This reflection was sparked by this tweet from Peter Steinberger, whom you should absolutely follow on Twitter and read on his blog. I came across his account recently and started reading his writing. For an amateur web coder, his blog is surprisingly rich with practical, thoughtful tips. I’ve also been poking around the tools he’s built and quietly incorporating some of the ideas he shares.

The tweet in question was about a Chrome extension he built that lets you summarize anything on a web page and then have a conversation with it. And that, in a sense, is the crux of what AI is enabling at scale: summarization.

Now, if you fall into the category of people who shout at clouds about the decline of reading, the decline of thinking, the decline of writing, the death of humanity, and so on—you know, the reflexive “kids these days” crowd—you are probably going to have a rough decade. Because AI now enables compression and summarization at a scale hitherto unseen in human history.

The absurdity of the technology is such that you can take a 600-page book, upload the PDF, and ask an LLM to produce a four-line summary. That is the world we are in. I can easily imagine the same people who have been crying hoarse about the decline of reading seeing AI as yet another death knell. Instead of grappling with a difficult text they were already inclined to abandon, people can now just paste it into an AI and get the CliffsNotes version of *War and Peace*, *Crime and Punishment*, or *The Brothers Karamazov*.

If you are deeply invested in the idea that longer-form reading is already dying, then yes, AI will probably accelerate that decline. Books, long essays, dense arguments. All of it.

But I’m not fully convinced by this story. My working thesis, which could very well be wrong, is that reading hasn’t declined so much as changed. Most people never read that much to begin with. Many wanted to read, but were intimidated by the sheer physical and cognitive bulk of certain works.

*War and Peace* at 1,300 pages is not an easy psychological commitment. *In Search of Lost Time* runs to roughly 3,000 pages. These are not books most people can even bring themselves to buy, let alone read.

What AI does, through compression and transformation, is lower the activation energy. It allows people to engage with difficult material in new ways. You can ask for a simpler rewrite. You can ask for explanations in plain English. Yes, something is inevitably lost in translation. But something, to me, is better than nothing. If people read even a mediated version of these ideas and take something away, that seems like a net positive.

More interestingly, AI can deepen reading. Traditionally, when you encounter a difficult passage, you have two choices. You either push through, half-understanding, or you stop and Google around, often landing on shallow or unsatisfying explanations. With AI, you have something closer to a tutor. A companion. You can ask questions in real time, clarify concepts, explore tangents, and connect ideas across disciplines. The reading experience itself becomes denser and more rewarding.

I had a firsthand experience of this while reading *The Poetics of Space*, which I found dense and occasionally nightmarish. I took screenshots, fed them into Claude Sonnet 3, which in AI time already feels prehistoric, and suddenly the book became navigable. Not easy, but navigable. That mattered.

Of course, there are real tensions here. I know people in media who are anxious about changing reader behavior. They publish long-form work, only to discover that many readers are pasting the entire article into an AI and reading a summary instead. That is not a trivial concern.

But historically, shifts in medium have always forced shifts in form. I suspect we will see new kinds of journalism, new creative forms, new narrative structures that are native to this AI-mediated environment. We are likely headed toward a Cambrian explosion of styles and formats, enabled by the capabilities of large language models.

If you care about the written word or the spoken word, we are entering a strange interregnum. I can’t help but think of the old Gramscian line: the old world is dying, the new world struggles to be born. In this interregnum, monsters appear.

Loose thoughts, no firm conclusions. But interesting times, undeniably.

*Speaking of new forms: I voice dictated this to ChatGPT, got a transcript, and had Claude clean it up. The medium is changing how we write, not just how we read.*