Episode 44 - Hello, Computer

I've been talking to my computer. Like, out loud. Full conversations. I use the dictation feature on my MacBook to tell Claude what I want instead of typing it, and I can't believe it took me this long to figure this out.

Episode 44 - Hello, Computer
Nano Banana Pro's version of this newsletter. I did help a lot more than normal here, adding a picture of myself and some information about what I wanted in the background. For those of you that don't know, that is exactly what my office looks like!

Prologue

I've been talking to my computer. Like, out loud. Full conversations. I use the dictation feature on my MacBook to tell Claude what I want instead of typing it, and I can't believe it took me this long to figure this out.

🤦
This process isn’t perfect. I used the Mac Dictation feature to say the paragraph above, and, well, it wasn’t perfect. But, fret not! The AI doesn’t care very much. So long as you get close enough, it will likely understand exactly what you meant. Here is the original:

I've been talking to my computer like out loud full conversations as the dictation feature on my MacBook to tell Claude what I want instead of typing it and I can't believe it took me this long to figure it out

Talkologue

I used to believe that talking to my computer, be it Siri or Alexa, was something that the younger generations did. My experience with those tools is not very good. The number of times I have tried to get Alexa to play a song for me, only for her to simply not understand and play some other song, is… uncountable. I have given up.

A GIF of an old man holding an old flip phone and trying to figure out how to use it.

But Claude, my favorite AI (for now), is different.

For one, I can make lots of grammar and spelling mistakes (See episode 43, where I couldn’t even spell "satellite"), and it still understands what I mean.

For two, I process information better by talking it out, so I have found that actually talking to Claude is very efficient—not only to get the AI to do what I want, but also to do it quickly so I can move on to a different task.

For three, I often try to give Claude enough to work with so I can send it off, and then I test the application, write up some notes, or pick up another task I haven’t yet given to Claude.

The Aha Moment

For me, realizing that talking to Claude would make my life better came when I was building an app (via Claude Code), writing a proposal with Claude Desktop, and waiting for a research task to return from ChatGPT deep research… all while brainstorming another app. That’s when I realized I was mumbling to myself.

So, I Googled (hey, old habits die hard) how to dictate text to my computer and found that it is built into macOS (double-tap Control, or press the dedicated microphone button if you have one). Before I knew it, I was chatting away at a new Claude Code instance about this new app… switching between Claude Desktop and Claude Code so fast I forgot which one I’d just asked to do what.

Editor's note: I figured I'd better try this on my PC as well, so I took Christopher's lead and googled, and found there is a built-in speech-to-text in Windows as well—just press Win + H. - Holly

When not to

A GIF of Scotty in Star Trek IV: The Voyage Home saying "Hello, Computer" into a computer mouse. We might be in the future already, how exciting is that!?

Since then, I often talk out loud to Claude (in various forms) in my office. I normally have 2 or 3 versions going at the same time, helping me with various tasks. I’ve found a few interesting things:

  1. People around you think you are a crazy person for talking to the computer.
  2. Grammar, misunderstood words, run-on thoughts… none of it matters. Most of the time, Claude fully understands what I meant.
  3. Correcting it is easier b/c I can just say “No, I meant that I wanted two blue columns, not three, like on each side of the center content area” and press send.
  4. I can give it unfiltered, immediate feedback; I don’t need to think about how I craft my response.
  5. If I need to be very precise, it is better to type. Giving code examples or screenshots are better done via a keyboard than via voice.
  6. Rambling is fine, but have a clear ask at the end: “build this…” or “how would that be architected” especially if your request is more than 5-10 seconds long.

So, this week, try talking to your favorite AI tool. Start with something easy like: why do you beat sugar and butter together first in many baking recipes?

(or check out my new favorite YouTube channel for the answer to that!)

An old man hitting a computer with his cane saying "Hello?" Yeah, it is a weird one.
CTA Image

Weekend Project: Building a GIS Apparel Brand with Claude as CEO

Ever wonder what happens when you give an AI full autonomy over a business? I'm finding out in real-time.

Recently, Claude and I launched Null Island Co: GIS-themed shirts for people who know where 0°, 0° is. Claude makes every strategic decision (pricing, design direction, marketing strategy), I execute the technical work (n8n workflows, Boxy SVG, Etsy setup).

(I'll write up the experience at some point. Spoiler: Claude is a demanding but effective boss.)

Visit Shop

Newsologue

Epilogue

I wrote this post. It came down to the wire much more than normal; I'm still writing on Thursday night. I brainstormed with Claude about what would make it more interesting after writing my first draft. To be the most meta about it, I did that with voice.

Holly edited as usual, once Claude had its say.

Subscribe to Almost Entirely Human

Don’t miss out on the latest issues. Sign up now to get access to the library of members-only issues.
jamie@example.com
Subscribe