AI can now generate whole pages of code from a single sentence. But as Luna just learned, 'one button and done' only holds until the screen turns red.

Coding in the AI era: why you still need to understand code

AI tools (code assistants, 'vibe coding') make writing and shipping much faster — and make it feel like coding is just pushing a button. But when an app runs for real, errors like a 500 Internal Server Error or a lost database connection still happen — and AI doesn't automatically know your system's context to fix them correctly. The person who understands what they just deployed is the one who can debug it.

So what should you learn in the AI era?

  • Read and understand the code AI generates — don't just copy it.
  • Know what happens when you hit 'Deploy': build, environment, config variables, database connections.
  • Read error logs (500, connection failed…) and trace back the cause.
  • Use version control (Git) to roll back to a working version when things break.

Frequently asked questions

  • Will AI replace developers? — AI is a huge accelerator, but when systems fail, the person who understands context and can read logs is still the one who fixes it. The skill shifts; it doesn't vanish.
  • What is 'vibe coding'? — Writing software mostly by describing what you want to an AI. Fast for prototypes, but you still need to understand the code to ship safely to production.
  • You hit Deploy and got a 500 — now what? — Read the logs to find the error (e.g., a lost database connection), check environment/config variables, and use Git to roll back to a working version.
  • What's the 'Undo' button of programming? — There's no magic Undo; the closest thing is Git — going back to a previous commit.

About this piece: this is a DETOON entertainment comic (characters Luna and Tau) with a few practical notes on coding in the AI era — not a detailed technical guide.