I've said this before: if you keep using AI, and don't write code yourself, you'll lose that particular skill.
But I think I was not entirely correct in that conclusion! 🤔
What you really lose if you keep using AI for everything, is not your skills, but your ability to focus.
Being able to focus, being able to use your brain deeply, i.e. being able to enter the flow state, is the true high level skill that empowers human ingenuity in every knowledge base field.
If you lose that, what do you actually have?
I hope you'll give it a thought! 🥰
Top comments (7)
When I am using AI, my focus is over the top 😂 It takes a serious amount of effort (and google search) to get what it's talking about.
Instead of letting the AI take the decisions for me, I explicitly ask it to tell me the available approaches first (i turn off agent mode so it can't touch the files while doing it). Then I dig deeper and choose the right approach, as well as add many more details to the prompt. Then I turn on agent mode and let it cook. I think I have started making fewer reverts since I adopted this workflow (I am not sure though), and learn much more.
Interesting!
However, that kind of focus is the focus of a teacher who is making sure the students are studying. What I'm talking about is the kind of focus that the students need to deeply understand the study materials. Makes sense? 🥰
Ah 😅 I mean, the teacher also has to think about the material to ensure the students are studying, right? Did i get it correctly? Anyway, is this too much analogies right now?
No, the analogy is still correct. The type of focus and the type of thinking is still very different 🥰
Yes, that's correct!
The real risk isn't forgetting syntax — it's atrophying the debugging instinct. When you write code yourself, you build a mental map of failure modes: off-by-one errors, race conditions, memory leaks, the specific way this ORM lazy-loads relations. AI gives you the "happy path" implementation. But when production breaks at 3 AM, you need to navigate the failure, not just generate another plausible snippet.
I've seen juniors who can prompt a decent REST endpoint but freeze when the latency spikes because they never felt the N+1 query problem in their fingers. The muscle memory isn't typing
fetch— it's recognizing the smell of a missing index.AI shifts the skill from "writing code" to "reviewing code at scale." That's a valid skill, but it requires more domain knowledge to verify, not less. If you skip the struggle of writing the first draft, you never earn the intuition to critique the second (site: labagent .tech)
Developers before AI used to write all the code themselves, lol Now developers use AI and code a little on their own. Hahaha, old-school developers were built different! 🤣🔥
Some comments may only be visible to logged-in visitors. Sign in to view all comments. Some comments have been hidden by the post's author - find out more