The editor that writes tests for your code
Here's a fun one. Autocomplete used to guess your next line. The latest update goes further: point it at a function and it proposes a whole set of tests — including the edge cases you'd probably have skipped.
It reads what the code is supposed to do, then asks the awkward questions: what happens with an empty input, a negative number, a value that's exactly on the boundary?
Why it matters
Tests are the part everyone agrees is important and nobody enjoys writing. Moving that work to the editor doesn't just save time — it changes the habit. When a test is one keystroke away, you actually write it.
- Edge cases first. It surfaces the inputs you forget under deadline pressure.
- Readable by design. Generated tests double as documentation of intent.
- A safety net for agents. When an AI agent later refactors the code, the tests catch regressions instantly.
The catch is the same as ever: a generated test is a suggestion, not a guarantee. Read it, make sure it asserts what you actually mean — and then let it guard your code while you move on.