I started this project with a practical problem: I don't have time to edit long conversations, and I don't have the equipment to make them sound good. So the one-sentence format was a workaround. Ask a short question, get a short answer, do something with it. That was the whole plan.
What I didn't expect is that the constraint would turn out to be the most interesting part.
Here's what I've noticed. When I ask people the longer version of the question — "how do you think AI is going to change your industry, what are you seeing, what are you worried about" — I get a certain kind of answer. Considered. Structured. Usually it starts with something like "well, on one hand" or "I think it really depends." Good-faith answers. But they're answers that have been partially pre-assembled. People have thought about this, or read something about it, and they reach for that when you ask an open-ended question.
When you tell someone they have to give you one sentence, something different happens. They stop reaching for the assembled answer and start reaching for the thing they actually think. It's not always clean — sometimes people give me a run-on sentence that's doing the work of a paragraph, and I have to gently say "okay but make it one sentence." But even in the negotiation, even in the pushback, you get something real. The resistance itself is often more revealing than the polished version would have been.
A teacher I spoke with in November spent about forty-five seconds staring at the recorder before she said anything. Then she said: "I keep hoping my students will use it badly enough that I can tell." That sentence took her most of a minute to arrive at, and it's a better description of a specific kind of classroom anxiety than anything I've read in the coverage of AI and education. Because it's honest about the dynamic — the teacher's hope that the tool will fail visibly, so the role of the teacher stays clear.
She wouldn't have said that in a longer interview. She would have said something about learning styles and academic integrity and critical thinking. True things. But not that thing.
I think the reason the constraint works is that it removes the permission structure of the considered answer. When you have time and space to answer, you answer in a way that represents you fairly. When you have one sentence, you just answer. And what comes out is usually the thing that's been sitting closest to the surface, waiting.
I'm not claiming this format is journalism or research. It's a recorder and a question and a minute of someone's time. But I've started to think the brevity is doing actual work — not just making a shorter product, but creating a different kind of honesty. One that longer formats accidentally edit out.
I'll keep going until I run out of people willing to talk to me, or until I have enough to know what I'm actually trying to say. I'm not sure which comes first.