Creating balintg.com
This is the very first post on this website. Right now I am sitting by my desk in Budapest, writing, thinking about the next words and sentences. This post is going on my own website. The website which I know is simple - by design - but it’s already finished. Might not be perfect, but it is finished. And by finished I mean it has a good looking UI (I hope we won’t argue on taste), has every text in place, no lorem ipsum, and no TO DO anywhere. It is already deployed under my own domain, it is working, it is reachable from anywhere. All I have to do left is fill it with content, and use it how I want.
12 hours ago all I had was an empty folder, and an idea from the day before. It is not a secret: this is my first “vibe coded” project, and I am pleasently surprised. I started experimenting with Claude Code yesterday. Paid for the pro subscription, installed from command line, then started the first course on Anthropics website: “Claude Code in Action”. It is a short, but helpful course which goes over installation, the basic functions and hooks. The course is practice-oriented, and offers an example project to try Claude on. However you can also use any other programming project, just like I did with merker4.com (more about this later). I was impressed by the functionalities presented, but I also felt a bit dump: like many others, I also assumed that vibe-coding is just mindlessly throwing instructions at Claude, then hoping it generates code that works. The course showed how Claude Code is more than just an LLM you can chat with. Now I see there is depth to this.
After finishing the course I asked Claude to analyze merker4.com, think and plan to improve it. It found and solved bugs that I don’t think I would have found, like ever. At that moment I knew that I can trust it with a basic project like balintg.com, especially because this one doesn’t handle any sensitive data. I went to bed knowing that I will have a productive day after waking up.
My philosophy before
Before any vibe coding I always chose the harder ways. I believed in learning by doing, but I am unfortunately also a perfectionist. Since the day I was born I always make sure I am ready to do something to a high standard before I do it. Web development was not different. I learned the basics during university, but since my profession is something very dfferent, I don’t really have meaningful experience with websites. I can see how webdev is powerful: one person can create something valuable without big investments, and share it with the world instantly. I’ve had the urge to do the same for a long time.
I researched tech-stacks, thought a lot about what I would like to create (then had different ideas a few days later), then started learning. However without a clear direction and belief it was hard to stick to it. Especially because even if I had the baic knowledge of the languages used, I did not have the confidence. I didn’t have confidence because I didn’t have proof: I have never went as far as deploying and maintaining a website. It was always just a short course, a quick demo for university or something along those lines.
With merker4.com it was something else: I decided to follow Corey Shafers Django course on youtube. This is the course that even after many years and a few django versions later is still coming up in searches when you try to find the best way to start Django. During the course you can put together a full blog website and deploy it. I decided to go along with it, but also tweak it a little, and modify it so that it can serve as my professional website.
Why I am writing all this: it took me a long time to finish that course, learn the concepts, and create an own project that I can use. It took me more time than I would like to admit. During this time I only used chatgpt as if it was a more experienced friend who can answer questions, but also I didn’t want to be disturbed everytime a problem comes up. I tought this is the way: grinding, learning by doing, and being proud about not using every advantage possible. In the end I finished and deployed merker4.com, and went as far as creating a separate email address for it (like a pro).
What I see now
Deploying my first website made me see that I am capable of it. However it also showed how many different things can come up during development and also in deployment. Starting any other project from scratch in that little freetime I had sounded exhausting. As if I had to have the greatest idea because it took months to finalise them.
But the times of learning one single tool for months or years is over. It is not affordable to be stuck in one place while everything is changing so fast. The way electricians don’t have to explore electrons to wire the lamp in your living room, or bakers don’t have to get into agriculture to bake bread - we also don’t have to write every single html line ourselves.
The work shouldn’t be the goal: with work we should create value.