The More AI Personalizes, the More Alone You Become
When content is created just for you, there's no one left to share it with.
Imagine opening TikTok and watching a video that was generated — not recommended, not filtered, but created — just for you. The pacing matches your attention span. The humor is calibrated to your taste. The background music shifts based on your mood that morning.
No one else will ever see this video. It doesn’t exist for anyone but you.
Now imagine the same thing happens to movies. You watch Titanic, but in your version, Rose dies too. Your friend watches the same title — but Jack survives. Your coworker’s version? The iceberg never appears. Same movie poster. Same title. Completely different stories.
This isn’t science fiction. AI-generated video is already here — and the economics of personalization push toward exactly this future. When the cost of producing content approaches zero, there’s no reason not to generate a unique version for every viewer.
And when that happens, something fundamental breaks.
You finish the movie. You’re moved. You want to talk about it. But there’s no one to talk to — because no one saw what you saw.
You can’t even scroll through the comments, because there are no shared comments for a film that only you experienced.
Over a decade ago, Eli Pariser, internet activist and author of The Filter Bubble, warned that personalization algorithms were creating a unique informational universe for each user. His famous example: two people Googled “BP” and got entirely different results — one saw investment news, the other saw the Deepwater Horizon oil spill.
That was about filtering. The same pool of content, sorted differently for each person. Concerning, but manageable — you could still find the same article if you looked hard enough.
What’s coming is qualitatively different. AI doesn’t rearrange existing content. It creates new content that never existed before — for an audience of one. There’s nothing to find, because there’s nothing shared to begin with.
Pariser’s filter bubble hid things from you. The AI creation bubble eliminates the very idea of a common experience.
You lose the ability to say, “Did you see that?” — not because the other person missed it, but because “that” only ever existed for you.
The Loneliness of the Perfectly Curated Feed
Think about what you do after watching a YouTube video that moves you. You scroll down to the comments — not for more information, but to find someone who felt the same way. You’re scanning for the comment that says exactly what you were thinking, and when you find it, you hit the thumbs up. That tiny act — “yes, someone else saw what I saw” — is deeply satisfying in a way that’s hard to explain.
It’s the same impulse that sends you to Reddit or X after breaking news. You’re not looking for facts. You already know what happened. You’re looking for reactions — for someone whose response mirrors yours, or challenges it in a way that sharpens your own thinking.
That spark of recognition — “You read that too?” — is one of the most basic forms of human connection. It’s why book clubs exist. It’s why colleagues forward articles to each other. It’s why your friend texts you a link at midnight with no context except “THIS.”
In a hyper-personalized world, these moments vanish.
Not because people stop consuming content. They consume more than ever. But because no one else has seen what you’ve seen. The documentary was generated for you. The essay was adapted to your reading level and philosophical leanings. Even the examples were swapped to match your cultural background.
The content is perfect. And perfectly isolating.
William Nicholson, the screenwriter of Shadowlands, gave the character of C.S. Lewis a line that has since become one of the most quoted phrases about reading: “We read to know we’re not alone.” The line resonates because it captures something true — reading is not just about acquiring information. It’s about discovering that someone else has felt what you felt, noticed what you noticed, struggled with what you struggled with.
But this only works when multiple people read the same thing.
Why We Crave Shared Experience
Research in social psychology confirms what we intuitively know: shared experiences are more intense than solitary ones. A study from Yale University (Boothby, Clark, & Bargh, Psychological Science, 2014) found that even something as simple as tasting chocolate feels more vivid when another person is tasting the same chocolate at the same time. The experience isn’t better because of conversation — it’s better because of the mere fact of co-experiencing.
We are, at our core, creatures who want to compare notes. We want to know: Did you catch that detail? What did you think of the ending? Did it change how you see things?
This is not a minor social nicety. It’s a fundamental driver of how culture forms. Shared texts — the Bible, the Analects of Confucius, Shakespeare, the Constitution — became civilizational touchstones not because they were the “best” texts, but because enough people read them to create a common language.
Classics endure for this reason. When Clifton Fadiman, the literary critic and longtime editor of the Book-of-the-Month Club, wrote about rereading classics, he observed that the book doesn’t change — but the reader does. Each reading reveals something new in you. And when millions of people across generations read the same text, the accumulated interpretations create something larger than any single reading experience.
AI-generated content, by definition, cannot do this. It has no shared audience. There is no community of readers. There is only you.
What Stays When Everything Is Personalized
If AI-generated content fragments the shared experience, what holds it together? Two things.
First, canonical works become more valuable, not less. In a world where most content is ephemeral and individualized, the books, essays, and ideas that millions of people have read become rare common ground. Sapiens. Thinking, Fast and Slow. The Analects. These aren’t just good books — they’re shared reference points. The more personalized everything else becomes, the more we’ll gravitate toward works that let us say, “I read that too.”
Second, the traces we leave behind start to matter more than the content itself. Shawn Wang (swyx), developer and author of the influential essay Learn In Public, coined the idea of “learning exhaust” — the byproduct of learning that, when shared publicly, becomes fuel for others. His point was about career development, but the principle runs deeper. When you highlight a passage, write a note in the margin, or share a reaction — you’re creating a signal that says: I was here. This mattered to me.
In a hyper-personalized world, these signals become the connective tissue that AI can’t generate for you. A highlight is proof that another human read the same sentence and stopped. A margin note is evidence that someone else wrestled with the same idea. These aren’t content. They’re presence.
This is what we’re building at Glasp — a place where the highlights, notes, and reactions of real readers are visible to each other. Not algorithmically generated. Not personalized into isolation. Just one person’s honest mark on a passage, available for anyone else who reads the same text to discover.
The internet’s 1:9:90 rule tells us that in any online community, 1% create, 9% contribute, and 90% lurk. But highlighting changes the math. You don’t have to write a blog post or record a video. You just have to read — and leave a trace. That trace becomes someone else’s moment of recognition: “Someone else underlined this exact sentence.”
AI can personalize your feed. It can generate content tailored to your taste. What it cannot do is tell you that another person — a real, specific, unpredictable human being — found the same paragraph meaningful.
That’s not a feature. That’s the foundation of connection itself.
→ Browse what people are highlighting on Glasp
→ Start highlighting the web with Glasp
What’s something you learned recently that you almost didn’t share — and what held you back? Leave a comment. We read every one.
Forwarded this email? Subscribe to Glasp Newsletter
Start highlighting the web with Glasp