to compensate. The way out, in my opinion, is to use AI differently.

Radical Quality: What It Really Means

Radical quality means texts that carry a signature, ours, that could not have been written by anyone else, that add something that was not there before. Length and formal refinement are secondary. What matters is the difference between using AI as a printer or using it as a panel of assistants to direct, real assistants so that we do not become the "copy-pasters of AI," to quote Marco Camisani Calzolari.

In my work, I have developed an approach that I call the 3C Protocol: Compare, Challenge, Curate. I compare the outputs of different models on the same task and make them challenge each other when they converge, because convergence can be automation. But I decide how to do it, what to keep, what to discard, what to rewrite. Because the final peak of Curate makes the difference: it means taking responsibility for the result. I think that those who use AI as an oracle to accept answers are delegating the most important part of the work: judgment.

Generic content answers questions that perhaps even the author has never asked, with information available everywhere. Quality content starts from a specific, personal, human angle, and adds a perspective that comes from experience to take a stand. And I am not saying not to use artificial intelligence.

The Choice of Side

The web, in addition to filling up with spam, popups, and Xs to click, has filled up with content that no one really reads, written by AI to be read by other AIs. Think about it: AI-enhanced browsers automatically navigating pages generated by other artificial intelligences, agents interacting with services managed by other agents. Artificial navigators visiting artificial content in an artificial way. It’s a toxic spiral that poisons the wells for those seeking real value.

Those who work with AI professionally can decide whether to ride the mass production wave, hoping that the numbers will compensate for the lack of substance, or to focus on radical quality, accepting that it requires more time, more method, more responsibility. And no, it’s not a shortcut; it’s a choice of side.