Uninformed response: if two different universal approximators can have different modes/rates of learning, then even if they're usable in theory, they could be l

Uninformed response: if two different universal approximators can have different modes/rates of learning, then even if they're usable in theory, they could be less good in practice.