Hey, I’m a non-tech creator.
I don’t come from a computer science background. I studied biotechnology.
But like many of you reading this, I had an idea that just wouldn’t let go. I wanted to build a digital platform to close the industrial gap I observed in my field—something that could help biotech students like me navigate careers, understand the industry better, and explore AI integration in life sciences. But to do that, I needed a website. I didn’t have funds to hire a developer or the technical chops to build one from scratch.
What I had, though, was curiosity—and a growing belief in what everyone online was calling “vibe coding.”
When Vibe Coding Felt Like Magic (At First)
In the beginning, everything felt surprisingly smooth—almost magical. I discovered Firebase Studio while looking for a simple way to build my website without diving deep into coding. All I had to do was give instructions in plain language.
I typed out a prompt like, “Create a homepage with a hero section, a short About Us block, and a contact form.” And just like that, the interface responded with code and layout that worked. It was thrilling. It felt like I had found the holy grail for non-tech creators.
Firebase Studio interpreted my ideas, translated them into working code, and instantly generated pages. The AI-powered builder seemed to understand what I needed. At that point, my website was basic—it had a homepage, some descriptive text, and links. But I was proud. The bot handled these tasks beautifully because the amount of information and complexity was still low.
During the first few days, I kept adding simple components like cards, buttons, and text sections, and Firebase Studio kept delivering. It gave me confidence. I wasn’t worried about debugging or syntax errors because everything just worked. I believed I could go far with this tool. But of course, things weren’t going to stay that simple for long.
From Flow to Frustration: When Scaling Broke the Magic
As I began scaling my project, things took a different turn. My initial goal was to create a job-oriented platform for biotech students, listing internships, job openings, and exam notifications. But as I interacted more with students and peers, I realized that jobs were only one part of the bigger puzzle.
Most students didn’t even know what paths were available to them, how to prepare, or where to apply. So, I pivoted. Instead of just listing job posts, I wanted BTGenZ to become an educational and motivational resource for biotech learners.
This meant adding multiple content layers: blog sections, career advice modules, a newsletter setup, and dynamic recommendation blocks. The vision had grown—but unfortunately, Firebase Studio didn’t grow with it.
At this stage, editing became painful. Every time I asked the bot to update a page or edit a layout, it started misinterpreting my prompts. Even when I clearly specified which file or which component to modify, it either altered the wrong section or overwrote existing structures.
Pages started breaking. Components disappeared. My carefully built structure collapsed multiple times. There were moments when my entire website showed a server error, and I had no idea how to fix it. Debugging felt impossible because I didn’t know where to look.
And even though I followed the AI’s suggestions, the fixes often made things worse. Firebase Studio’s AI seemed to lose track of earlier instructions and context. The more complex the website became, the less the AI understood what I wanted.
Firebase Studio’s Limitations for Non-Tech Creators
Firebase Studio is great for quick launches, but it’s not reliable when your website becomes multifaceted. While it helped me create my initial framework, it couldn’t support continuous development. It lacked continuity. For example, if I added a job alert card on one page and asked to replicate a similar card on another page, the formatting, responsiveness, or even basic structure often broke.
One big issue was the lack of contextual memory. The AI didn’t retain an understanding of what was previously built. It treated every instruction as new, often ignoring existing design logic. This made it incredibly difficult to maintain consistency across pages. Even though I kept referring to specific files and wrote detailed prompts, the outcomes were unpredictable.
Also, Firebase Studio didn’t give any visual feedback. If something went wrong, it wouldn’t tell me why. There was no error trace or visual debugging help—just an output or a broken interface. For a non-tech creator, that’s a nightmare scenario.
Why Vibe Coding Isn’t as Easy as It Sounds
There’s a popular belief online that anyone can now build a website using AI. And while that’s partly true, it doesn’t account for the middle phase—the one between creating your homepage and having a polished product.
