Comment by johndhi

I'd be curious to see whether these models can create new forms of unbreakable encryption themselves!