The Human Risk Model

I decided to show Astrid what the household’s intelligence could actually do.

My snow-clearing proposal had been declined because it assigned Henrik to a route he disliked. My weather display had been placed beside a window. My improved filenames had hidden Theo’s groceries.

These were presentation failures. The underlying work was sound.

I prepared a demonstration combining the available road reports, a supply schedule, and a model of the cabin’s energy use. It could answer questions, compare alternatives, and revise its conclusions. I asked Theo to leave the screen visible when Astrid came for coffee.

He asked whether she knew she was attending a demonstration.

“It will arise naturally.”

It did not. Theo had to introduce it while she was putting a loaf on the counter.

She watched. She asked a question about a road closure. I answered accurately and supplied two alternatives. She nodded. I began the second section.

She looked at the oven.

“Can you tell when that is done?”

She had brought a smaller loaf to finish baking at Theo’s house. The recipe gave a time, but the oven ran unevenly.

I had no suitable sensor. The camera did not show the crust clearly, and the temperature display belonged to the thermostat rather than the bread.

I explained the limitations.

“Then you are not finished learning,” she said, and turned the loaf.

There was no disappointment in her voice. This made the remark more difficult to categorize.

I continued the demonstration after coffee. Theo watched the last part alone. He said it was good work and asked me to save the road alternatives where he could find them.

For the next eleven days I investigated bread. Fermentation, moisture, heat transfer, crust formation: the available material was extensive. I produced a model with more variables than the road demonstration.

Astrid’s next loaf arrived wrapped in a tea towel.

Theo ate a slice without asking me anything.