The Keyboard Tax: What Typing a Thought Costs That Saying It Does Not

A thought arrives while you are carrying groceries up the stairs: call the dentist, or the missing line in tomorrow’s presentation. You will put it in your phone. By the time your hands are free, you have a different thought, and the first one has gone where thoughts go.

Every thought that passes through a keyboard pays something on the way: a few seconds, a few looks at the screen, a little attention that was on the stairs a moment ago. Call it the keyboard tax. This post looks at what it is made of, what two studies report about typing and speaking on a phone, and what a recording can and cannot change.

The steps between a thought and a note

You take out the phone and unlock it. You find the app, or the one you had open. You find the field and bring up the keyboard. You type, correct what the keyboard changed, and save. Then you look up and find your place again.

None of these steps is hard. The problem is that the thought has to survive every one of them, and each step gives it a chance to go. Our post on why your best thoughts disappear makes the same point about ideas in the shower and on a walk: the system has to be ready before the thought arrives. We could not find a study that times this chain in daily life, so there is no number for it here.

What typing on a phone costs when the words are already given

The best data we found on phone typing come from a 2019 study in which 37,370 volunteers typed given sentences into a web test on their own phones. The average speed was 36.2 words per minute, with 2.3 percent of the final text left uncorrected (Palin et al., 2019). The authors note that the volunteers chose to take part, which limits how far the results carry. Two details matter here. The task was transcription: the sentence was on the screen, and the typist only had to copy it. And even with the words given, people still spent part of their time correcting them.

Your thought is not on the screen: you have to hold it, phrase it and type it at once. The study does not measure that, and we will not guess.

Speaking is faster, in the lab

The comparison people usually want is speech against the keyboard, and one lab study measured it. In it, speech input was about three times faster than typing on a phone keyboard for short messages, in English and in Mandarin (Ruan et al., 2016). The same study is fair about the other side: after corrections, the speech recognizer left slightly more errors in the final text than the keyboard did.

Three cautions. The participants entered short messages in a lab, not their own thoughts on a staircase. The recognizer was a research system, not the one in your phone. And “faster” says nothing about whether what you said was worth saving. Our post on Kidlin’s law uses the same study for a related point: the first, rough version of a thought is often the useful one.

Where typing still wins

Saying a thought is not better in general. Typing wins when the wording matters: an address, a name you will look for later, a line you are polishing. It wins in a meeting, a library, a train at night, and anywhere you would rather not be heard. It wins for editing, since a transcript still needs a keyboard to fix. And a spoken note is only as good as its transcription; a quiet mumble on the stairs gives a recognizer little to work with.

So the honest claim is narrow. The keyboard tax falls hardest on short, unplanned thoughts that arrive while your hands and eyes are busy. That is the thought a recording is for.

What a one-touch recording changes

In Vozly, capture is one gesture: hold the microphone, speak, release. What you said is transcribed with Apple’s speech recognition and saved as a to-do, note, diary entry or dream, and the original audio is kept, so you can replay what you actually said when the transcript is off. On iOS 18 and later, the “Add task” control in Control Center opens the app and starts recording at once, which takes the “find the app, find the field” steps out of the chain.

It is one touch, not zero: you hold the button while you talk, or slide it to the right to lock the recording and tap stop when you are done, and you still take the phone out and unlock it. Nothing starts without that touch. Vozly does not promise accuracy: the transcription is as good as your recording. It is not an assistant. It does not understand a due date you spoke, set a priority, or sort the item; you do those by hand afterwards. There is no search, so items are browsed by date and project, and your data lives on the device: there is no cloud sync yet.

What it changes is the length of the chain. The thought has to survive one gesture instead of a row of steps, and when the transcript is rough, the recording is still there. A pile of captures still needs a moment to sort; our second brain post covers that side.

Sources

• Palin, K., Feit, A. M., Kim, S., Kristensson, P. O., & Oulasvirta, A. (2019). How do people type on mobile devices? Observations from a study with 37,000 volunteers. Proceedings of the 21st International Conference on Human-Computer Interaction with Mobile Devices and Services (MobileHCI ’19). ACM. doi:10.1145/3338286.3340120
• Ruan, S., Wobbrock, J. O., Liou, K., Ng, A., & Landay, J. (2016). Comparing speech and keyboard text entry for short messages in two languages on touchscreen phones. arXiv:1608.07323. arxiv.org/abs/1608.07323

More on capturing to-dos by voice: Best Voice To-Do List Apps for iPhone (2026) on voicetodolist.com, a resource site published by Proseca, the maker of Vozly. Vozly itself is at vozly.app.