I Stopped Typing for a Week. Here's What Actually Happened.

I didn't do it because I had to. My hands were fine, my typing speed was decent, and I wasn't looking for a new system. I did it because I kept hearing that dictation was faster and I didn't believe it. So I ran the experiment myself.

The rules were simple: no typing for content creation for seven days. Emails, articles, notes, messages, all spoken. Passwords and search queries got an exception. Everything else was voice.

Day One Was Embarrassing

I talked to my computer like I was leaving a voicemail for someone I didn't know well. Stiff, over-enunciated, weird. I kept stopping to re-read what had been transcribed instead of just continuing to speak. The output was choppy and I hated it.

I was using VoiceInk, which transcribes locally and fast, so the latency wasn't the problem. The problem was me. I had no idea how much of my "writing process" was actually me reading my last sentence and using it as a springboard for the next one. Without a keyboard, that loop felt broken.

By Day Three, Something Shifted

I stopped trying to write out loud and started trying to talk through what I meant. The distinction sounds small but it changed everything. Instead of composing sentences, I was explaining ideas. The prose got messier but the ideas got clearer.

I wrote a 900-word piece that morning in about 22 minutes. That same piece would have taken me closer to 50 minutes at the keyboard. The draft needed more editing than usual, but there was genuinely more to work with.

The Surprising Hard Part

Short-form text was harder than long-form, not easier. Dashing off a two-line Slack message felt absurd to speak aloud. I caught myself typing those anyway and not counting it as cheating, which tells you something about how context-dependent the tool is.

Long-form content, articles, outlines, detailed emails, that's where voice earned its place. The longer the thought, the more the speed advantage compounded.

What Broke

Code. I write occasional scripts and nothing about dictating code felt natural. Variable names, syntax, brackets, it was a mess. I made an exception by day four and stopped pretending otherwise.

Also, open offices. I work from home, so this was mostly fine, but I had one afternoon at a coffee shop and I simply did not dictate in public. Some tools are context-specific.

What Surprised Me

I talked more than I typed. Not slightly more, significantly more. My daily word output roughly doubled. A lot of that extra volume was noise, hedges and repetitions that would have been trimmed by the act of typing. But about half of it was real content I would have simply left out before because the typing felt like too much effort.

I also noticed that voice captured a more natural version of how I actually think. My typed writing tends toward shorter, more clipped sentences. My spoken writing was warmer and more direct. Not always better, but different in a way I hadn't expected.

What I Kept

I still type. But I now dictate anything that's going to be more than three paragraphs. Articles, long emails, meeting notes, rough outlines, all spoken. The keyboard handles everything short, technical, or public.

The split took about a week to feel natural and another week to feel obvious. I don't think of it as a replacement anymore. I think of it as the right tool for a specific kind of output, the kind where volume and velocity matter more than precision in the first pass.

If you've been curious about dictation but haven't committed to a real trial, a week is long enough to get past the awkward phase. That phase is real, but it's also short.