World to Gif uses Advanced Artificial Intelligence (really!) to do a very simple thing: transform your pictures into the most appropriate GIF, using the magic of giphy.
How does World to Gif work?
Well, when you take a picture, or select a picture from the phone’s library, World to Gif analyses the picture using machine vision and AI. First, it analyses the style of the picture using a model based on this paper and on this great book about Swift and AI. Then it looks for the most salient objects in the picture using the Resnet50 model available at Apple’s webpage.
Once the app knows the dominant style of the picture, and which objects it can recognise, it will search on giphy for the most appropriate gif that uses those keywords. The app then displays the gif, and if you tap on it you can see what the computer saw.
What is the point of this app?
Well, the website is called ridiculous software, so …
Can I save my pictures?
Well, in an age in which we want to burn the environment to have ownership of fleeing beauty, I decided that World to Gif saves nothings. Everything is temporary. Enjoy that.
Why is this app so ugly?
Well, the point is not making something pretty, and [INSERT LAME ARTISTIC EXCUSE HERE]. But more honestly, I suck at UI design. If you want to make it prettier, contact me.
Well, there are mainly two arguments: first, I thought it would be stupid and funny to use the powers of the internet and AI to find the hidden gifs everywhere.
But there may be a bit more to it: GIFs are such an online thing, and photography is being so often used as an access to the natural in our permanently online world, that I thought it would be fun to connect an artificial construct (of the past) with an artificial construct (of the future). Media rewriting media, I guess.
How was this made?
100% Swift, using Sebastian Boldt’s JellyGif and the Giphy API.
World to Gif does not collect any data at all.