The proposed solution involved changing the validation algorithm so that transactions could not be replayed on both networks. This would be done by requiring a unique child transaction on each network that would be necessary for the transaction to be valid. By implementing this policy change simultaneously in both networks, old pre-signed transactions could only be replayed by the owner of the coins, as they would have to spend them immediately in the child transaction. Once spent, the UTXO sets would diverge across the two networks, effectively preventing replay attacks.