I'm not a professional coder, just a computer enthusiast who tries to learn and experiment with general C and embedded-specific coding. You can check
my repo and see for yourself what I've done with AI help in one year as a side project in my own time. What do you think about my code?
My point of view after one year of using a general-purpose AI, it's accelerates learning, but it does not replace understanding. It's not a magic tool. For example, hard-to-find bugs are out of its scope. For the simple fact that AI never has the whole context and all the data.
Another example is that AI can't make architecture choices, because it'll never assume long-term cost and management. But you can get guidance with questions like: in the scope of x, how can I implement y? Why is z good for general-purpose systems and bad for an embedded one?
Maybe professionals and/or seasoned developers don't need such guidance because they have already learned the required knowledge. If you work in a team, you probably don't need AI. I work alone, and I find in AI a virtual companion able to check spelling mistakes, review code, write documentation, and talk about architecture choices.
Human-written text, AI grammar/spell checked.
