I was in a strategy meeting in Shibuya last month, watching a senior product manager take notes on his laptop. Brilliant guy—ten years of experience, deep instinct for user psychology, the kind of person who can look at a feature mockup and tell you exactly where the friction is.
But I couldn't stop watching his hands. Two-finger typing. Hunt and peck. Eyes flicking between the screen and the keyboard every three seconds. Every time someone said something important, you could see him tense up: do I keep listening, or do I finish this sentence?
He was losing the thread of the conversation because his brain was stuck in traffic.
I've managed hundreds of people over two decades. I've seen every work style imaginable. The slow walkers who produce masterpieces. The fast talkers who ship nothing. I would never claim that typing speed equals productivity, or that slow fingers mean a slow mind.
But I will say this: typing too slowly is a fatal structural flaw for a knowledge worker. And we pretend it isn't because it feels too basic to be important.
The Attention Tax
We think of typing as simultaneous—thinking and doing at once. But it's not. For a slow typist, the brain isn't processing strategy while the fingers move. It's managing a bottleneck.
You have a thought. A good one. A connection between two ideas that might change the project. But your fingers are still hunting for the "k" key. By the time you finish the sentence, the thought has evaporated. Working memory is fragile. It doesn't wait for your pinky to find its home row.
I feel this myself, and I type fast—deliberately practiced since I was a teenager. Even now, my mind is often three sentences ahead of my hands. I can feel the gap. The frustration of having to slow my thinking to the speed of my fingers. If I feel it at 80 words per minute, I can't imagine what it feels like at 25.
Flow state requires a seamless API between brain and machine. When that API is clogged by mechanical friction, you never get there. You're constantly context-switching between what you want to say and how to make the letters appear. It's like trying to compose music while learning the piano.
The Professional Development Blind Spot
Here's what baffles me. The same PM who hunts and pecks will spend $3,000 on a Python course. He'll get certified in Agile, learn advanced Excel, attend storytelling workshops. He optimizes everything except the baseline mechanism he uses for eight hours a day.
We've accepted a bizarre cultural norm: If you type slowly, you just type slowly. Like it's a personality trait. Like being left-handed. We don't treat it as a skill gap because it feels too elementary—something you should have fixed in high school, so admitting you need to fix it now feels embarrassing.
But the math is unforgiving. If you spend four hours a day writing emails, specs, and Slack messages, and you type at half the speed of a fluent typist, you're losing two hours a day to mechanical friction. That's ten hours a week. Five hundred hours a year. An entire quarter of your working life, spent waiting for your fingers to catch up.
The Voice Workaround
For most of my career, if you were a slow typist, you were trapped. That was your ceiling. You could take typing classes (boring, slow, embarrassing) or you could accept the tax.
The AI era just blew the ceiling off.
At Mercury, we've been experimenting with a different workflow. Not traditional dictation—none of that "period, new paragraph, comma" nonsense that makes you sound like a robot having a stroke. I'm talking about Voice-to-AI: dumping your raw, unfiltered stream of consciousness directly into an LLM and letting it handle the structure.
You speak like a human. Messy. Repetitive. Circling back to points. The AI listens, cleans, organizes, formats. You bypass the keyboard entirely.
I watched that same PM try it last week. He was hesitant at first—"It feels weird talking to my laptop"—but he gave it five minutes. He rambled for three minutes about a user onboarding problem. Pauses, false starts, "no wait, actually..." The kind of messy thinking that would have taken him twenty minutes to type.
Claude turned it into a clean, structured spec in thirty seconds. He stared at the screen like he'd seen a magic trick.
The Real Excuse
I know the objections. "I can't talk out loud in an open office." "I feel ridiculous dictating to a machine." Fair. But let's be honest about what's really happening.
Most people aren't avoiding voice input because of the environment. They're avoiding it because speaking your thoughts raw feels vulnerable. When you type, you can edit as you go. You can backspace, rephrase, polish. You present a finished thought. Voice forces you to expose the messy process—the half-formed ideas, the contradictions, the thinking that doesn't know where it's going yet.
But that's exactly where the value lives. The polished sentence is usually the dead one. The raw, chaotic monologue contains the insight.
The B2A Shift
In the Business-to-Agent economy, this isn't a gimmick or an accessibility feature. It's a fundamental shift in how human value gets extracted.
For years, brilliant strategists have been paralyzed by input friction. They have the ideas. They have the pattern recognition. But the keyboard acts as a gatekeeper, filtering out everyone who can't think in QWERTY.
Voice-to-AI removes the gate. Your output speed no longer depends on your finger dexterity. It depends on the clarity of your thinking and your willingness to let the machine handle the syntax.
If you've spent your career feeling like your internal monologue was always three laps ahead of your deliverables, you don't need to learn to type faster. You need to learn to think out loud.
The bottleneck was never your brain. It was your hands.
— James, Mercury Technology Solutions, Tokyo, May 2026
