Can I output two streams of text?

While my post on dictation in Emacs was focused on real-time writing with voice, my main motivation with the tool is to blend text and voice inputs in ways that can give me higher Words Per Minute (WPM) speed than any of them individually¹.

Speech is around 1.5-2 times faster than keyboard². But then you might need to do post-processing for mistakes which could increase the total input time drastically (probably lower if you put LLM post-processor for general post-processing). With real-time edits, this post-processing time goes to zero so I could argue that my current version already achieves the set goal.

But I think we should also explore the independence of voice and text cursors. This is what led me to find and try use cases where I could meaningfully write and voice two things at the same time. Here is how this looks like while speaking out docstring for a class in Python:

Figure 1: Trying to write docstring using voice while I write code using keyboard. I don't feel I was able to multitask well here as both inputs were cognitively demanding. Apologies for the poor audio quality.

Since my last post, I added a few fixes for conflicts between voice and text cursors. Even then, you could see me stumbling since my mind couldn't put attention in two places both of which needed healthy amount of effort. People do talk while typing, given one of the two is less of an attention sink. This means pairing something like writing emails while programming might work out better. Or maybe there is no point in parallelizing attention.

Footnotes:

Although convenience is a good enough reason to use this, I wanted to play around with pure speed gains.

~150-170 WPM for speech and 70-90 for keyboards. Without factoring in pondering time.