Can and cannot

AI's strengths and weaknesses are almost the mirror image of your Python programs.

AI is strong at things with no exact answer — recognising a face, suggesting a video, summarising an essay, translating a sentence, spotting a pattern in huge amounts of data.

AI is weak at things needing exactness — precise arithmetic, guaranteeing a fact is true, explaining why it gave an answer, staying identical between runs.

Your calculator is better at arithmetic than the most advanced AI ever built. That is not a joke.

Here's the part that matters most for you as a user.

When your Python program fails, it crashes — a red error, a line number, an obvious problem. When AI fails, it produces a fluent, confident, completely wrong answer, in exactly the same tone as a right one.

There is no error message. There is no line number. Nothing looks wrong.

Made-up AI output has a name — a hallucination — and it's not rare. It invents book titles, historical dates, and functions that don't exist, all delivered with total confidence.