Hacker News Reader

Google has some serious chops when it comes to image recognition and machine learning. Telling the difference between a photo of a car and a photo of a train is much more difficult that telling the difference between a real photo and some random file encoded as a photo.

Pretty sure it's like one afternoon's work for any google engineer to write some code to figure out when something like this is happening.

They can check the amount of "randomness" in each photo to determine whether it's an actual photo or a disguised file.

You could probably use some steganographic techniques to embed your data in what looks like a real photo prior to uploading though c: