You need rules for AI (Socrates had one for writing)
1. The potion
What if I told you I have a potion that will make you wiser and improve your memory?
And I’m no sleazy car salesman. I’m the inventor of numbers and astronomy.
But you still say no.
I am Theuth, the Egyptian god of knowledge, and you are Thamus, king of Egypt, trying to improve the lives of your people.
You say no because you claim the potion will have the opposite effect. I’m selling writing, but you say it will introduce forgetfulness into the souls of those who learn it. They won’t practice using their memory, because they’ll put their trust in writing.
You claim I have discovered a potion not for remembering, but for reminding. It’s a pseudo-intellectualism that looks like wisdom without being properly taught.
And what if I told you, plot twist, that I’m no god? I’m actually just Socrates, trying to convince you that I hate writing.
I’m just arguing with myself, making up an Egyptian myth, and you’re my buddy Phaedrus, who actually calls me on my BS.
I wish I’d had more buddies like you, because I was later executed, making me the first person in recorded history to be put to death for being a troll.
I also compare writing to a painting. A painting may look real, and you may want to talk to it and understand it further, but it isn’t, and it won’t talk back.
Writing is the same: it won’t reveal anything more than what’s already on the page.
Socrates never liked writing, and he never wrote anything down. It was his students, Plato among them, who loved his ideas and also saw the value in writing.
They wrote his teachings down, some of it decades later, so they would outlast his life. And they have.
Socrates had his own rules for the technology of his day, and he backed it up with myths and stories that made sense to him.
Technological shifts throughout history seem to create a split in how people perceive the change and how they adopt it.
The split is really about their own rules: where they draw the line. Every person draws a different line for how much of the new thing they’ll take on.
You may have already drawn some lines of your own with AI, at least in your head. For me, it’s that I don’t use AI to write a text for my partner. For you, it might be that you don’t let it make purchases on your behalf. Wherever you land, now is the best time to start writing your own rules.
2. The farmer’s list
Two thousand four hundred years later, near the turn of the century, farmer and poet Wendell Berry wrote an essay called “Why I Am Not Going to Buy a Computer.”
The year is 1987. It’s the start of the personal computer boom, and writers are being told to switch. But Wendell Berry draws a line in the dirt, in his own farm soil. He writes with a pencil, and his wife, Tanya, types his drafts on a typewriter and edits them.
Farmer, wife does the typing, refuses new technology: you might be thinking of a certain capital-C political label. You’re right. He’s a capital-C Conservationist.
Berry goes on to explain why he refuses to adopt the computer as a writing tool, and his reason is a moral one. He doesn’t trust the energy corporations, especially the coal companies that have stripped the mountains of Appalachia and the earth in general. He writes:
“I would hate to think that my work as a writer could not be done without a direct dependence on strip-mined coal.”
Finally, he sets standards for himself for adopting any new technology in his own work. He gives nine of them, but they come down to three ideas that stood out to me:
- Cost: it should be cheaper, smaller, and use less energy.
- Better: it should do the work clearly and demonstrably better than what it replaces.
- Relationships: it shouldn’t disrupt anything good, including family and community.
Not everyone liked his list. When Harper’s reprinted the essay, readers wrote in, and they went straight for the first rule.
You know what’s cheaper, smaller, and uses less energy than a computer? A pen. Another suggested we should all just get a wife, so we wouldn’t need a vacuum cleaner or a washing machine either.
They were a bit harsh, and I like Berry’s pushback. He wanted to keep his marriage private, which is a hard thing to defend in public. But he made clear that he and Tanya both find meaning in the work they do together.
That’s the important part: they’re his rules, and they don’t really have to matter to anyone else.
We could argue about whether these are good rules. What matters is that he wrote them down. And you may already be holding them up against how you use the new technology of our time: AI.
AI keeps getting better at more skills. Lately people have been posting demos of the new Claude models editing video, which is a scary thing to watch if you’re a video editor.
And by Berry’s first rule, it passes: it’s demonstrably cheaper than hiring a professional.
I use AI too. I’ve used AI video generations in my cold intros to visualize what I’m talking about, and on Reddit some people called them slop. Honestly, a few of them didn’t meet my own standard either.
So here’s my new rule: if I don’t have time to make them good, I don’t use them.
Berry’s second rule asks whether the new tool does the work clearly and demonstrably better. Take AI coding. People like to call AI code slop, but they forget that a lot of human code was slop too. I’ve read and reviewed plenty of it.
If you coded before AI, you know the feeling: you leave review comments and ask for changes, then feel bad asking for even more, until you eventually surrender and type “Looks good to me, ship it.”
Berry’s third rule, that a new tool shouldn’t disrupt anything good, including family and community, is probably his best. People shouldn’t be outsourcing their loved ones.
Which brings me to another rule of mine: companion AI isn’t good for anyone under 60 who isn’t in a retirement home.
For older people who are lonely, these tools can be great. They can offer connection, a bit of stimulation, a nudge to engage with the real world, and help thinking things through. Outside of that, and outside a learning environment, which most people aren’t in, I don’t think people should use them as a replacement for real friends.
3. The red water
Neil Postman was a media and technology critic and a professor at NYU. In 1998, he gave a talk called “Five Things We Need to Know About Technological Change.”
Speaking to a group of faith leaders, he gave five ideas on how to deal with new technology, and a few of them stand out for anyone making rules.
One of his ideas is that technological change isn’t additive; it’s ecological. Take the car. Berry would see a trade-off, but it’s bigger than that. It didn’t just add a faster way to get around. It polluted our air, choked our cities, and degraded the beauty of our natural landscape.
Put more clearly, drop red dye into a beaker of water: you don’t get clear water with one spot red. All of it turns red.
In Postman’s words, you get “a new coloration to every molecule.”
Writing your rules is deciding what you’ll use the water for. Maybe some of you will cook with it. Maybe some of you will drink it.
But there’s no undoing the red dye. The labs have already added it; we’re left with the change, and you have to make your own rules.
The fifth idea is that technology becomes mythic. Once it changes everything, we forget it was ever a choice. It starts to feel like part of the natural order of things, the way the alphabet does.
This is why Socrates’ argument sounds a bit absurd to us: writing has gone mythic. It’s only because his own students drew a different line that we know his.
Berry rejecting the computer in 1987 may sound just as absurd now. But it made sense to him at the time; he lived by it, and he got the last word.
AI hasn’t gone mythic yet. Now is the time to draw our lines and set our own rules, before we stop questioning it at all.
So write your rules. Share them if you want, and expect the snobby replies from people with their own rules and values. But at least you’ll have yours, and you’ll have thought them through. The existential dread gets a little quieter once you know where your lines are.
Are you going to sit back and let AI reach mythic, maybe even godlike, status? Or are you going to be like the king of Egypt and push back at the inventor of math and dice?
Gods can invent the dice. You choose the rules.