I should start with a confession: Looking back from where I stand today, I realize that at my core, I am a terminal programmer. In university, I took algorithm courses, learned C++, and reached a point where I could translate anything I designed mathematically in my mind into code. From Fibonacci sequences to cryptography, from matrix operations to numerical analysis, I was solving everything on a black terminal screen with white text. But I had a small problem: I didn’t even know how to show a simple button to a user on the screen.

In this article, I want to tell the story of my journey from that dark terminal screen to my current web projects, the AI systems I worked alongside during this process, and how my “Vibe Hardcoder” identity came to life.

Terminal Days: C++, Python, and the Obsession with Shorter Code

Everything started with C++. The language I learned in algorithm classes fit perfectly with my mathematical way of thinking: strict rules, clear syntax, and the feeling of having everything under control...

Then I taught myself Python. My first reaction was: "Wow, C++ feels so heavy." Something that took fifteen lines in C++ could be done in three lines in Python. That sense of flexibility was completely different.

Over time, something else took over: an obsession with shortening code. I started compressing my own C++ code as much as possible. Then I moved into Python, using dictionaries, list comprehensions, and lambda functions to compete in code golf challenges. Testing how much logic I could fit into a single line felt perfectly aligned with the optimization spirit of mathematics.

During that time, my relationship with code changed. It was no longer just about making it work—it was about making it work in the shortest and most elegant way possible.

First Disappointment: “You Can’t Code with This”

When the AI wave first began, I was excited like everyone else. I asked ChatGPT to fix a 150-line piece of code. After a long and frustrating process, it produced something… but the result was terrible.

"You can't write real code with this." I said, closed the tab, and moved on.

At that time, AI was just a “cool toy” for me.

Until I met Claude.

“I Found the Best Coding AI”

Back then, Claude wasn’t directly accessible from my country. I connected through Poe. In our first serious technical conversation, I noticed something: this tool didn’t just write code—it understood it.

It explained why it suggested certain architectures, offered alternatives, and spoke directly to my analytical mindset.

That day, I told myself: "I found the best coding AI."

And I wasn’t wrong.

Token Limits and Timeouts

But there was a problem: limits. Tokens ran out, sessions expired, and I had to wait. Being forced to stop right in the middle of a flow was incredibly frustrating.

I couldn’t just sit still. While Claude rested, I had to keep working.

I tried Bard as an alternative… it was terrible. It generated code, but had no idea what it was doing. After experiencing Claude’s elegance, it was unbearable.

Then Gemini arrived.

I approached it with skepticism, but I was shocked. It was fast, consistent, and in some tasks, almost flawless.

That day, I said:

"I found my second hand."

And at that moment, I realized I wasn’t using a single AI anymore—I was building a system.

From Terminal to Web

In March, together with Claude, I built my current website from scratch. No WordPress. No templates. Everything was written manually.

As someone used to a black terminal screen, I found myself dealing with routing, session management, and database optimization for the first time in a real sense.

The only thing I lacked was the web paradigm—and Claude filled that gap perfectly.

It didn’t tell me to take a 40-hour course. It gave me exactly what I needed, right when I needed it.

Then things grew.

CheckMate was born.

The System Became a Team

Over time, this evolved into a natural division of labor. I design architecture with Claude. Gemini runs in the background.

One day, I received an email: "Your Claude API access has been suspended."

I panicked. Opened the system… everything was still running.

Because Gemini had taken over the load.

That’s when I realized: this was no longer just a tool. It was a system.

mod_security and Reality

The most valuable lessons came when everything broke.

mod_security was blocking even the word “SELECT.” The solution: Base64 encoding.

It worked… until I saw encrypted text in my own admin panel.

"Did we get hacked?"

No.

I was just afraid of my own code.

AI Is Neither God Nor Devil

What I clearly understood is this: AI is a tool.

It does nothing on its own. But used correctly, it expands your limits.

CheckMate is the best example of that.

Is the Code Beautiful? No.

But it works.

And sometimes, that’s enough.

The Future

When I look back, I see this:

A person trying to shorten lines of code in a terminal… now building systems.

And the most important thing I’ve learned is this:

Big problems are solved by breaking them into smaller pieces.

The rest… AI handles.

In the end, I’m still a mathematician from the terminal.

Just now… a bit of a Vibe Hardcoder.