What Efficiency Takes From Us
On AI, and the human practices worth protecting.
I use AI every day in my personal life and at work as an engineer.
At work, we’re strongly encouraged to use it, to the point where it has become almost a checklist. Individuals and teams that leverage it to solve a problem, are not only recognized, but rewarded.
And I totally get why.
I’ve witnessed AI models complete convoluted tasks in seconds, that would have otherwise taken hours, days, or even weeks.
I’m not against AI. But I have started to wonder about the cost of optimization and efficiency. Because every tool that makes something easier, changes how we practice something. And in that practice, we may not know what we are losing.
Efficiency Has a Shadow
I was fortunate enough to purchase a vehicle with self-driving capabilities. It has been a game-changer, especially on long commutes, or after long days when I’m mentally exhausted. I have absolutely no regrets, but it has created an interesting problem:
My driving has actually gotten a little worse.
My wife was quick to point this out at dinner this past weekend. I struggled to keep the SUV cleanly between the lines.
Although technology had made one part of driving easier, it had also quietly reduced the reps I was getting behind the wheel. And this is the part I keep thinking about.
Efficiency has a cost. And efficiency also does not always announce what it is taking from us. Sometimes it simply removes the practice.
The Practice We May Be Losing
While AI can answer an analytical question in a matter of seconds, sometimes the real work is in actually asking the right question, or actually understanding the situation.
That kind of discernment does not come from speed alone. It comes from staying close enough to the work to understand what is actually happening. It comes from context, curiosity, repetition, and even friction. The slow process of noticing patterns that are not always obvious at first glance.
I recently took on a new role at work. AI has already helped me make sense of hundreds of data points. It’s helped create patterns. It’s helped create signals. It’s also suggested actions to take. It’s been incredibly useful.
But the output isn’t the same as the work.
I still have to meet with teams. Build relationships with teams. I still have to understand the history, the personalities, the frustrations, the excitement, and the undercurrents that do not always show up cleanly in a report.
And while AI has helped me prepare, it cannot build trust for me with other colleagues.
It can point me toward a possible issue. But it cannot walk into the room, listen carefully, and understand what people are really trying to say.
It can flag a risk from thousands of data points, but it can’t give me the words, in the moment, to express empathy to a team member having a rough day.
That part still requires practice. That part still requires being a human.
And if we outsource too much of the thinking, we may eventually lose the reps that help us know what matters.
What We Still Have to Practice
This is not an article against AI. In fact, I’m for it.
The friction AI can eliminate is tangible. But I don’t think our goal should be that all friction is removed. Some friction actually forms us.
Frederick Douglass once said:
If there is no struggle, there is no progress.
He was speaking in a much deeper context than technology or efficiency, but the principle still gives me pause.
Progress has always required something from us such as attention, effort, courage, practice, and failure. And if we’re always searching for the most optimized solution, we could be robbing ourselves of the very thing progress requires: grit, perseverance.
There is no perfect way to move forward, but one thing is for certain. Technology will continue to evolve. LLM models will get smarter. Tools will become faster, more capable, and harder to ignore.
So maybe the stand we will need to take is not whether we use them. We will use them.
The real stand is deciding what we are willing to hand over. And what we are not willing to lose: presence, imperfection, understanding. Or the ability to struggle with something long enough for it to shape us.
I want the help technology has to offer. I just don’t want to lose the parts of me that still need the practice.
Both / And is about making decisions you can live with.
Subscribe for weekly reflections on decisions, doubt, and living aligned.
I agree. Understanding rarely arrives all at once. It grows through the slow accumulation of patterns.
I love AI and use it daily, but yes, I am worried that some people, especially young ones (god, I swear I am not THAT old), will use it as a shortcut and will loose critical thinking in the process. And that is a scary cost for me. We read books and were asked to write essays about main characters - what they thought, what they intentions were - not because teachers cared what characters were actually thinking about, but because the process of finding out is what mattered, that's where development lived. AI drastically cuts execution time, but I have concerns when it comes to handing over tasks that require actual thinking. I am afraid that cost might be too high.