Getting into every tech stack

Life has always been about learning, and my tech journey was going smooth before the agentic AI wave hit. When all these agentic AI tools came out, the coding world changed completely. Things aren’t the way they used to be. Life was moving well — I was learning and doing a lot of things.

But when the agentic AI era came, everything changed for me — and I’m sure for you too. At first, I tried to stay away. But with how fast it was developing and how everything shifted in that direction, I felt like if I didn’t go with that momentum, maybe I’d be left behind. Maybe there wouldn’t be a place for me anymore. So I jumped into vibe coding. Luckily, I got access to top-tier AI agents at the time — Opus 4.6, Claude 5.3 — the peak ones. I got hooked on models that made me feel like, “Man, come on, you have to get used to this and move on, otherwise you’ll fall behind.” And I did. I went along with where everyone was headed. But deep down, I wanted to be that guy who didn’t give a damn about AI and just stayed true to himself.

And it’s not just me. Look around the internet — our favorite StackOverflow traffic, everything is down. Even Linus Torvalds came out and admitted to vibe coding. DHH made Omarchy and said he couldn’t have come this far without AI.

Talking about projects — I was doing everything. From complex backends, mobile apps, web apps, desktop apps, to even developing an AI/ML model during this period. I implemented things in projects that I never thought I could do this fast. I got involved in GPS stuff for delivery projects on web and mobile. GPU things, Redis, WebSockets — I was everywhere. And at that time, I was feeling really GREAT. Like I didn’t need anything or anyone.

But I ignored so many things. I didn’t go out for 8 or 9 months, I think. I was completely lost. I built desktop applications and all. I figured out Docker, CI/CD — deployed all my projects on my own. All I’ve gained is a pile of resources. But now I feel like this isn’t what I should be doing. I hate code nowadays. I hate AI. I hate AI-generated things. And yet I still have to deal with it.

Not only that — I was deep into Linux stuff too. Exploring TUIs, open source, spending more time in the terminal than in other apps. TUIs became a genuine love. Learning was never the problem for me. Before, I used to learn fewer things but for longer periods. Now I’m learning more things, but I don’t know how much of it will stick. As I know — and as you probably feel too — there’s a big gap between what I’m actually learning and what I feel I’m learning.

Now I’m trying to step away from all this. I will never use agentic AI for learning projects or for building my knowledge. And I recommend you do the same. If your goal is to be the best frontend developer, don’t use AI for frontend. I don’t mean don’t use it at all — just don’t use agentic AI. Take ideas from generative AI, but implement them yourself. That way, if someone asks you to do something, you won’t need to ask anyone — you can do it on your own. Same goes for everything.

But right now, if you’re diving into lots of things, you’ll have too many to choose from. One reality for me is that nowadays, code is cheap. Code is so cheap. Anyone can create it with a prompt. You can find millions of lines of code. And if you can’t find the code, you can just screenshot what you want, ask AI to write it, and take a look. Take ideas. But don’t depend on that code — especially for system design and backend architecture.

But I’m literally laughing here, because I’m teaching you to do this and that, while I’m also a victim like you. I’m trying hard to get out.

So here’s my plan. If I get a project related to frontend, I’ll use agentic AI for it — because I don’t want to see myself as a frontend developer anymore. I don’t want to waste more time there. And this is where I found the real problem. I was quite good at frontend in my prime. But now, my friend, I can barely design one thing on my own. I have to ask AI. (Although I can still put together a decent design because I know a few things — typography and all that.) But in other areas, I feel bad. You know why? Because AI has changed how we think in the programming world. Before, we had principles and an architecture in mind, and we stuck to them. But nowadays, we don’t stay with our own thinking. If the AI says, “Hey, don’t do this, do this instead, this will be better,” we shift immediately and believe it. And that’s the issue, my friend — you and I are both facing this. We can’t go beyond our imagination because AI is imagining for us. Because of this, you stop imagining. The ability to imagine is shrinking. And this is a problem across all of life — not just tech, but everywhere. It’s hurting us.

For me, I’ve had fewer real connections lately than before — less time spent with people. Imagine if you had a friend or assistant who did everything for you — you just give them a hint and they deliver. Can you go beyond that? No, you can’t. That’s the issue. And we need to step out of this.

As I said, I did almost everything. I wanted to learn everything — servers, backend, Linux, data, networking, and all. But can I remember what I was thinking at the beginning? Barely. I can’t remember anything. And you too — remember what your plan was when you started working on that project? Do you remember your plan? And look at where you are now. Is this good? Is this necessary?

So think about this. And yes — if you somehow slip back into using agentic AI while learning new things and end up in that same loop, turn off the internet. Read from a book. Code from a book. That’s what I’m starting to do, and I’ve already learned a lot.

We are all consuming tech instead of owning it.

TLDR: AI is allowed for: syntax lookup, boilerplate, docs. AI is NOT allowed for: architecture decisions, algorithms, core logic, debugging.

Thanks.