MirthTale
Reserve
Workshop

The first prototype was a printer, a microphone, and a stapler

We built the ugliest possible version to find out whether the idea survived contact with a four-year-old. It did, mostly.

The rule for a first prototype is that it should be embarrassing. Ours was. A second-hand thermal printer, a USB microphone, a laptop under a tea towel, and a stapler. It sat on a kitchen table in Berlin for four weeks and made about ninety books for nine children.

What we wanted to learn

We had three questions. Would a small child talk to a box? Would they accept a story the AI had mostly written and illustrated, as long as their choices were kept? And would they want to read it again the next day, on paper, with an adult?

What happened

The first question answered itself in the first hour. Children talk to anything that has a button. The second took longer. Some children were suspicious of the parts they hadn't said, and told us so. But when their own characters showed up on the page, doing what they'd said, the suspicion mostly turned into pride. One child spent ten minutes finding "her" fox in every picture.

The third question was the one that mattered, and the answer was clear. Books came back the next day, and the next. Parents told us the stapled thing lived on the bedside table alongside the real books. Two children asked for a specific book by title a week later. That was the evidence we needed that paper was right.

What broke

The staples. Constantly. Thermal paper fades. The microphone missed quiet voices. The printer was slow enough that children wandered off before the book appeared. Every one of these became a line in the specification for the real device: a sealed spine, permanent ink, a microphone tuned for small voices, and a book in the hand within a few minutes.

The laptop under the tea towel also had a screen, and children found it. That led directly to the decision we'll write about next: the display that goes dark.

All posts