WolfI have found voice input to be super helpful. I use it entirely for prose: notes, tasks, and reminders. I **don't** use it to make the machine "do things" or for coding (but that's just me ... you do you; and also, I don't use it for coding **yet**).
I currently use two tools:
* Spokenly (phone + laptop)
* Just Press Record (watch + phone)
Each has advantages. Both produce transcriptions which I then "route" to the appropriate place (Things, Obsidian, Fantastical) with Sharing.
Just Press Record also runs on my watch, so I can use it while driving or other situations where the phone is inconvenient, illegal, or unavailable. I can later look at a list and handle them one at a time or in bulk.
Spokenly is significantly more accurate. You pick the underlying model, so you can decide if you want local-only (as I do, so no fees of any kind), size (controls memory used and speed of translation, at the cost of accuracy), and what spoken languages it knows. You can switch at will; and you don't have to have the same model on your phone as on your laptop. You typically handle results in Spokenly immediately (for a while I thought this was the only choice), but they are saved and you can look at them all in History and deal with them from there.
I'm still playing with which models are the best balance of speed and accuracy for my use case. On my phone I'm using "Distil-Whisper Small (English Only)". On my Mac, the same but "Medium".
This doesn't **sound** like a huge win. I certainly didn't expect much when I decided to try it. But it turns out to punch far above its weight.
#VoiceInput #Spokenly #JustPressRecord #Things #Obsidian #Fantastical #Productivity
Joseph King
Kyle Smith
Diva