Touching Oil
Do what AI can’t
Welcome! As we continue to understand and evolve with AI, my conclusion to date is to do what AI can’t. We’ll see where that goes, but I imagine these essays will continue to focus on the elemental aspects of human experience. This past week, paid subscribers received their first field note. A preview of the next note: highlights from 2025 game camera photos.
Thanks for reading.
If AI can create or do whatever you do, why would you keep doing it?
What if AI can’t quite replicate what you do, but it’s good enough, so much so that your audience or superiors can’t tell the difference between your work and that of Claude or ChatGPT?
For a writer:
If AI can create writing that is as good as yours, making the two indistinguishable, why keep writing?
What if AI can’t quite replicate your words, but the writing it does create is actually preferred by readers?
If what I spend hours generating can be indecipherably replicated by ChatGPT in minutes, then I should stop writing that type of content.
Refusing to use AI out of principle, keeping blinders on, is ignorant. AI is a tool. It’s here.
Avoiding it and not using it would be tantamount to not using the internet, another powerful tool. Digitized content saves hundreds of hours of driving to/from the library and sifting through materials. Therefore, not using the internet is a waste of time when the same content is readily available on your laptop screen.
Not understanding its limitations or how to distinguish yourself from it means your writing (or whatever your work is) will become indistinguishable from slop.
Use AI and evolve accordingly.
The writing craft consumed by AI is like a table saw or pocket screw replacing a dovetail joint. Christopher Schwarz tells the story of Troy Sexton in his book, The Anarchist’s Toolchest. Sexton ran an extremely efficient woodshop that reflected ingenuity and hard work. Yet, in his decision to “compete with the machines on their own turf,” he lost.
Even “the most efficient one-man factory cannot compete with a factory driven by a hundred men.”
His business withered.
Schwartz blames us: “We have become a culture so obsessed with price more than any other attribute of the things we buy,” so for the cabinet-maker, “It’s difficult to compete against furniture that costs less than what you pay for your raw materials.”
I can’t help but feel Schwartz’s radical desire to resist the corporatization of a craft. It’s very romantic. Conversely, a laissez-faire approach to tech adoption is appealing since we can conveniently ignore self-restraint and mindfulness.
I look around my living room, and there’s a lot of plastic. It will be thrown out. It won’t last a lifetime.
Is there a parallel for the written word?
Writing on this paper takes time, and it’s inefficient. I’ve tried to dictate. I’ve tried to scan and OCR. Digital notepads almost work, but there’s nothing like this pen and paper.
The current AI moment is unique for two reasons.
AI capabilities are very good, almost good enough to fool most people. But it’s also good enough for people not to care if they prefer AI writing for whatever reason. My biased thinking tells me it’s because of its simplicity; it’s the laminate flooring of the writing world. No maintenance, durable, and it looks OK.
This is what, understandably, many people prefer for its convenience and functionality. The opposite is the floor my feet currently rest on: wide pine boards with gaps between edges that collect dirt. The finish is chipped away in high-traffic areas: in front of the woodstove is hardest hit, where the dog sleeps and plays. This floor is impractical in that it marks easily, needs to be refinished, traps dirt, and is soft. The scars from Maya’s paws are artifacts I don’t want to sand away; I want them preserved. They would never have appeared on a scratch-resistant modern faux-wood floor.
When I put my hand on the plastic floor, I feel the artificial wood-grain texture stamped into the wood at an Asian factory. I’m touching oil that’s been extracted and processed into this durable piece of flooring. The pine, alternatively, originated from a mill and trees within fifty miles. I touch it, and the character warms: rough undulations that feel like wood. Both plastic and wood are products of the earth; it’s just that one is less removed, closer to its end user, than the other.
While AI is getting close, like the plastic flooring, to replicating wood, I don’t believe it ever will. That doesn’t mean wooden floors will continue to exist; it just means we may choose plastic floors as being good enough.
The second reason this is a unique time period is that a very capable AI has a full 10,000 years’ worth of human-generated data to work with. This virgin dataset is rich with nuance, experience, authenticity, and patterning. There will be no other time when AI will have such an authentic dataset to work with. Starting now, we have an increasing amount of AI-generated content. There’s the boring stuff: listicles, reviews, bland reporting. Then there are the essays and books written by AI.
So begins a cycle in which we are at the beginning of: AI-generated content that is used to generate additional AI content. What will this eventually look like? Will it be as rich as human-generated content? My feeling is that it will not. Will people care if something is AI-generated, as long as it gets across the information and is affordable, like the written version of an Ikea sofa? I’m not sure, but artificial content seems to me to represent a watering down of our collective corpus of human-made words.
I once wrote about my greatest procrastination attempt (to date): I convinced myself that I needed to build my own desk before I could begin a paid writing gig. And to do that, I needed to not only buy wood for the project, but harvest, mill, and season my own wood before building said desk.
What if I actually did that and produced an heirloom-quality desk instead of this crappy pressboard I’m slouching over?
What’s the point of creating human-generated writing or craft furniture?
I’ve racked my brain for days answering this question.
The point is that the value lies in the purpose and the form itself, almost like the concept of religion. Illogical as it is, religion serves a purpose, despite its tenets defying the objectivity that underpins our modern technological prowess.
Non-AI creation requires faith in the form we choose. And consumers of this art will know what’s genuine as long as they are committed to form. Posts from X and Notes lead to essays shrinking, an attempt to better satisfy an appetite for brevity. The conniving smoothness and cadence of an AI-generated sentence doesn’t flex my brain, but that may be OK with (or even preferred by) some people.
My inclination is to resist the complacency that comes with letting all writing eventually be authored by AI. I don’t see this necessarily as anarchy, but it does overlap with Schwarz’s description of anarchy in the context of craft: “a tendency to question the institutions that make craftsmanship and well-made furniture impossible in this modern age.”
I’d add to this the desire to keep sharing human thoughts and experience, not just that of a bot. It’s convenient to buy a chair from Ikea without maxing out my credit card, but I also like using the turned bowl my friend made from a piece of mountain ash.
Perhaps it’s futile if audiences end up preferring the too-good-to-be-true cleanliness of LLMs.
Figure out AI’s capabilities and limitations. In the end, whether by choice or necessity, all that will be left is for us to do and make what AI can’t.
Thanks for reading. Subscribe (free) for future essays.
To support the work, paid—or a one-time contribution—helps.
It's hard to stay human and create human work when we're surrounded by so much AI content, but then again, being human has always been hard and will always be hard, we were made from hardship and thrive in it! So it's really fitting that in order to create human work, we simply do what we have always done - endure and create.
“Non-AI creation requires faith in the form we choose” – perhaps that’s the most important part. And the hardest.