Why Prompting Is Really About Understanding Ourselves

Not long ago, people thought “prompt engineer” would be a niche role — like getting a degree in MySpace page design: looked promising until the world moved on. A handful of specialists who knew the secret phrases to coax good answers out of AI.

That view was too narrow. The reality is much bigger: we’re all going to be prompt engineers. Managers, designers, analysts, writers, students. And yes, machines too — prompting other machines in a chain of interactions to complete tasks. Prompts won’t be a side skill. They’ll be how we work.

I often describe good prompts as “thinking about thinking.” They aren’t random lines of text. They reflect our mental models — the frameworks we’ve built over a lifetime.

But many of us don’t know why we think the things we do or how those damn thoughts got in our head. Our brains are neural networks — yes, just like the ones used to build LLMs. From the moment we’re born, they’re shaped by experiences, by what we read, by our successes and failures, by our families, politics, and culture — most of all by our parents, who somehow manage to sneak their opinions into our lives like unrequested software updates. And then there are jingles — seriously, who let “I’m lovin’ it” take up permanent residence in there? All of that becomes our “training data”.

Here’s a simple illustration. Take two friends: one grew up watching his parents lose their home to failed business loans. For him, loans are reckless and unnecessary. Another had a father who built a successful career in finance. For him, loans are just one of many financial tools for growth and opportunity. Both see their views as obvious. Neither fully recognizes the beliefs and experiences that shaped them.

That’s what we’ll be encoding into prompts — our hidden assumptions and biases. And right now, many of the prompts we see are overly simple: “Summarize this page. Keep important points.” What feels essential to one person might be background noise to another — like someone’s favorite song becoming your least favorite earworm. They work, but they don’t capture the depth of how we think, or why.

This is why psychology will matter more than ever. The hard part won’t just be writing a clear instruction — it will be understanding ourselves well enough to know what we’re actually asking the machine to do. In other words, we’re really teaching it to borrow our way of thinking and reasoning — warts, quirks, biases and all.

At the same time, the technology will give us something back: time. Machines will take on the structured, repetitive tasks, the summaries, the status reports, the endless small requests that eat away at our days.

That leaves us with more space for what can’t be automated — at least not fully. Human conversation. Body language. Shared laughter. The subtle moments of connection. But even as technology advances, 55% of communication is still body language. Machines may approximate it, but people will always value connection with other people.

So no, prompt engineering isn’t going away. It’s becoming universal — like the computer or email before it: not a specialty, but the everyday tool we all lean on to get our work done. We’ll all be writing prompts, and machines will too. The real question is: how aware are we of the thinking behind those prompts?

The better we understand ourselves, the better we’ll guide the machines. And the more we let them carry the routine, the more time we’ll have to focus on what’s most human.