Three lessons of typing, and you've gotten good at it. Today you put the keyboard down. In this lesson you'll press one button, say something out loud, and hear it answer you - in a voice, like a person on the phone. This is the lesson where the whole thing stops feeling like a computer program.
And here's the secret about why this one's easy: you've been practicing for it your entire life. Every phone call, every chat over a fence, every question you've ever asked out loud - that's the skill. There's nothing new to learn. There's only a button to find.
Find the button
Sign in at chatgpt.com, like always. Look at the box where you normally type. At its right-hand end live two small pictures: a plain microphone, and a set of little sound waves. They do different things, and we'll use both today. The sound waves are the phone call - a spoken conversation, voices both ways. That's where we start. (The microphone gets its turn in a few minutes.) On a phone or tablet with the ChatGPT app, both live in the same spot. The pictures get redrawn from time to time - when in doubt, rest your pointer on a button and it will tell you its name.
The question your computer will ask you
The very first time you press it, before anything else happens, your computer will pop up a little question: may chatgpt.com use your microphone?
Stop and appreciate that popup. That's your computer doing its job - nothing on the internet can hear you until you say so. Remember last lesson: it only knows what you send it. This popup is that promise, with a doorman. Click Allow, and the microphone works for this conversation the way a phone works during a call: while you're on, and not otherwise.
Now just talk
- Press the sound waves.
- Say something. Out loud. A full sentence, like you'd say to a person: "I'm thinking about planting tomatoes this fall - is it too late?"
- Pause. It answers, in a voice.
- Answer back. Interrupt if you want. Ask it to slow down, speed up, start over. It doesn't mind.
- When you're finished, press the button that ends the conversation - it's right there on the screen, like hanging up.
The first ten seconds feel strange - everyone talks too formally at first, like they're leaving a message on an answering machine in 1987. That passes. By the third exchange you'll just be talking.
When voice beats typing
Some conversations are better out loud, and you'll find your own list fast. When your hands are busy - cooking, folding, fixing. When your eyes are tired. When what you want to say is long and rambling and would be miserable to type - voice loves rambling; say it messy and let it sort you out. And when you just want the company of thinking something through with somebody, out loud, the old way.
You don't have to go all in
Here's the nuance most guides skip: talking and listening are two separate choices, and you get to make them separately.
Remember the plain microphone, the other button in the typing box? That one is dictation: press it, talk, and your words appear as typed text in the box, as if you'd typed them. Press send, and the answer comes back in writing, for you to read at whatever pace you please. You talk your half; you read its half.
And it works the other way around, too. Under any written answer there's a small speaker. Press it, and that answer gets read to you out loud, in a pleasant voice, while you do something else.
So really there are four ways to have this conversation: type and read, talk and read, type and listen, talk and listen. None of them is the right one. There's only the right one for this moment, this task, this mood - and now you own all four.
One of those four, as it happens, is how this very site gets made: one half talked out loud, the other half read off the screen in writing. Try all four and see which one is yours.
A few things you might be wondering
Is it listening when I'm not using it? No. The microphone runs between the moment you start a voice conversation and the moment you end it - that's the whole window. Your browser even shows a little indicator while the microphone is on, so you can see it with your own eyes.
Do I need special equipment? No. Laptops, phones, and tablets all have a microphone and speaker built in. If it can do a video call, it can do this.
What if it mishears me? Correct it, out loud, like you'd correct a person: "no, I said tomatoes, not potatoes." It takes correction gracefully - better than most people, honestly.
Can I reread a spoken conversation? Yes. When you end a voice conversation, its words are waiting in your chat in writing - like minutes from a meeting: faithful to what was said, if not always word for word.
What this looks like in real life
Nikki found her reason the first week. Bread day, both hands in the dough, phone propped against the flour canister:
Nikki, hands full, out loud
"My dough is way stickier than usual and I've already added flour twice. Same recipe as always. It's humid today - could that be it?"
And it talked her through it while she kneaded - yes, the humidity; no, don't add more flour; wet hands instead. She never touched the phone. The bread came out fine, and she got to keep both hands where they belonged: in the dough.
Notice she didn't compose a tidy question. She thought out loud, the way people do. That's not a workaround - that's the feature.
That's the lesson
That's part one of the road, finished - all four lessons. You can talk to an AI, it answers your way, you've read its notes and hold the off switch, and now you can do the whole thing out loud with your hands in the dough. You're not getting comfortable anymore. You're comfortable. Part two is where we put it to work: the everyday dozen - a letter, a trip, tonight's dinner, and more.