>> The world's foremost proponent of not trusting standardized parameters wrote and repeatedly submitted a paper demonstrating a model attack that targeted curv

>> The world's foremost proponent of not trusting standardized parameters wrote and repeatedly submitted a paper demonstrating a model attack that targeted curves built from mathematical constants; when you've "backdoored pi", there's really no place to go from there. That sounds like an argument against standard parameters. Even if we trust an algorithm, we can't trust others to choose parameters for them? The solution seems obvious - find a better way to choose (non-standard) parameters.