For programming, yes, because the text is formatted, but not for plain English. The average person can easily speak 130 to 150 words per minute, but they certainly can't type that. If you personally can, then alright, that's cool. You're maybe more efficient with a keyboard.
And regardless of your opinion on it (or mine) vibe-coding is plain English, so what they're saying is true for that one use case. Besides, they can use both. The keyboard is still there.
This comment was written using speech-to-text.
When you're not thinking, and you know what you're going to say, you can talk at 130 words per minute very easily.
When using a keyboard, you're going to pause just as much to think. The comparison is this: when you are in text entering mode (not thinking mode), are you faster talking on the microphone or the keyboard?