Lately, I’ve noticed how quickly I search for answers. Not because I need them urgently, but because I’m reluctant to sit with my questions for long.
In the past, uncertainty felt tolerable to me.
When I needed answers, Googling was always an option, but It demanded something from me. My Patience. My judgement, and the ability to search without knowing what I was looking for.
I had to sift through endless pages, follow dead ends and often accept that the “right answer” might never come. I was okay with not knowing. I accepted that the pursuit itself was often enough. That I 'kind of' got the answer I was looking for. Sometimes my struggle was the point.
Through this process, my questions often had time to boil. It gave me a sense of agency. I wasn’t just retrieving information anymore, I was forming a deeper level of understanding by exploring. Thinking. Pursuing.
Now, I ask an LLM prematurely.
I want answers. Fast. This is a pattern I’ve recognized too often lately and I wonder if it’s helping me or breaking me.
Before AI, I could be wrong slowly. Ideas lingered enough to be examined. Mistakes had texture. Now, Fast learning short circuits that deep learning. It reinforces my thoughts and what I already believe.
AI doesn’t punish shallow thinking, it camouflages it. Someone who barely understands a topic can now produce fluent explanations, confident summaries & plausible arguments. The danger isn’t that people stop thinking, it’s that they stop noticing when they haven’t thought. To me, that’s far more unsettling than the ‘AI is making us dumb’ narrative.
When answers arrive too quickly, questions stop maturing. Clarity arrives before understanding has time to form.
Don’t get me wrong. AI has helped me in ways I couldn’t imagine. I use it daily, often as a genuine thinking companion. But I’ve started to wonder: when does help become indulgence? Where do I draw the line? And when?
So I asked myself, What kind of thinking emerges from friction?
Is it the kind of thinking that can live without answers? Or the thinking that tolerates being wrong long enough to change?
This shift matters more than the outcome. It changes how we think. It gives us a way to arrive at our own answers, rather than accepting the first convincing one offered to us.
That’s when it became clear to me: The issue isn’t AI. It’s how & when we invite it to our thinking.
AI isn’t removing thought from the process, it’s rearranging it. It offers answers before we’ve decided what to believe.
Used deliberately, it sharpens our thinking. Used prematurely, it replaces that part that would ideally do the sharpening.
A mirror, not a crutch
AI is best understood as a mirror, not a crutch. It reflects our ideas back to us, often clearer, more confident and more complete than they were in our heads. That reflection can be useful, but it can also be revealing.
When our thoughts return to us effortlessly, we see how unfinished they were to begin with. We mistake fluency for insight, output for progress. Not because AI misleads us, but because it reveals how little friction our thinking passed through.
The question of complacency
When answers are cheap, it’s tempting to skip the struggle that gives us meaning. Verification feels optional, Intuition hardens to belief. Fluency begins to feel like truth.
AI doesn’t take our thinking away. It asks us, quietly, whether we still want to do it.
So maybe the question isn’t whether AI is helping or harming us.
Maybe it’s simpler than that.
What would it look like to delay AI by 10 minutes?
To sit with a question long enough to get it wrong once.
To let an answer feel incomplete before asking for a better one.
Not as resistance. Not as purity.
But as a way of reminding ourselves that thinking still begins before clarity arrives.
AI will be there when we’re ready.
The question is whether our thinking will be.
