I've always liked autometically generated avatars in theory. But in practice they never seem that useful because there's nothing recognizable in the image for m

I've always liked autometically generated avatars in theory. But in practice they never seem that useful because there's nothing recognizable in the image for me to subconsciously recognize. Secret had an interesting take on this. They had a limited set of simple icons, like boats or rockets, and then dynamically colored them. They were easy enough to recognize so patterns became more apparent.