I Tried Building a Blog with AI and Discovered What It Doesn’t Teach You
Web DevelopmentCarrer & Soft SkillsArtificial Intelligence
Last updated 7/2/2025, 12:08:00 PM5 min read

I Tried Building a Blog with AI and Discovered What It Doesn’t Teach You

The real journey behind my blog in Next.js: mistakes, AI, and rebuilding.

There’s a price I paid for focusing my studies and my experiencies exclusively on ServiceNow. Emerging web development technologies weren’t really necessary in my daily work, since the platform is largely based on Low/No Code solutions.

Then, in early 2024, with the rising popularity of artificial intelligence, I had an idea that (although it now feels a bit cliché) changed my path: learning by asking AI. I decided to create a blog that would serve as my portfolio, my knowledge base, and my testing and learning playground.

I opened a new project in Figma and a diagramming tool, and began to think. That’s my favorite way to kick off any planning.

My goal was this:

Since ServiceNow mostly used AngularJS, JavaScript, HTML, and CSS in its portals, I had a new challenge ahead: getting up to speed with modern web development technologies.

My first target was React.

Prompt after prompt, I started building my blog and learning the tech at a very surface level. In just a few days, the project was nearly ready. That’s when I asked for feedback and ended up receiving some incredibly valuable mentorship (for which I’ll be forever grateful) from one of my greatest friends and mentors, Yuhri Bernardes.

He explained the current landscape to me: component lifecycle, SSR vs CSR, and most importantly, how for a blog-oriented project, Server Side Rendering (SSR) was essential for SEO optimization.

So, a new obstacle appeared: migrating my React project to Next.js, where SSR is easier to implement.

Once again, I went with what seemed like the most efficient path: ask AI, copy, paste, test. Got an error? No worries, AI would fix it.

Until… it didn’t.

AI started to get confused, and I couldn’t fix a critical lifecycle issue. It was frustrating. While you feel like you’re in control, AI’s limitations remind you that, deep down, you don’t actually understand what you’re doing.

That’s when it hit me.

I couldn’t depend entirely on AI, or keep going with this “vibe coding” approach, if I really wanted to build something solid, secure, and high-quality.

So I started over. From scratch.

I studied React and Next.js the old-school way: reading official documentation, watching technical videos, digging into StackOverflow, making mistakes, fixing them, and consolidating what I learned.

After a few weeks and combining the knowledge I’d gained with all my experience in ServiceNow Service Portals, I reached the result you now see live. And it was so much more rewarding.

I subscribed to GitHub Copilot, which helped a lot with suggestions and auto-completion, but this time, I was the brain behind the process. AI assisted me, it didn’t lead.

The biggest lesson from all this was realizing that AI is not (yet) the core of progress, but rather a powerful tool, an extension of our consciousness.

We need to be careful not to overlook our humanity in this process. Because it’s exactly that (our human side) that will solve the problems when everything goes wrong: when there’s no power, no internet, or when a system crashes.

AI does significantly reduce costs in the short term, but it still can’t fully grasp the broader context or design optimized architectures that take into account the true cost, all risks, and at the same time deliver a high-performance solution.

We can’t build our creations, decisions, or businesses entirely on something that depends on so many external and unstable variables. Not yet.

That doesn’t mean AI won’t someday be as reliable as electricity or the internet. I believe we’re headed in that direction. But until then, it’s our human capability that upholds what we’re building.

To sum it up

To truly build something with solidity and quality, you need more than copy-pasting AI results. You need experience, failure, deep comprehension, and most of all, a human presence in the process.

If you’re just starting out or thinking about getting back into development, know that using AI is great, but don’t forget to develop what matters most: yourself.

Solid knowledge still is, and always will be, the foundation of true innovation.

I hope you enjoyed this article. If you can share the blog, I’ll be grateful. That way it keeps growing, and we grow together!

See you in the next post!