18 min · up to 80 XP

When AI is confidently wrong

The classic failure patterns of generated code.

Sam asked her AI for a small feature: remove the newest book from the list. What came back looks immaculate, tidy name, helpful comment, five confident lines.

She's about to run it. But this is Module 4, so you go first: one of these lines cannot work. Not "works but gives the wrong answer", it will stop the program dead. Find it before the red text does.

Spot the bug

Sam's list is a plain JavaScript array of titles. One of these lines calls a method JavaScript does not have, the AI made it up. Click the impostor.

Click the line you think is wrong.

1function dropNewest(titles) {
2  // The newest title sits at the end
3  const removed = titles.removeLast();
4  console.log("Dropped: " + removed);
5  return titles;
6}

Failure pattern one: invented names. It's the code version of a chat AI inventing a citation, the same machinery, predicting what plausibly comes next, with no lookup against a list of real methods. And it never, ever writes "I'm not sure this method exists."

The good news: invented names fail loudly. One run, one clear error message, fixed in a minute. The patterns that follow are not so polite. Sam's next feature did run.

This lesson is part of the full course

The free lessons are just the opening chapter. Unlock the rest of Code with AI to keep going.

Every Code with AI lesson, as it ships
New interactive lessons added monthly
Founding-member pricing, locked in
Learn at your own pace, forever

or keep browsing the free lessons