Are You Using Technology, or Is Technology Using You?

I just finished James S.A. Corey’s excellent The Mercy of Gods (I was rushing last night so I could bring the second book in the series, The Faith of Beasts, on my trip to Italy for a conference today). The book is, like the last book I wrote about, a first contact sci-fi novel, but here first contact goes awfully wrong. In short, the aliens, referring to themselves as the Carryx, arrive, the war is over before it starts, and the surviving humans are absorbed into a vast and terrifying civilizational machine.

The remainder of the book, detailing humanity’s quest to survive within the cruel galactic civilization that has absorbed them is claustrophobic and does a good job of capturing the sensation of being a small working part inside a mechanism whose purposes are not your own. The humans are given tasks to perform with unclear instructions, and as the book goes on you, as the reader, understand more about how these tasks fit into the larger social mechanism created by the Carryx.

As a metaphor for our claustrophobic relationship to screens and addictive technologies the book hits eerily close to home. Jonathan Haidt and Catherine Price make a variant of this argument in The Amazing Generation: are you using technology or is technology using you?

That book, which my wife picked up recently, is geared towards kids and pre-teens, and provides useful advice on how to avoid addictive technologies and how to build a rich adolescence. It also, I think, speaks to adults working inside the emerging AI ecosystem, because we are constructing the next layer of a vast socio-technical machine right now, and the question of which direction the agency flows in is largely being settled by default.

The ways in which contemporary digital technologies capture our attention in everyday life and fuel addiction, seems to work something like this: I have a task to perform, such as checking the weather for the day so I know how to dress my kids. This task requires a device, and in general the device is general-purpose: a smartphone or a laptop. This bundles the capability to perform the task together with capabilities to perform myriads of other tasks, which means the path to checking the weather leads to checking my email, which leads to googling something, which then leads to a long session surfing the web (yes, I know doomscrolling social media is the new thing to be addicted to, but that’s not my poison). Of course, this is the point—contemporary digital technologies are constantly being iterated upon to increase “engagement.”

I realize, in the same way that an alcoholic can’t get sober hanging out in a bar every day, I can only maintain a healthy relationship to digital devices through re-architecting my environment. Hence, I increasingly find myself bifurcating my approaches to technology. On the one hand, I am increasingly re-adopting single-purpose, lo-fi devices. I wear Casio digital watches that tells the time and nothing else. I recently bought a Walkman-style cassette player, which is great when I want to listen to some music in the evening while winding down by reading a novel. I draft a lot of writing (such as this post!) on a reMarkable e-ink tablet that cannot show me a notification because it just isn’t into that kind of thing.

On the other hand, I use AI (mainly Claude) maximally for production work: websites, interactive artifacts, syllabi, slide decks, program proposals, literature reviews, code for data collection and analysis, drafting prose, etc. Increasingly, leveraging AI is becoming central to the knowledge work that I do. The more I use AI, the better results I am getting. I am finding that the clearer I am about what I want out of a large language model, the easier it is to set up the context and the prompt to generate what I want. This is changing how I do research at a fundamental level.

There are also areas where the lo-fi and AI-enabled workflows are starting to hybridize. For example, I wrote the first draft of this post by pen on the reMarkable tablet, I then fed it to Claude, which generated a draft, which I am now editing by hand on my laptop at the airport. I expect more of these hybrid workflows, where each tool is being used for what it is best at, to emerge over time.

I am not a Luddite, nor am I anti-AI. The goal is to automate a machinery that runs below the surface of my life, doing the production work, so that the surface itself provides a rich texture where curiosity, exploration, learning, and achievement can happen together with other humans. In Corey’s novel, the horror is that humanity ends up inside the machine, serving its purposes. My project is keeping the arrow pointed the other way through adopting and designing technologies that serve human life.