This idea is marked as "live" but I'm not seeing which of the suggestions below were implemented into the editing experience. Are there any release notes explaining this update?

Make transcripts or any documents a single source of truth. Highlighted text generates notes that can be out of sync when the original document is updated. For example, a transcript may have some spelling errors that, when fixed, will not reflect in previously generated notes.

Mass fix for misspelings, as mentioned by Floyd and Christensen, is an alternative approach for this issue.

There only seems to be a way to designate a speaker/change the speaker's name one by one by each line on transcripts; there should be a way to bulk update the speakers' names( e.g., spk_0 should all be chaged to 'john'), while keeping the option to update individually if there's a mistake etc.