Why No One is Right about AI

And how to assess a new technology

Sep 01, 2025

There’s a new technology that you guys may have heard of: “Artificial Intelligence.” It seems to me that most normal people (i.e. not tech investors) are underestimating the importance of AI.

I want to offer 2 frameworks for asking good questions of new technology. The first comes from Andy Crouch. The second comes from Neil Postman. While neither author is specifically addressing AI (their theories predate the rise of modern AI), they offer an orientation to thinking about new tech that is helpful and relevant. Importantly, both Crouch and Postman offer insight that challenges the way we typically think about new innovations, and problematize the way we’ve been thinking about AI.

Before looking at their insights, it’s worth pointing out that in our culture (although this might be changing), the baseline assumption is that if a new device improves efficiency or extends human capacity, it’s good. But this assumption is pretty obviously simplistic.

All technologies extend or amplify human capabilities. So, it’s not enough to ask whether a device extends our capacities; we have to ask Which capacities? In which directions? To what end? You can see how, once you get beyond the efficiency-worship, there are lots of complicated questions to be asked.

These are the questions that many people are asking today about AI: Who should use it? What hard work will it make easy? How will it affect X industry? These questions make sense: every new invention promises to make something easier to do. What you don’t see upfront is that they also make other things harder. This is where Crouch and Postman are so helpful.

Both Andy Crouch and Neil Postman argue that technology/innovation is not merely additive. New innovations don’t only make things easier, they make some things harder. These frameworks are helpful because they point to the effects of technology that you won’t see on the posters. They help you see the losses and changes that are often overlooked with the introduction of a flashy new device.

I think before we start asking the normal questions about AI (should I use it for my research paper? Should I use it for poetry?, etc.), we need this broader framework: What losses and changes does this technology introduce? Here are the two ways these theorists think about innovation.

Crouch

Crouch’s insight is that every innovation or culture change makes something obsolete. Innovation doesn’t merely extend our capabilities, it shifts them. While making some new things possible, it makes other things nearly impossible.

Crouch gives the example of the US highway system. It’s a stunning innovation (that most of us think of as normal or uninteresting). Think of what it makes possible. If I want to drive the 2000 miles from Chicago to LA, I just hop in the car. I barely even need a map. Everything I need is literally laid out in front of me to make that journey, and the obstacles have been blasted away with dynamite or bridged-over by a team of engineers.

More than that, the highway system offers systematized roads that have enabled us to build fleets of shipping containers that attach to trucks that are constantly striping our country like ants, carrying stuff. 2-day shipping (but also always-stocked stores and inland sushi) doesn’t happen without the highway system.

On the other hand… Crouch wonders, What if I wanted to ride a horse from Philly to New York? That’s a journey that many people made before the highway system was invented. Now it’s nearly impossible. The trails the horses used to use have been paved over. The roadside inns you’d find on your journey likely don’t serve hay.

Crouch’s point is that innovation doesn’t simply expand, but shifts the window of possibility.

Postman

Postman’s insight is similar. He argues that technology isn’t merely additive; it’s ecological.

The best example is the invention of central heating. Central heating is amazing. Now you don’t have to load your stove with wood and start a fire every day. Now you don’t have to drag a tree into the backyard or chop firewood. Now you control the temperature of every room in the house with one button.

But this changes the environment. Whereas there used to be a central gathering place in the home (around the fire), now everyone in the family can comfortably go their separate ways. The kids can hide out in their bedrooms rather than gathering around the fire. Central heating creates a new environment, and in this new environment, it’s easier to be isolated.

Or we could apply Postman’s insights to the highway system. The systematized roads that cross our country have literally changed its ecology. For starters, the interstate system covered over 1 million acres of land in concrete.1 That affects more than just water drainage and the heat profile of the landscape. It also cuts through migratory bird flyways. It also creates demand for roadside food and gas. It also spurs more road construction to connect the highways to other roads.2

It also changes our neighborhoods. Did you know that the US highway system intentionally razed and paved over neighborhoods that the developers didn’t like (read: black/brown neighborhoods)?3

More than simply making it hard to ride horses or walk in our cities, the highway system is a technology that changes the environment. Urban sprawl, mall culture, fast food, the erosion of regional distinctives, the death of small towns, family road trips, rural manufacturing, Disney World, housing shortage, … the list of unintended consequences goes on and on. Innovation reshapes the world. And changes like that—changes to the shape of our ecology—always have unforeseen and unintended consequences.

AI

I won’t waste your time with predictions about the changes AI will bring. But I will say this: I think no one knows what consequences the mass adoption of AI will cause.

When the internet was invented, no one was thinking that the medium-term consequences would be that there would be one store where you buy everything, and middle school children would be far more anxious. In hindsight, Amazon and social media seem inevitable. But when the internet was a way of sharing data among research institutions, how could we have known?

No doubt, AI will transform the way we communicate and access information. In these narrow arenas, the changes may be net positive. But—following Crouch’s insight—AI will also make some things much harder. In 10 years, will it still be possible to learn through slow investigation of sources? Will it still be possible to communicate privately? Will it still be possible to publish news or research that people trust?

AI will build fast lanes for efficiency, but in doing so, it will tear up the slower routes where much human development, trust, and authenticity have traditionally been cultivated.

And, ultimately, AI will change our environment. One positive effect in this regard is that it seems that AI will democratize access to expertise. Knowledge (medical, institutional, etc.) won’t be segregated to cities or sequestered among elites.

On the flip side, there’s no obvious reason to think that the long-term ecological shifts AI will bring will be positive. What is it like to live in a world where everyone knows everything and no one knows anything? In the very near term, we may live in an environment where it’s extremely easy never to talk to another person face-to-face. Humans may do the equivalent of hiding in their bedrooms once we no longer need to gather around our proverbial fireplaces. What happens to a society when people no longer interact?

When the highway system was built, the engineers weren’t thinking about migratory birds or monarch butterflies, but their technologies changed the whole environment. This type of ecological shift will lead to consequences that we can’t foresee.

Tidbit Pile

Discussion about this post

Ready for more?