if it’s img hash comparison 1. no need to upload the img (calculate hash client-side and send that) 2. can be bypassed by changing 1 pixelhttps://twitter.com/lorenzofb/status/928014255466668032 …
apparently they're planning on using photodna, which is tolerant to some degree (not aware of a real whitepaper) of trivial mods
-
-
if done clientside there's nothing stopping me from submitting arbitrary hash-alikes to blacklist but I guess they could check...
-
...whatever it would end up blocking if/when it got a match
-
interesting twist is when someone inevitably submits legitimately risque photos of someone else to get their posts removed
-
and fb has to decide whether to apply its autotagging system for authentication purposes
End of conversation
New conversation -
-
-
why would they use that instead of a neural network embedding and some kind of LSH? isn't that why they have FAIR?
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.