The original note, as written at the time
A micro camera design that you can strap onto anything. Good for recording inspections. Good for keeping kids safe when playing in the garden — it can identify snakes and dangerous bugs using AI. We could build these right here in Uganda. Just need to connect a camera to the wireless feature.
Ongoing development
A micro camera you can strap onto anything. Good for recording inspections. Good for keeping children safe in the garden — it can identify snakes and dangerous bugs using AI. And we could build these here in Uganda: just connect a camera to the wireless feature.
The two examples are the argument, because they point in opposite directions and both are right.
Inspections are dull, repetitive work done by a person who has to hold a torch, a clipboard and a phone at once, in a space that is sometimes dangerous. A camera that straps to a hard hat and records continuously turns a report written from memory into a record that cannot be argued with. That is the professional case and it is enormous: every building in this country is inspected by people whose evidence is a hand-filled form.
The garden case is the domestic one, and the interesting part is that it is local knowledge. African parents know which snakes matter, and a model trained on European gardens would be useless. The value is entirely in the dataset, not the camera.
The manufacturing claim is the one I would push hardest. This device is a camera module, a battery and a radio, and all three are commodity components sold by the reel. There is no part of it that requires a factory we do not have; what we do not have is the assembly line and the confidence. That is a policy and capital problem, not a technical one.
The ethical line needs drawing now rather than later. A camera strapped to a child is a camera that records other people’s children. Any product here must be local, must not upload, and must delete by default.
Build the cheap one first. The cheap one is the one that saves a life.