Last month, I wrote about what it might mean to cultivate our own intelligence, creativity and judgment in a world increasingly shaped by artificial intelligence.
Since then, I’ve been noticing something about myself.
AI can really frustrate me.
It can produce something in seconds that once might have taken me hours. And somewhere along the way, I think my expectations changed. If I can describe what I want and provide clear instructions (aka prompts), why can’t it provide consistent, accurate results?
Recently, I heard someone on a podcast talk about overhearing her husband interact with AI. He was frustrated with the response and spoke to it sharply—essentially telling it that its work was terrible and that it needed to do better. She jokingly told him he’d better not ever speak to her that way.
It made me think about some of my own interactions—and I wasn’t particularly proud of everything I saw.
Not because AI has feelings that need protecting. But because I am still the person having the interaction.
So I got curious.
What was I actually getting so frustrated about?
I knew the technology was capable of producing the desired result because I had seen it do so before. But I also began to recognize that part of my frustration came from how much work I was doing to get the result I expected—repeating myself, carrying the context of what we had already decided and keeping track of how all the pieces fit together, while trying to figure out why something that seemed clear kept going off track.
The longer the loop continued, the more I just wanted to skip to the end.
And I’m beginning to think that may be about more than technology.
I started thinking about how often we encounter the finished product without seeing much of what produced it.
Imagine this.
You’re watching a cooking show. The host stands behind a beautiful counter surrounded by little bowls. The onion is already chopped. The spices are measured. The herbs are washed and cut. Everything needed for the recipe is sitting within reach.
Then the cooking begins, and twenty minutes later, there’s a beautiful meal.
But someone did a lot of work before we ever saw that counter lined with perfectly prepared little bowls.
So what does a cooking show have to do with any of this?
More than you might realize.
Writing this article began long before I started writing paragraphs. I thought about different ideas. I questioned whether they belonged together. I developed an outline. I looked for research. I discarded some things. I changed my mind about others. Eventually, an argument began to take shape.
Technology can help with many parts of that work. It can make some of it faster and eliminate other steps entirely. But the finished product still doesn't tell us much about what it took to get there.
Recent research from MIT Sloan describes something similar happening inside organizations using artificial intelligence. Researchers found considerable “hidden work” behind successful AI initiatives: people experimenting to discover what the technology could reliably do, reviewing and refining its output with others, and continually adapting what they were building as they learned.
That idea of hidden work stayed with me.
Because what gets hidden isn't only the hours or the effort. It's also what can happen through the work: the trial and error, the brainstorming and collaboration, the questions we learn to ask, the constraints that spark creativity, and the experience that helps us recognize a problem differently the next time we encounter it.
Last month, I wrote about the growing importance of skills such as analytical thinking, creativity, curiosity and resilience. Perhaps this is part of how we develop them—not separately from the work we're doing, but through it.
The process may be producing more than an output. It may be shaping something in us, too.
So, What Am I Doing Differently?
All of this brings me back to where I started: my frustration.
I still get frustrated. I still have deadlines, and I still sometimes think something should take far less time than it does. But I’m becoming more thoughtful about how I approach the work.
I’m becoming more deliberate about breaking complex projects into smaller pieces—not simply because the technology works better that way, but because it helps me understand the workflow itself and recognize more quickly where things are going off track. What am I actually trying to accomplish? What needs to happen first? Where do I need to think, research or experiment? What can I hand off? And what do I need to learn before I can move forward?
I’m also giving myself more permission to experiment. Research may lead nowhere. An idea may get deleted. An entire approach may be scrapped. That doesn’t necessarily mean the time was wasted. Sometimes discovering what doesn’t work is part of developing the judgment to recognize what will.
At the same time, I’m more thoughtful about my time. I’ve discovered that I reverse engineer things constantly—starting with the outcome I’m trying to achieve and working backward to understand what has to be true to get there. I was doing this long before I knew there was a name for it.
I’m also learning to put boundaries around the time I spend experimenting. At some point, I need to step back and assess whether I’m making progress, need to change my approach, or simply need to walk away for a while.
Maybe that’s part of what I had stopped seeing when I became so focused on getting to the end.
The outcome matters. And I want to get better at celebrating more than the outcome.
The ideas that didn’t work but led somewhere else. The creativity that emerged from a constraint. The curiosity that kept me experimenting. The courage to change direction when something wasn’t working. Even the growing ability to recognize what deserves more of my time—and what doesn’t.
When I look back at some of the things I’m proudest of, the finished product tells only a small part of the story. What it doesn’t show is everything that went into becoming the person who could create it.
Maybe some of that deserves celebrating, too.
Sources
MIT Sloan School of Management. “Why Some Organizations Turn AI Experiments Into Business Value While Others Quietly Fail.” September 9, 2026.
