The new default
I was debugging a query last week. Saw a familiar error. Knew exactly what was wrong - a date filter wasn't casting correctly. Something I've fixed dozens of times.
But instead of just fixing it, my first instinct was to paste the error into an AI tool. The fix took ten seconds once I actually looked at it. Writing the prompt would have taken longer.
That keeps happening. A quick SQL filter. A config change I've made before. A formula I could write with my eyes closed. Stuff I already know how to do. And instead of doing it, I'm reaching for AI out of habit.
I use AI constantly. It makes me faster at a lot of things. But there's a growing category of tasks where the detour is actually slower than just doing the thing.
Skipping reps
Every time you outsource a simple task to AI, you skip a rep. Reps are how you build intuition.
Intuition is the thing that lets a senior data person glance at a dashboard and say "that number looks wrong" before anyone runs a test. Pattern recognition from years of doing the boring stuff. Seeing the same errors. Writing the same joins. Noticing when something doesn't feel right.
You don't build that by describing problems to a chatbot. You build it by doing the work yourself. And if we stop doing the simple stuff, we stop building the thing that makes us good at the hard stuff.
A simple filter
I've started asking one question before I reach for AI: do I already know how to do this?
If yes - just do it. No prompt. No context window. Just the thing, done.
If "sort of, but I'd need to look something up" - that's where AI is great. Better than Stack Overflow. Faster, more contextual, less junk to sift through.
If "I have no idea" - obviously, use the tool.
The problem is the first category. The stuff you know cold. That's where defaulting to AI costs you the most. Not just time. It trains you to not trust your own judgment.
Start from the human version
Before you prompt, think. Do you actually need help? Or are you just reaching for the tool because it's open in the next tab?
Your own brain is often the fastest tool available. And the more you use it for the straightforward stuff, the sharper it stays for everything else.
The GPS problem
It's like using GPS for a route you drive every day. You'll get there. But at some point you realize you couldn't describe the route to someone else. The knowledge never stuck because you never needed it to.
AI will make most of us faster overall. But inside that speed gain, a specific skill is quietly disappearing - the ability to just do the simple thing without help.
The people who will be best at using AI are the ones who know when not to. Who recognize the easy problems instantly. Who can tell the difference between "I need help" and "I'm just being lazy."
AI intuition isn't knowing how to use AI. It's knowing when not to.