Everything you might want to know about SeeFood, the working Not Hotdog app inspired by HBO's Silicon Valley.
Structurally, yes. A filling inside a partially sliced bun fits most definitions of a sandwich. Culturally, almost nobody calls it one, and neither do we. SeeFood sidesteps the debate entirely by only caring about one thing: is it a hotdog, or is it not a hotdog?
SeeFood is a free web app that looks at a photo from your camera and tells you whether it is a hotdog or not a hotdog. That is the entire product. Point it at food, get a verdict.
It is a fan-made tribute to it. In HBO's Silicon Valley, the character Jian Yang builds an app called SeeFood, pitched to investors as the Shazam for food, that only ever learned to recognise hotdogs. This is a real, working version of that joke.
It is a parody and is not affiliated with HBO.
Jian Yang is the character on Silicon Valley who builds the Not Hotdog app. Asked to make the Shazam for food, he trains it only on hotdogs, so it can confidently identify a hotdog and label absolutely everything else not hotdog.
In spirit. A true Shazam for food would name any dish you point it at. SeeFood proudly recognises exactly two categories: hotdog and not hotdog. Think of it as a Shazam for food with very strong boundaries.
SeeFood runs an image classifier directly in your browser using TensorFlow.js and a MobileNet neural network. When you take a photo, the model scores it on your own device and decides whether it is looking at a hotdog.
It is free and runs in any modern browser. There is nothing to install. On a phone you can add it to your home screen and it will open like a normal app.
That one is on purpose. The dachshund is the single dog SeeFood is willing to break its own rules for. Point it at one and it returns a hotdog verdict, complete with a bark. It is the app's only known bug that is also a feature.