aiopinionpersonal experience
AI can generate code and designs for anyone. But without deep expertise, you can't judge if what it produces is actually good.
May 15, 2026
Without a doubt, AI is excellent at generating code and designs. We have officially entered an era where almost anyone can produce software. Since last year, several landing pages in my organization have been successfully created by non-technical people with the help of AI.
If you spend any time on X or Threads, you’ve likely seen the loud, sweeping claims: Programmers are dead. Designers are no longer needed. But there is one interesting thing these posts have in common: almost all of them come from non-tech people.
They are swept up in the era of “vibe coding”, the trend of prompting an AI to piece together an app based on a feeling or general idea rather than actually engineering it. While a fully functioning, vibe-coded app is a neat starting point, seasoned programmers and designers know the truth: building a good, scalable digital product isn’t that simple. Getting an app to run is just step one. Maintaining it is an entirely different thing.
I am a believer that someone with deep expertise in a specific field will always produce superior results when using AI compared to a novice.
Take my own experience as an example. No matter how hard I try, and no matter what advanced tools I use when generating UI with AI, my designs cannot beat those created by a seasoned designer. A professional’s design just feels right, whereas mine usually ends up looking like a generic AI slop product.
On the other hand, no matter how hard a designer tries to vibe code an application, their AI-generated code quality will not beat the architecture designed by a seasoned programmer.
So, what exactly differentiates a seasoned professional from a non-tech person who is vibe coding?
It is the ability to judge the quality of the output.
When a seasoned programmer looks at AI-generated code, they have the cultivated sense to evaluate it. We don’t just see lines of text. We can instantly assess whether it is secure, scalable, and maintainable. AI can do the typing, but it cannot provide the judgment.
This brings up a crucial question: how do you develop that sense of judgment?
It all comes down to developing your taste. When you are starting out, your taste is great, but your abilities haven’t caught up yet. You know what you’re making isn’t quite good enough. That frustration is actually a good sign. It means your taste is functioning properly.
To close that gap, you have to expose yourself to great work. This is how you learn what greatness looks and feels like:
Find people who are respected in their field and surround yourself with their work. Use their apps, study their repositories, and read their insights. Learning from the best is the only way to internalize what “good” actually means.
Most importantly, you have to practice your craft. A designer must design. A writer must write. Creating things manually makes you not just a good judge of taste, but eventually, a taste maker.
While practicing, seek feedback. A good critique from the right person can accelerate your learning process far more than trial and error alone. Your early creations probably won’t be great, and that phase is entirely normal. Don’t quit. It gets better over time.
Despite the incredible advancements in AI, I find myself still actively learning new things the old-fashioned way. For example, I have been learning UI design by replicating layouts in Figma, entirely for the sake of building my own design sense. It’s slow and sometimes humbling, but that’s the point.
AI hasn’t made expertise obsolete. If anything, the fundamental concepts of whatever craft we’re working on have become more important than ever. AI can execute your vision instantly, but it is your taste that dictates whether that vision is actually worth building.
The people who will get the most out of AI aren’t those who use it the most. They’re the ones who can tell when it gets things wrong.
Thank you for reading 👋
Comments