While building my own voice capture workflows, I noticed something: speaking to the machine works much better when you don’t see your words being transcribed in real time.
The moment text appears on screen, you shift from thinking to editing. You notice a weird phrasing. You pause. You self-correct. The focus moves from what you’re trying to say to how it looks written down.
Turning off live transcription was a small change that made a surprising difference. Without the visual feedback, you stay in your head. The thought completes itself before your inner editor gets a chance to interfere.
Let the transcription happen later. That’s the machine’s job.