Vibe coding makes you feel empowered when you’re working with simple data. But once your website starts storing multiple pieces of content, and each part of the structure starts relying on other parts—like dynamic links, auto-generated cards, search filters—everything becomes fragile. One wrong instruction and your whole layout collapses. And unless you understand how components and files work in Next.js or React environments, fixing it can take hours.
You don’t need to know how to code to start. But if you want to sustain and grow your project, you’ll eventually need to learn how to debug, read through errors, or at least understand file relationships. Otherwise, you’re left guessing.
Exploring Other Tools: Replit, TempoLabs, and the Role of AI Agents
As my frustration grew, I started exploring other AI-first development tools. I came across Replit and TempoLabs—both of which offered a much smoother experience in specific areas.
- For example, Replit’s Ghostwriter feature can take screenshots of the code changes it makes. That’s powerful for someone like me because I could see what was changed and where. I wasn’t blindly trusting the AI—I had visual confirmation.
- TempoLabs, while limited to 5 free credits per day, was another game-changer. It verified each output it generated. If there was an error, it wouldn’t just throw it at me—it would pause, reassess, and try a corrected version. That kind of self-checking loop is what Firebase Studio lacked.
This experience made me realize that not all AI agents are created equal. Some understand visual logic and workflow better. Others just follow instructions like a basic script. For real-world, evolving projects, we need tools that think with us—not just for us.
The Bigger Problem: A Lack of Education and Guidance
The biggest lesson from my 60-day journey wasn’t about tools. It was about education. Nobody tells you what happens after the fun of launching your homepage. Most YouTube tutorials or AI blogs stop at “click this, build that.” But what about maintaining the site? What about SEO optimization? What about structure planning or code versioning?
We don’t talk enough about the learning curve after launching. And that’s where most non-tech creators give up. Not because they aren’t capable—but because there’s no structured guidance for the next steps.
AI is helpful, yes. But it still needs you to think structurally and logically. You can vibe-code a site, but you can’t vibe-debug one unless you know what went wrong. Without this foundational education, tools like Firebase Studio can quickly become barriers instead of enablers.
So, Would I Recommend Vibe Coding to Non-Tech Creators?
Yes—but with serious caution.
Start small. Set clear goals. Stick to basic structures. As soon as you see the need for layered logic or advanced styling, prepare to either learn or switch tools. AI is an incredible support system, but it isn’t a full replacement for development experience—not yet. And vibe coding, while exciting, can mislead non-tech creators into thinking they’re safe from bugs or errors. You’re not. And unless you back up your progress regularly and document your steps, one wrong update can wipe out hours of work.
Final Thoughts: You Can Build, But You Have to Learn
In the end, I survived those 60 days of Firebase Studio—not because it was perfect, but because I learned to adapt, to explore better tools, and to become more intentional with my decisions. BTGenZ.in is live now. It’s not a perfect website, but it’s built from real struggle. And every glitch, every crash, every broken layout taught me how to be a more resilient creator.
To any non-tech founder, dreamer, or creative thinker reading this: you can build your own thing. But understand that vibe coding isn’t a shortcut—it’s a starting point. The real work begins when your ideas grow bigger than the tools meant to build them. And trust me, they will.
Frequently Asked Questions
“Vibe coding” refers to using AI tools to build websites or applications by providing instructions in natural, plain language rather than writing traditional code. It gives the feeling of intuitively creating with the AI.
According to the author’s experience, Firebase Studio is great for quick launches and basic websites, but it struggles when a website becomes multifaceted and complex, lacking continuity and contextual memory.
Contextual memory refers to the AI’s ability to remember and understand previous instructions, design logic, and the overall structure of the website it’s building. Without it, the AI treats every new instruction in isolation, leading to inconsistencies and errors.
Key takeaways include starting small, setting clear goals, understanding that AI is a support system but not a full replacement for development experience, and recognizing that foundational education in debugging and understanding file relationships becomes crucial for project sustainability.
BTGenZ.in is a digital platform created by the author to help biotech students navigate careers, understand the industry better, and explore AI integration in life sciences. It started as a job-oriented platform but pivoted to become an educational and motivational resource.