Or it could be similar to how image generators generate text in images (by making things that look like text but is generally unintelligible nonsense).
Though if it trained on enough keys, it might have picked up on some of the correlations. I’d be surprised if they don’t use a database instead of just a checksum to determine valid keys, but if it was just a checksum, it’s possible that a NN could figure out how to generate valid keys.
They have a proprietary algorithm that generates keys and shoves them into a DB, in many cases they have a form of checksum to quickly check the validity of keys in-form to reduce DB lookups. Credit/Debit card numbers also have a checksum built in
If you had enough keys, it might be possible to train a NN to generate new ones, they are after all fantastic at pattern recognition it’s their whole schtick
Or it could be similar to how image generators generate text in images (by making things that look like text but is generally unintelligible nonsense).
Though if it trained on enough keys, it might have picked up on some of the correlations. I’d be surprised if they don’t use a database instead of just a checksum to determine valid keys, but if it was just a checksum, it’s possible that a NN could figure out how to generate valid keys.
They have a proprietary algorithm that generates keys and shoves them into a DB, in many cases they have a form of checksum to quickly check the validity of keys in-form to reduce DB lookups. Credit/Debit card numbers also have a checksum built in
If you had enough keys, it might be possible to train a NN to generate new ones, they are after all fantastic at pattern recognition it’s their whole schtick