Every year somebody declares touch typing obsolete. Smartphones were supposed to finish it off, then dictation, and now language models that write on our behalf. The argument never changes: why learn to type fast when so little text gets typed by hand anymore?

The argument is sound. It just isn’t arguing with what touch typing actually gives you.

Speed has nothing to do with it

Start with a fact that every “ten-finger method” course would rather you didn’t know. In 2016, a team at Aalto University in Finland (Anna Feit, Daryl Weir, and Antti Oulasvirta) sat 30 ordinary people down at a keyboard, pointed motion-capture cameras at their hands, and recorded how they typed. Some had learned the classic method at some point in their lives. Others were self-taught and hunted around with five or six fingers, whatever worked.

There was no difference in speed. The self-taught typists were just as fast as the ones who placed their hands by the book. The number of fingers involved had no bearing on the result at all.

But the same study turned up a second finding, and it’s the one people tend to forget.

The difference is where your eyes are

The fastest typists in the group had two things in common: they kept their hands still, and they barely looked at the keyboard. Not the fingers, not the layout, not the technique. The gaze.

Think about what actually happens when someone types looking down. They see the screen in snatches, lifting their head every few words to check what came out. A typo registers not at the keystroke but half a line later. A sentence gets judged after it’s complete, not while it’s forming.

Put another way, the screen and the thought fall out of sync. You think one thing, type a second, check a third, and everything runs late. In that sense touch typing isn’t about the hands at all. It’s about keeping your eyes where the text is being made.

Why the cost lands on thinking, not on the clock

It looks like all of this costs seconds. It doesn’t. It costs attention.

Working memory is small. By Nelson Cowan’s estimate it holds about four items at once, and that is the entire budget you have while carrying a thought you’re about to write down. Every switch from keyboard to screen and back to the thought spends some of that budget on finding your place rather than on the content. Part of the sentence simply goes missing on the way.

Cognitive psychology has long distinguished two modes of doing a task: controlled, which demands attention, and automatic, which demands none (Schneider and Shiffrin, 1977; and the three-stage model of skill acquisition from Fitts and Posner). Adult speech is automatic. You don’t think about how to move your tongue; you think about what to say. That is the only reason conversation is possible at all.

Touch typing moves typing from the first mode to the second. Not “type faster” but “type without giving it any thought.” The working-memory slots that frees up go to the thought itself.

What AI changed

The main way people work with language models today is a text conversation. Not one document polished over an hour but dozens of short messages: clarify, ask again differently, push back. Text has stopped being the product of the work and become the medium it happens in. Like talking.

In that medium, typing turns into the bottleneck between the thought and the result. The closer it gets to the speed of thinking, the less gets lost on the way to the screen, and the sharper the prompt. And the quality of the model’s answer depends almost entirely on the quality of the prompt.

There’s a second, less obvious effect. AI is taking over the production of text, which is the very thing people once learned to type for. What’s left for the human? Forming the intent: working out exactly what you want and saying it precisely. That is the most attention-hungry job there is, and it’s still done by hand, on a keyboard.

The skill, then, hasn’t lost its value. It has changed jobs. It used to save a typist’s time. Now it protects the attention of whoever is doing the thinking.

What about dictation?

It would be dishonest to wrap up without a word on it. Voice input exists, it works, and in some situations it wins outright: dictating a rough draft when there’s no one around is faster than typing one.

But editing, precise terminology, code, variable names, plus anything done in company or in a room that needs to stay quiet, all of that stays on the keyboard. And those are exactly the tasks where losing your train of thought costs the most.

So what’s worth learning isn’t the “ten-finger method” with its correct hand placement. It’s one specific ability: typing without your eyes taking part. Not for the words per minute. So that nothing stands between the thought and the screen anymore.