Working in tech (primarily in IT & InfoSec), I’ve found myself wearing many hats that involve coding. While I’m not a software engineer by trade, I’ve built everything from WordPress sites from scratch to microSaaS applications, automation, and creative coding projects. Lately, something’s been eating at me: this nagging guilt every time I use AI code generation to generate more than half of my code.
🤔 A Year of Wrestling with AI and Code
I’ve felt this weird moral obligation to write code by hand, and I’m starting to recognize how much this limiting belief has held me back. Take my NUXT project, for example. It’s taken me almost a year of on-and-off work because I kept going back and forth about whether I should code my own boilerplate by hand (well also because life events got in the way 😅).
After exploring various frameworks like Laravel, Ruby, and React, I found my sweet spot with the VUE ecosystem. The syntax just clicked – it’s easier to read and write, more so it works beautifully with AI coding tools like Copilot. (Don’t get me wrong, I’m still a Laravel Lady at heart, but AI tools haven’t quite mastered PHP frameworks yet). Despite finding this perfect match, I’ve been holding myself back with a self-imposed rule about AI usage that’s seriously slowing my progress.
💡 The Reality Check I Needed
Here’s what hit me this week while working on my hackathon projects: if my focus is getting out of homelessness and generating income, why am I obsessing over the percentage of AI-generated code? The truth is, you still need a solid understanding of technology infrastructure and code to get a working application up and running, deployed, and secure.
I know how to set up a VPS, handle hosting, and manage the overall technical infrastructure. Using an AI assistant doesn’t erase that knowledge or make me less competent. In fact, this is what separates traditional developers from founders – entrepreneurs will do whatever it takes to succeed, even if it means getting help (aka whether from humans, bots, or AI).
🚀 Shifting from Developer to Founder Mindset
Developers trying to generate income could benefit from this mindset shift: it doesn’t matter how many GitHub commits you have or if half your codebase was AI-generated. At the end of the day, customers don’t care how the application was built. They care about one thing: does this bring value and is it worth paying for?
🛠️ The Reality of Working with AI Code Generation
I’m on day four of my hackathon, aiming to launch several products (or at least get to MVP stage), and I’m finally shifting my mindset. I’ve noticed something about my relationship with tools like Cursor, bolt.new, and other AI code generation tools – I often get frustrated because I feel like I could work faster doing it myself.
And honestly? Sometimes that’s true. When you code by hand, you have more autonomy and things tend to go your way (aside from major bugs). But using AI code generation requires a completely different approach. From my point of view thus far, you have to think more like a product developer or project manager.
💡 A New Way of Building
When approaching a product build with AI code generation, you need to think like someone who’s technically aware but not necessarily a full-time developer. You become more of a conductor – you need excellent communication skills and a solid understanding of the infrastructure backbone.
In my opinion, what really matters is:
- Having a technical understanding of what you’re building
- Creating genuine value
- Protecting user data
- Building secure applications that serve a real purpose
🌟 Finding Permission to Embrace AI Code Generation
What really helped shift my perspective was hearing that successful developers like Marc Lou use AI code generation. Here’s someone who’s made $1 million from building applications, openly uses GitHub Copilot and AI tools – and people don’t care! They care that his products help them save time and money.
Take Dan Shipper, for example. Watching his podcast over the months, and more so recently opened my eyes to something interesting. He has many entrepreneurs-in-residence or I guess developers in prominent positions working at his company (I’m not quite sure how this works 😅), but what struck me was how he actively encourages them to use AI code generation. It’s all about, let’s just build and ship! In a recent episode “He Built an AI App With 10,000 Signups in 1 Month” with Kieran Klaassen, the GM of Cora.Computer and Brandon Gell, Dan was genuinely excited about his team using tools like Cursor to build faster and more efficiently, and even to improve writing! This really challenged my perspective – here’s a successful founder not just accepting but enthusiastically promoting AI tools to accelerate development. It reinforces what matters most: creating valuable products that help others save time and money while building their own success stories.
🎯 Moving Forward
So today, I’m letting go of the limiting belief that using AI makes me less capable. If AI can help me build faster while still maintaining security and quality, why not embrace it? The goal isn’t to prove I can write every line of code – it’s to create valuable solutions that help people while building a sustainable income for myself.
Have you struggled with similar thoughts about AI code generation? I’d love to hear your perspective on balancing traditional development with AI tools.