Out of 16 years in school, and thousands of lessons, there’s only one I remember word-for-word.  

It was one sentence long. 

It was taught to me by Professor Owen Jenkins, the most feared English professor at Carleton College.  

I'd heard rumors of his harsh criticism since Freshman year. He (allegedly) shredded one student’s paper and handed it back in a paper bag. He (rumor had it) ripped a chunk off somebody else’s paper and returned it with the comment “my cat wouldn’t swallow this.”  

AI is good at a lot of things. But compelling, original writing with a unique point of view still requires a human. 

Don’t believe me? Just ask an AI.  

I did, and ChatGPT was refreshingly honest. It warned me that without a human partner, it can sound generic, offer shallow insights that aren’t grounded in lived experience, and deliver made-up facts with confidence. All good reasons to be cautious. 

Here's one more, from the human perspective: if you're serious about your writing, and you want people to engage with your ideas, don't outsource your original opinions, insights, and voice. Huge mistake. If you hand those over to a robot, you're missing the point of writing. 

I recently took a side quest in my "AI Mastery Project" to build a Project Management system in Notion with ChatGPT’s help. That taught me so much about promptcraft, I turned the postmortem into a blog post. 

However, the experience sparked several insights about another subject that warrants a deeper examination: the Human–AI partnership.  

After wrestling with this subject, I believe I'll still be writing about it long after I've mastered promptcraft. If we assume AI will be a permanent part of our future at work, then we'll have to get skilled at directing, and working with, non-human partners.  

Sometimes the right conversation finds you just when you need it. 

It recently happened for me at a winery, where my wife and I wound up sitting next to two brothers. Our small talk eventually turned to AI, at which point it stopped being small talk. 

I shared how I write with AI: I let it prompt me, interview me, organize my thoughts, and give sound editorial advice. But the moment it tries to write a passage in my voice, I push back.

Ever since I started using Claude and ChatGPT, I've been fascinated by their differences. I use ChatGPT to organize my life, to build systems, to give advice, for prompt engineering -- my left-brain work.

But I always sensed Claude was the more sensitive writer. As one friend commented a few months back, "The ghost in Claude's machine feels more human."

A couple weeks ago, I kicked off a big project – master the AI tools reshaping the creative landscape. I’d written a V2MOM (Salesforce-speak for “master plan”). I’d built a lesson plan for week one. I had dozens of ideas for blog posts and plans to catch up with dozens of people. 

And I didn’t know where to start. I was paralyzed.