AI has recently been teaching me how to do a job I no longer hold, but one that defined a large portion of my professional life. I wasn’t searching for this lesson; it simply emerged from a prompt.
Have you ever heard of the “Peter Principle”?
Developed by Laurence J. Peter, the principle observes that in a hierarchy, people tend to rise to “a level of respective incompetence.” The logic is simple: employees are promoted based on their success in their current role until they reach a level where their skills no longer translate to the new responsibilities. They eventually plateau at a position where they are no longer competent, and there they stay.
I lived this.
Around 1983, I blundered into a trade school where I learned to program in BASIC, IBM Assembler, RPG, and COBOL. For someone who grew up seeing computers only in sci-fi shows, it was the place I belonged. My career took me from a third-shift operator at a hospital in Ft. Oglethorpe to a long tenure with the City of Chattanooga.
Because I was good at the technical work, I was promoted. I climbed the ladder all the way to Deputy Chief Information Officer. After a few years, I realized I missed the craft, and I asked to go back to being a programmer. We made that happen, but as management changed over the years, I was once again asked to step into a leadership role.
From the outside, I seemed like a capable manager. In reality, I suspect my teams simply succeeded despite me. Delegation is the cornerstone of management, and I struggled with it. I existed in a constant state of tension: I feared micromanaging my team and stifling their creativity, but I also feared that without rigid control, the results wouldn’t meet the mark.
Then came the AI.
Recently, I was crafting a system prompt to create an AI editor for my blog. I spent a significant amount of time detailing exactly what I wanted:
I asked the AI to create a prompt template for an editor, providing a markdown example of my existing SEO template. I instructed it to keep specific headings, use provided bullets as examples, and feel free to add necessary sections—provided it explained why any existing sections were removed.
As I hit “Enter” and waited for the response, I looked back at what I had written. I had a sudden, sharp realization: This is exactly how I should have been managing people.
In that prompt, I had provided a detailed specification of the “what” and the “why” without dictating the “how.” I had set clear boundaries and defined success, yet left ample room for the performer—human or AI—to bring their own unique value to the end product. I had achieved the perfect balance between guidance and autonomy.
The results were fantastic; in fact, I used the resulting prompt to help edit this very post.
I’m not sure why it took a Large Language Model to teach me this, but I wish I had discovered this clarity sooner. I wish I had given my teams these kinds of clear, empowering instructions instead of hovering or withdrawing.
Live and learn. Adapt or perish. No matter when the lesson arrives, I’m just glad it finally did.
