I’ve been trying to learn Go. Writing small programs, doing coding challenges, trying to build something real.
And I keep failing.
Not because Go is hard, but because AI is always right there.
The Pattern I Can’t Break
The dangerous part isn’t that AI gives answers.
It’s that it gives them before you’ve struggled enough to need them.
It starts the same way every time. I open up a challenge, something simple like rotating an array or parsing input. I try for a bit, hit a wall, and almost without thinking, I reach for AI.
Just a quick check. Just a hint. Just one answer.
It Feels Like Progress
There’s a dopamine loop here that’s hard to ignore.
I paste the problem in and get clean, working code back. When I read it, it makes sense. I tell myself I understand it, then move on.
Another problem done. Another small win.
A green checkmark. A passing test. A quick hit of “I’m getting better.”
But it's not real, I didn't earn it, I didn't learn anything.
But I Didn’t Learn It
That’s the part that’s been bothering me.
I didn’t struggle through it. I didn’t sit with it long enough to understand why it worked. I skipped the part where my brain actually builds a model of what’s going on.
And somehow, it still feels like progress.
The Moment It Hit Me
I tried to write something in Go without AI. Nothing complicated, something I had already done”before in another language.
And I froze.
I knew I had seen the solution. I knew I had written something like it. But I couldn’t recreate it.
Because I never actually owned it, and with languages like Go, ownership matters.
It’s not just concepts, it’s the muscle memory.
Typing if err != nil a hundred times isn’t exciting, but that repetition is how the language actually sticks.
The Chip Addiction
Using AI feels like opening a bag of chips. You tell yourself you’ll just have one, one answer, one suggestion, one shortcut.
But you don’t stop at one.
Because it’s easy. It feels good. It keeps you moving. So you grab another, and another, and before you realize it, you’ve finished the whole bag.
You solved five problems. You feel productive.
But if someone asked you to cook the meal yourself, you couldn’t.
What Learning Used to Feel Like
Before this, getting stuck meant something. You’d read docs, try things that didn’t work, and sit with the problem longer than you wanted to.
That friction wasn’t a problem.
It was the learning.
The Part I Didn’t Have Words For
I didn’t have a name for this before.
But that uncomfortable part, the getting stuck, trying something wrong, figuring it out, that’s actually the point.
It’s where the learning happens.
When I skip that, I’m not saving time, I’m skipping the part that makes it stick in my head.
What It Feels Like Now
Now the friction is optional, and I keep choosing to skip it.
Because why struggle for 20 minutes when I can get the answer in 10 seconds?
But that tradeoff is starting to show.
The Real Cost
Speed without understanding is fragile.
AI is making me faster, but it’s also making it easier to avoid thinking deeply.
And when I step away from it, I can feel the gap. The understanding isn’t there. The instincts aren’t there. The confidence isn’t there.
What I’m Trying Now
“I’m not allowed to use AI until I’ve failed properly.”
My Learning Protocol
- The 15-Minute Rule: No AI for the first 15 minutes of being stuck
- Docs First: Read the official Go docs/spec before asking AI
- Write It Wrong: Attempt a full solution even if I know it’s broken
- The Socratic Prompt: Ask for hints, not answers
This isn’t about avoiding AI, but about putting it in the right place in the flow of learning.
To reiterate
“I’m not allowed to use AI until I’ve failed properly.”
This doesn't mean I’m quitting AI, but I have to try something first, because if I reach for it too early, I don’t learn.
So I changed one rule.
I write the solution I think might work. I run it, even if I know it’s wrong. I follow the error instead of avoiding it.
Only after that do I open AI.
Sometimes I still ask it for help, but I’ve changed how I ask. I don’t ask for the answer. I don’t ask for the code.
- I ask it to explain the concept.
- To point me in the right direction without solving it for me.
And when I do, I don’t ask for the answer. I ask:
- “What am I missing?”
Or I’ll say:
- “Explain this like I’m trying to figure it out, not like you’re trying to solve it for me.”
That shift matters more than I expected, because now I’m comparing my thinking to the solution instead of replacing it.
Where I’ve Landed
AI didn’t make learning worse.
But it can make it optional.
If I’m not careful, I’m going to get really good at finishing problems… without ever actually understanding them.
If you’re learning something new right now, and AI is part of your workflow, I’m curious , do you feel this too?
Top comments (0)