<

Hello World

Hello, world. Let’s first talk a little about this website: how it was built, why it was built that way, and how I build all of my projects in general.

Let’s start with this: AI. Everyone always talks about AI; it’s all anybody talks about these days. We live amidst an AI bubble. Some would say, “No, it’s not a bubble and you are just behind if you do not use AI,” and I would disagree, partially. Companies like NVIDIA and OpenAI all sort of circle-jerk each other in the circle of compute and selling AI to regular people. But in the world of programming, AI has created a divide between us. Some code ONLY with AI; they flex being “vibe coders” and they want to show off to everyone the constant slop projects they built with AI, put up on GitHub, and forget about to move onto the next thing that they will build.

The position I hold on AI is that, yes, if you do not utilise AI in some way you are kind of falling behind, be it autocomplete or agentic coding. I don’t think there is anything wrong with agentic coding, but it all comes down to the user and how they use AI to write their code. You could be a complete newbie to coding, having never built a project in your life, then with Codex or Claude you build a sloppy website and show it off on Reddit. But people will ask questions, and they will want you to address those questions like a developer would… But you are not the owner of the product; you are not actually the creator. You do not know anything about what your agent made. How would you answer questions about what you have built if you know nothing about it?

And that is where I chime in my two cents: the only reasonable way to use AI is that you are able to read AND understand what it produces. This is how I built this website. Cursor did most of the work; I then read through it, made sure everything was to spec (it was not), found the issues, and addressed them, because I know what it is doing and I can read the code it is writing. I actually own my creation even if it was mostly built by an LLM “intern” who doesn’t really understand anything. It is similar to the real world: a higher-up tells you what to do, you do it, and you show it to them. They point out the mistakes you have made and tell you to go and fix them (or fix it themselves in my case). This is a pattern I have experienced in the real world and in my own projects. AI will never (at least within a close time frame) be able to produce and code like a human can. It does not have common sense; it may write completely stupid code that works for a while but breaks upon an update.

Vibe coding sucks. Do not vibe code, if you are going to use AI for your projects, use it wisely, don’t be stupid.