With AI writing so much code now, I’ve been wondering: is learning to code from scratch still worth it for someone starting out?
On one hand, you can get a working app without writing much code yourself. On the other, it feels like you still need to understand what’s going on to debug things, judge whether the AI’s output is actually good, and build something that’s really yours.
What do you all think? If you were starting today, would you still learn to code properly, or would you focus more on learning how to direct AI tools? Curious what changed your mind either way.
I think learning to code remains essential. While AI can generate code, the discernment required to evaluate its quality, debug effectively, and truly innovate still stems from a foundational understanding of programming principles. Directing AI tools effectively also requires a grasp of what is computationally feasible and how to structure a problem for an optimal solution.
Yeah, that’s the part I keep coming back to too. It’s not really about writing every line yourself anymore, it’s about knowing enough to spot when the output is wrong, or just subtly off.
I think the shift is less “learn to code or don’t” and more “learn to read code, then learn to direct it.” The people who struggle most with AI tools, in my experience, aren’t the ones who type slowly — they’re the ones who can’t tell a confident-sounding bad answer from a good one.
So yeah, still worth learning properly. Just maybe with less time spent memorizing syntax and more time spent on debugging and judgment.
