Ty Daniels
Superstar
as a non-programmer who has been been running various code created by LLM's since chatgpt debuted I strongly disagree. If you're inquisitive and have the mindset to troubleshoot then you can start off small with small projects to get productive results.
I understand your position, but....
There are certain things about knowing how to code and App architecture in general that is needed to be really effective.
Even when prompting, there is a "Not Knowing" what "You Don't Know" aspect of it, where someone "Who Knows" can "Prompt" for more specific results.
Sure you can have the AI build "app code", how is your app structured? are you "separating concerns"?
Is the UI/UX correct? is a function "over-engineered"?, is the LLM referring/using imaginary "ghost functions" or props?
What about "Re-Rolls", where you are then tasked to add additions to your existing code, how do you, not really understanding app architecture/coding going to "Safe Guard" from the LLMs inevitable "Drift" and lack of coherence?
To me is sounds like you are learning as you go along, which isn't an issue.
But for the average person who knows nothing about coding, or app architecture in general, will not find consistent results on a project "Vibe Coding".
To get the most out of it, you really have to understand coding, and application architecture.
Last edited:


"
