Saturday, half past seven in the evening. Tomasz and Izabela have popped out for a couple of hours; grandma Elżbieta is minding the children. Basia, five, is running a temperature. Grandma goes to the drawer for the fever syrup, picks up Iza’s phone — and before she’s opened anything, one line on the notification: “Warning: allergy to this medicine (Basia)”. This scene really happened. This episode takes it apart.
How did the phone “know”? It didn’t. Three weeks earlier Basia came out in a slight rash after ibuprofen, and the paediatrician said to write it down. Iza opened the medicine, tapped “Flag allergy”, chose Basia, typed in the substance — a few seconds in the waiting room. The app didn’t diagnose anything. The doctor did that, and a person typed it in. The app only remembers — and reminds you at the moment it’s easiest to forget.
The catch: the box from the drawer was a completely different brand from the one a month before. Different name, different logo. And yet the warning came up — because the app doesn’t look at the name on the box, it looks at the active ingredient inside. A different box, the same substance. That’s exactly when it’s easiest to slip up: you buy “something for a temperature” under a new name and you don’t connect it in your head with a rash from a month ago.
The episode is honest about the limits. Matching by substance is the word “recognises”, not magic — a second pair of eyes, not the only pair; it doesn’t let you off reading the packet. The app warns about exactly what you entered, and about the same substance in other boxes — but it doesn’t guess about whole families of medicines it hasn’t been told about. Joining up related medicines is for the GP or the pharmacist. And one hard rule: the app will never block giving a medicine. Warn, don’t block — so you make the decision seeing the warning, not next to it.
The second half is the mechanics you don’t spot at first glance:
The conclusion is the same all the way through: the app is memory. The knowledge is the GP and the pharmacist. It wasn’t the phone that saved anyone — it was Iza spending a few seconds to write something down, and the app handing it back at the moment it mattered.
Topics:
Materials:
⸻
About AI in this episode. The dialogue was written by Tomasz Szuster. The voices you hear are AI-generated synthetic voices — not recordings of the hosts. We say this straight out, because we think it should be the standard for any podcast made this way. Full details in the show notes and in the episode outro.