One of the most important things to know about Descript is that there are two main typing modes: Editing Media (default) and Correcting Text.
In Editing Media mode, cut/copy/paste/delete affects the underlying audio/video in addition to the text - and you can only type punctuation.
If you want to make corrections to typos made by the automatic transcription, enter Correcting Text mode by pressing
Command + E on Mac, or
Ctrl + E on Windows, or clicking the dropdown in the upper left of your script: