Over the past few weeks, I've been paying attention to something. Whenever I read about artificial intelligence, the same narrative patterns keep coming up. I want to approach one of them carefully, because I haven't fully grasped it myself yet.
It has to do with the stories that follow new technologies. With AI, too, I have the sense that however new these stories feel, they probably aren't new at all.
When I think of earlier technologies, photography, film, television, social media, the same movement keeps repeating. First a promise. Then a threat. And then a reassurance that quietly names the very fear it's trying to dismiss.
Photography was going to displace painting. Sound film was going to ruin the silent art of the moving image. Television was going to endanger family conversation. And social media first promised connection, before it became a story about isolation and frayed attention.
The easy lesson I feel uncomfortable with
I notice how tempting it is to turn this into a soothing moral: always the same excitement, and in the end things turned out just fine. That reading has a certain charm. Which is exactly why I don't trust it. It's too smooth, and I think it misses what's genuinely interesting.
Maybe the repetition itself isn't the point. Maybe it's a measuring line. If the narrative form stays constant across decades, whatever deviates becomes visible against it. The ever-same story isn't proof that nothing happens. It's a tool for seeing what does. And it measures in both directions. That television would change our attention, that social media could shake the shared ground of facts, was in hindsight not mere projection. It was remarkably accurate.
What might be different about AI
Earlier technologies were almost always described as doing something to something. Photography multiplied the image. Film set it in motion. Television broadcast it. Social media distributed it. The narrative of replacement was there, but it always meant one tool taking the place of another.
With AI, we speak differently. We don't say it transmits or amplifies. We say it thinks, decides, understands. And with that, the old worry about replacement is aimed, perhaps for the first time, not at one skill among many, but at the very faculty we'd use to judge the replacement in the first place.
I'm not sure that's true. But if it is, then what's new isn't the excitement. It's what the excitement is about.
A word that says two things
At this point I get caught on a word, as I so often do: performance.
In its first sense, it means achievement, the output we deliver. It's the question of a society tuned to compression and yield, one that asks of everything new what it means for output. The technology is then the surface. The worry about performance is us.
In its second sense, performance means the enacting quality of language. The idea that saying something doesn't just describe reality, it shapes it. The talk of performance doesn't only describe whether AI makes us more productive. It presupposes that productivity is the right measure. Whoever speaks first of relief or replacement has already conceded, at bottom, that the matter is one of performance.
The judgment sits in the frame, before the argument even begins.
The historical trace can show why the question of performance feels so hard to avoid. What it can't take is the choice of whether to make that question the only one. In my view, the point where I can still decide something is smaller and more concrete than the word "discourse" suggests. Because I'm not always aware of how the terms I reach for transform reality.