Neil Postman wrote his book about television in 1985, but if you read Amusing Ourselves to Death now, it's hard to escape the feeling that he was actually writing about something that hadn't been invented yet. The book is about how the dominant medium of an era shapes the way the culture inside that era is able to think. He was warning about TV. Forty years later the warning lands differently, because the medium has changed again, and this version of the medium generates the content itself.
The medium is the metaphor
Postman's argument runs roughly like this. Print encouraged habits of mind that worked across hundreds of pages: linear reasoning, abstraction, sustained attention, the patience to follow a long chain of thought to wherever it ended up. Television demanded a completely different set of habits. Short bursts, image processing, emotional response, entertainment value as the price of admission for any serious topic to even make it on screen. His claim was that you can't have a culture do serious political discourse when the dominant medium for that discourse is a 30-second commercial spot. The medium doesn't just deliver content. It shapes what kinds of content can survive at all.
Now think about what happens when the dominant medium becomes one that generates content instead of transmitting it. When the place we go to encounter writing, draft emails, summarize meetings, hold conversations, and work out our own thinking is a model making the output up on the fly based on what has worked statistically in the past. That's a different shift than Postman was tracking, but the structural question is the same. What habits of mind does this new medium reward, and what does it quietly erode?
Huxley over Orwell
Postman opens the book with a contrast worth pulling forward. Orwell warned about a future where information gets suppressed and the truth is something the people in power can lock away. Huxley warned about a different future, one where information would be so abundant and so easy and so entertaining that people would simply lose the capacity to care about anything that required effort. Postman thought Huxley had the better forecast, and at the time he was talking about cable news and game shows.
That's the bet that turned out to be right. We didn't lose access to information. We got drowned in it. The version we're living through now, with AI pushing the marginal cost of producing content toward zero, makes the original Huxley problem look quaint. Every channel that used to have a quality floor because writing took time has had that floor pulled out from under it. Every form that used to mean something because somebody bothered to do it can now be done in three seconds with nobody in particular doing it.
What's different about LLMs now
What's been changing about LLMs over the last couple of years is how personable they're getting. The models remember you across sessions, they write in your voice, and each release pushes the interface a few more inches toward feeling like a person who actually cares about you.
Postman noticed that television had a tendency to take serious things and dress them up as entertainment, because the medium itself couldn't really do much else. LLMs flip that move on its head, dressing generated text up as serious things, producing output that reads like a thoughtful essay or a personal letter from a friend who's been thinking about your situation. The form has been disconnected from any actual source, and what reads like a person who cares is a model arranging tokens that statistically resemble a person who cares.
What this means for the work I do
This is where it gets practical for the kind of work I actually get hired for. When the cost of producing competent-looking content drops to zero, the value of competent-looking content drops with it. Every brand can now generate on-brand copy, on-brand imagery, on-brand campaigns at a volume that was unthinkable two years ago. The flood is already here. Anyone who's spent five minutes on LinkedIn this week knows what it looks like.
The premium for the kind of work I'm hired to do has shifted hard toward the things AI can't fake convincingly. Voice that's recognizably yours, and the kind of editorial taste that only gets built up over years of putting things into the world and watching what actually lands. Both of those come out of being a specific person who's been in the room and made the calls. The brands I'm working with right now are starting to ask the harder version of Postman's question for themselves: what should we be paying attention to as a company, and what do we want the people on the other end to feel as actually coming from us versus coming from a system wearing our brand as a costume?
The question that survives the next model release
Postman wasn't trying to talk anyone out of television. He was a teacher and a critic. His project was to make people conscious of what the dominant medium was actually costing them, so they could choose differently in the places where it mattered. The book has held up because the question doesn't go away when the technology changes.
Forty years later, the only useful version of that question is what we want to keep doing the slow way, on purpose, even when AI can do it faster. The writing we want to do in our own words because the writing is part of how we figure out what we actually think, and the work we want the people on the other end to encounter as something that genuinely came from us and not from a system trained on us. The medium is going to keep shifting. The question of what we're paying attention to, and what we're letting get quietly hollowed out, is the part that doesn't change.