- Arcbound
- Posts
- Is this still my newsletter?
Is this still my newsletter?
AI wrote this. I’m keeping my name on it.

I might lose a few subscribers over this one: I think AI makes me more authentic.
I realize how that sounds. I’ve built a business around helping people understand who they are and put it into words. Authenticity matters to me. So does the possibility that someone has something meaningful to say and never quite finds the language to say it.
That’s why I think AI is one of the greatest things to happen to writing.
Shower thoughts, finished
Imagine I have an idea in the shower. Something clicks. It stays with me all day, but when I try to explain it, the words don’t quite land.
Now imagine bringing that thought to a system that knows my work so damn well we can start developing it right away. It knows the experiences I’ve shared, the books that shaped my thinking, and the connections I keep coming back to.
It pulls in something I wrote before. It asks a question that makes me rethink part of the argument. It suggests language. Some of it lands. Some of it doesn’t, and I say so. Then I add the experience that explains why I care.
Eventually, I have something that says what I mean better than my first attempt ever could.
Every book has a ghost
Every leap in how we create has been met with suspicion. Socrates worried that writing itself would weaken our memory. Then came the printing press, the typewriter, spell check. The internet was supposed to rot our brains. Now we can’t run a business without it.
Getting help with writing isn’t new either. Authors work with editors who cut whole chapters. Publishing houses reshape manuscripts line by line. Leaders hire ghostwriters and speechwriters. Nobody says the book belongs to the editor.
So I struggle with the idea that help from a machine makes a thought less mine, especially when I’ve questioned it, shaped it, and decided to put my name behind it.
There is one real difference, though. A great editor will tell me when I’m full of shit. AI can be far too willing to tell me I’m brilliant. That gap is where most people get into trouble.
Are you even steering?
The problem I see is people accepting whatever comes back.
They let the system supply the perspective, the reasoning, and the conviction. Then they publish it because it sounds impressive.
That’s AI slop. You have to care enough about what you’re saying to reject a beautifully written paragraph you don’t believe.
But there’s a catch, and it applies to me too. My filter is simple: Does this sound like me? Does it feel true?
The better a system knows my voice, the weaker that filter gets. It could hand me an opinion I’ve never held and make it sound so familiar that I nod along.
That’s the risk I take seriously. I could mistake recognition for agreement. So “this feels like me” isn’t enough. I have to ask: Can I explain why I believe it? What would change my mind?
There’s a second risk: a system that knows who I’ve been so well that it keeps pulling me back there. I don’t want an echo. I want room to change, to realize something I believed last year no longer holds up.
That’s what I actually want from AI. Not a copy of my past thinking, but help generating intuition. Sometimes I sense something before I can explain it. Working through it shows me whether there’s actually an idea there, and sometimes it becomes one I wouldn’t have reached alone. AI helps me get there faster and go deeper.
But only if I’m present for it.
Argue with your AI
Take one half-formed idea you’ve been sitting on. Give it to your AI tool with a few lines about who you are and what you believe. Then go line by line. If you can’t explain why you believe a sentence, cut it or rewrite it.
Pay attention to what you reject. That’s where you’ll learn the most about what you actually think.
Full disclosure
I’d been taking notes on this idea for a week. I made bullets, built an outline, and wrote a few paragraphs. Then I opened ChatGPT and said, “You’re my newsletter writer. Here are my thoughts. Help me build a prompt to turn this into a piece.”
It already had context on my voice from thousands of my emails. I brought the idea, the perspective, and what I wanted you to think about. It generated a version. I asked for another because the first was too long and needed more of me in it. Then I edited it.
I probably typed about 35% of the words you just read. My editor, Hemali, wrote another 20%.
Does that mean I own it any less?
If a ghostwriter had drafted this, few people would blink. So what’s really different about AI?
Maybe nothing. Whether it’s a ghostwriter, an editor, or a machine, the help only works if you know yourself well enough to recognize what you actually believe. Most of what we call slop is really people approving words they never examined.
We worry AI will replace our voice. But how much of what we say already comes from wanting approval, looking successful, or meeting someone else’s expectations?
If we haven’t taken the time to understand ourselves, what exactly are we asking AI to preserve?
Perhaps the problem isn’t AI. It’s us.
I stand behind every word here. Where do you land? Hit reply and tell me.
Reply