The average reader spends 52 seconds on a blog post. Not 52 minutes. Seconds. Nearly three-quarters of people admit to skimming rather than reading. Scroll depth has dropped to about 55% of a page before visitors bounce. None of this is new. But something has changed the equation entirely: the supply side exploded. AI can generate a 2,000-word article in seconds. The cost of producing written content has collapsed to near zero. And the result is an internet where more than half of all published articles are now AI-generated, according to a 2025 analysis by SEO firm Graphite that examined 65,000 English-language URLs. Europol warned in 2022 that 90% of online content could be synthetically generated by 2026. Whether or not that exact number holds, the direction is unmistakable. We have created an ocean of words that nobody swims in.

The paradox of abundance

When content was scarce, reading was valuable. A well-written article in a magazine had weight because someone chose to research it, write it, edit it, and publish it. The scarcity was a quality signal. Now every company, tool, and bot can publish endlessly. Marketing teams use AI to produce dozens of blog posts per week. Startups auto-generate entire content libraries before they have customers. According to McKinsey's 2025 State of AI report, 73% of businesses now use AI for content creation, achieving a reported 4x increase in output. The paradox is this: as the supply of content approaches infinity, the value of any individual piece approaches zero. But human attention has not scaled. There are still only 24 hours in a day, and the average person reads about 200 words per minute. The bottleneck was never production. It was always consumption. Reading has become the scarce resource.

It is not just social media anymore

The attention economy is a familiar concept. Social media platforms compete for eyeballs. Short-form video eats into long-form text. This has been happening for over a decade. But the new pressure comes from a different direction. It is not just TikTok competing with blog posts. It is AI-generated blog posts competing with other AI-generated blog posts, all competing for the same finite attention. Every channel, from email newsletters to search results to LinkedIn feeds, is now flooded with machine-produced text optimized for engagement metrics rather than understanding. The sheer volume creates a noise floor so high that genuine signal struggles to break through. Readers, overwhelmed by the deluge, develop rational coping strategies: they skim faster, read less, and trust nothing.

The cost of shallow reading

This is where the real damage happens, not to content creators, but to readers themselves. Researchers have drawn a clear distinction between shallow reading, the fast scanning we do online, and deep reading, the slow, immersive process where we build mental models. UCLA professor Maryanne Wolf, one of the leading scholars on reading and the brain, describes deep reading as a process that promotes "vocabulary, reasoning, concentration, and critical thinking skills." It also builds empathy and social perception, cognitive processes linked to better health outcomes and longer life. A 14-year longitudinal study published in BMC Geriatrics found that people who read frequently (at least once a week) had a 46% reduced risk of cognitive decline compared to infrequent readers, regardless of education level. The mechanism is straightforward: deep reading exercises neural pathways that maintain cognitive function over time. Skimming builds none of this. When you scan a 2,000-word article in 52 seconds, you are not engaging in deep reading. You are pattern-matching for keywords. A generation trained on summaries and bullet points gradually loses the capacity to hold complex arguments in their heads.

The irony of summarizing the summary

Here is a pattern that captures the absurdity of the current moment: people use AI to summarize articles that were written by AI. An AI writes a blog post based on a prompt. A reader asks a different AI to give them the key takeaways. Each step introduces information loss, context collapse, and subtle distortions. The original nuance, if there was any, evaporates. What remains is a kind of telephone game played between language models, with humans barely involved. This is not a theoretical concern. It is the default workflow for millions of knowledge workers right now. The question is not whether fidelity degrades, it is how much, and whether anyone notices.

Why short and dense beats long and hollow

The response to content abundance is not to produce more content. It is to produce denser content. Orbit Media's 2025 blogging survey of 808 content marketers found that the average blog post is 1,333 words, and only 9% of writers publish posts longer than 2,000 words. But here is the twist: 39% of those who do publish longer posts report strong results, nearly double the 21% benchmark. Length alone is not the variable. Depth is. The posts that break through the noise tend to share a pattern: one clear idea, developed with specificity, delivered with respect for the reader's time. Not a 3,000-word SEO play padded with filler. Not a listicle generated in eight seconds. Something that rewards the act of reading by offering something you could not get from a summary. When the supply of words is infinite, the competitive advantage shifts to the quality of thinking behind them.

Choosing to read is an act of resistance

Books remain the last long-form medium standing. Even podcasts, ostensibly a slow medium, get consumed at accelerated speeds. Research from UCLA found that college-aged students could listen at double speed with minimal comprehension loss, but other studies show significant drops in learning beyond 1.5x. The instinct to compress, to get through more faster, is everywhere. Choosing to read deeply, to sit with a long argument and let it reshape your thinking, is now a deliberate act. It requires opting out of the default mode of consumption. It means accepting that you will process less information in exchange for understanding more of it. This is not a nostalgic argument for the good old days of print. The point is structural. In an attention economy flooded with zero-cost content, the ability to think carefully is the thing that compounds. Deep reading is how you build that ability. It is not a leisure activity. It is cognitive infrastructure.

The ecosystem has changed

None of this is about being a Luddite. AI-generated content is not inherently bad. It is useful for documentation, translation, first drafts, and a hundred other applications. The problem is not the technology. The problem is what happens when infinite supply meets finite attention without any new filtering mechanism. The old filters, editorial curation, publishing costs, reputation, have eroded. The new filters, algorithms, AI summaries, social signals, optimize for engagement rather than understanding. And readers, caught in the middle, default to the path of least resistance: skim, extract, move on. The humans have not adapted yet. But the ones who do, the ones who choose depth over volume, who read one excellent piece instead of scanning twenty mediocre ones, will have a compounding advantage. Not because reading is virtuous. Because in a world drowning in words, the ability to actually think about what you read is the scarcest skill of all.

References