it gets even more fun when you can generate images that are imperceivably different to the human eye, but produce different (and controlled) hash values. (see

it gets even more fun when you can generate images that are imperceivably different to the human eye, but produce different (and controlled) hash values. (see shazam decoys and adversarial images). not sure if collision is really the right term here though as similarity hashing intentionally hashes multiple images to the same hash. in other news, my google phone just helpfully scanned all client side images, rounded up everything it thought was a meme and suggested i delete them to save space. i wonder if that feature has telemetry built in...