Let me start by saying that I am probably not your stereotypical vibe coder that is totally reliant on AI for programming, etc. I recently graduated with a bachelors in cybersecurity, and currently work in IT as my 9-5. So while I wouldn't consider myself a great programmer or software engineer, I'm decent at using computers and know the ins and outs of them somewhat well.
After I came up with an MVP, I got to work on the actual programming. While I mentioned I am decent with computers, my background in actual coding doesn't go further than an intro to python class I took my sophomore year of college - enough to get a grasp on the basics of coding, but not much that would help me with this project I was about to embark on.
I started off by using ChatGPT to write my initial code. It was helpful, but it was hard going back and forth from my codebase to ChatGPT, but nevertheless, I made it work and I was able to design my MVP in about a week or so.
Then I found Cursor and I never looked back. When I first started using Cursor, I thought that I would be finished with my product within a couple of days. But as my codebase grew bigger, I realized that it was becoming more of a challenge. In particular, explaining what I wanted and how I wanted different features to react with the rest of the code.
I would often just revert changes and have it try again until I finally got things to work the way that I wanted. But I soon realized that I would have to learn the actual functionality of the code instead of mindlessly demanding the AI to make changes for me. So I learned the ins and outs of code, helping me to understand why things weren't working the way that I wanted and forcing me to design solutions myself.
So here's my big takeaway; vibe coding can be useful and the ability to create things with just a few sentences is quite fascinating and incredible. But to build useful, complex products, a solid understanding of programming and coding is absolutely necessary.
If you guys are interested in checking out the product I've attached a link to the site where you can install it. It's a free extension designed to detect and redact sensitive information from your AI prompts before sending them. The extension has a built in NER model and comprehensive pattern recognition, so all processing happens locally through the extension itself - meaning no sketchy data transfer across servers.
Thanks!
[link] [comments]